Commissioned

Personalized Health Management:
A Geisinger View

Walter F. Stewart, PhD, MPH
Associate Chief Research Officer
Director, Center for Health Research

J. B. Jones, PhD, MBA
Research Investigator I
Center for Health Research

Ron Paulus, MD, MBA
Executive Vice President Chief Technology and Innovation Officer
Clinical Innovation Department

Mark Selna, MD
Associate Chief Innovation Officer
Clinical Innovation Department

Geisinger Health Systems

INTRODUCTION

The US healthcare system is beset with persistent structural challenges that continue to erode the quality of care while simultaneously increasing cost and hindering access (1-4). Structural changes in how care is delivered and financed will be required to overcome these challenges, as will a fundamental shift away from the traditional role of what it means to be a patient in the U.S. healthcare system. We believe the latter shift will involve a cultural change in how the individual is viewed, transitioning from a classic “patient role” to an engaged “consumer role.” In this paper, we set forth the concept of “personal health management” (PHM) to describe a process, supported by a set of tools and technologies, in which consumers assume an increasingly active role in how their health is managed, in personalizing the value proposition of their purchased health care services, and in determining how and which health care services they use. With deliberate meaning and intent, we use the terms “consumer” and “patient”, respectively, to distinguish “a shift towards more control” versus “a passive participant in a care process”. In this paper, we describe the vision for PHM at Geisinger, where innovative processes, supported by a set of tools and technologies, are adopted based on sensible business principles and payor-linked incentives, all with the purpose of changing behavior to move from a provider-centric to a consumer-centric model.

We currently envision seven objectives essential to implementing a fully developed PHM model. Objectives 1 and 2 provide the foundation for all other objectives. Objectives 3-4 support the consumer. Objective 5 supports the provider. Objectives 6 and 7 facilitate consumer-provider communications.

All data relevant to consumer PHM should be valid, integrated, and readily accessible.
Health care and information access channels should be diversified to provide for timely and efficient encounters.
Encounters should be based on content specific, personalized, consumer relevant data.
Where sensible, encounters should be informed by consumer choice.
Clinicians should have ready access to the right data, information, and expert guidance.
Shared decision making should be seamless and routine during every encounter
PHM should be dynamic, guided by consumer preferences, current status, and evidence.

One of the hallmarks of the US healthcare system is the persistent lack of alignment among payors, providers, employers, and consumers to achieve high-quality care in a cost-effective way (5). Indeed, there is no single stakeholder who is focused on the overall value of the care process (i.e. optimizing the incentives and objectives of ALL stakeholders). Cost-shifting among stakeholders (i.e., employers, insurers, providers) has been the dominant means of solving periodic market or policy challenges (6). Consumers are the latest to “participate” in these cost shifting efforts via high deductible plans and so-called “consumer-directed” plans (7), a change that reflects the extreme pressures on US employers to reduce the impact of healthcare coverage on their bottom line, paralleling the shift from defined benefit to defined contribution retirement plans(8). These plans have been successful in realigning incentives, but without a concomitant change to the delivery system or tools needed to navigate it; furthermore, these consumer-directed initiatives do not inherently satisfy any of the primary objectives of PHM.

In many ways, the anatomy of Geisinger as an integrated delivery system – with its multiple hospitals, a geographically-dispersed multi-specialty group practice, and a separate (non-exclusive) insurer – make it a microcosm of the larger US healthcare system, struggling to meet the challenge of aligning incentives and processes to satisfy its major stakeholders. Specifically, throughout the Geisinger service area and throughout the US, the greatest near-term strategic/political challenges are financial - per–capita costs are increasing, patient average age is increasing, per-capita clinician supply is decreasing, and below-cost payers (e.g., Medicare, Medicaid) are becoming the increasingly predominant payor. As a business imperative, Geisinger is seeking to close the gap between what Medicare pays and what it costs to provide care. We believe that PHM is central to closing this gap.

Our belief in the potential to alter a complex marketplace through empowering the consumer is not without precedent. The financial planning market was once dominated by a paternalistic model in which high-quality advice and information were available through “knowledgeable experts” primarily to those with the resources to pay for it. With the transition from defined-benefit to defined-contribution pension plans, a consumer focused shift was inevitable. Subsequently, new tools and processes (e.g. web-based asset allocation guidance, online trading, etc) were developed that effectively allowed consumers to become, if they so desired, their own financial planners. In this environment, consumers are allowed to make “bad decisions” (e.g., sole selection of a money market account), if they want to, without a “parental influence” over what they do. At the same time, they are guided by increasingly sophisticated, but easy-to-use programs (e.g., risk profiling tools that map to fund allocations, automated rebalancing, age-based shifts in asset allocation, etc.) that rival even the most expert financial advisors in terms of performance. The focus is on providing the consumer with high quality information, tools and back-up human support; while the business seeks to influence the selection of an “optimal” choice, it doesn’t feel “responsible” for their ultimate choice. Although not without challenges, few can argue that the revolution in financial planning has increased access, reduced costs and dramatically improved financial performance for many millions of consumers.

In the sections that follow, we describe in greater detail the key elements, process and tools that will be important to our PHM model, followed by our five year plan.

PHM OBJECTIVES

In this section, we consider the primary needs of and challenges faced by both consumers and providers as their roles evolve in a PHM model. We also outline the processes and tools relevant to the PHM objectives, with a specific focus on tools and processes that integrate multiple data sources and integrate with our EHR. Throughout this paper we use the terms “tools” and “processes.” Tools serve a single function (e.g., data capture) whereas processes involve two or more tools that are linked and/or functionally dependent (e.g., patient data capture and clinical decision support) or represent an existing human process transformed by the introduction on a new tool.

Health care reform efforts over the past 30 years have essentially been focused on “tinkering” with the existing system, rather than systematically reengineering processes to support consumers and providers in a consumer centric model. Health information technology is critical to reform efforts (9, 10). Use of an electronic health record (EHR), in particular, is proving to be critical to re-engineering processes - importantly, though, not as a self-contained solution, but rather as a multi-process/tool integration platform. We view the EHR as an important focal point for the development of new processes and for the development of supporting tools that will diversify options for providing, managing, and monitoring care. Additionally, we believe that the development of tools that interact with, but which are distinct from, the EHR represent a market segment with the potential to spawn rapid, industry-changing innovations as individuals, companies, and health care systems seek to capitalize on the opportunity.

Objectives 1 and 2: Foundational

Real-time unfettered access to data from disparate sources, including an EHR, is essential for the continuous evolution towards a PHM model. We consider our own work in this regard and implications for PHM, recognizing that there is more than one solution. In the past 24 months, to satisfy several explicit business needs, Geisinger has created a comprehensive enterprise-level data warehouse, a resource that has transformed our thinking of what is possible. The warehouse receives feeds from multiple source systems, including our EHR, financial decision support, claims, patient satisfaction and high-use 3^rd-party reference datasets. The source data are transformed through a standard “Extract-Transform-Load” (ETL) process. Expansion to additional data source systems (e.g., niche specialty systems such as oncology) is being accomplished in stages and will eventually encompass all high-value data sources. While the warehouse will significantly advance our ability to perform traditional activities such as performance reporting, trending, self-service data access and other similar tasks, the most important value of the data warehouse is as a foundation for advanced analytics and as an automated real-time data source for PHM processes, enabling what we describe as our Clinical Decision Intelligence System (“CDIS”). Because CDIS represents a standardized, normalized, multi-year dataset for our entire population that is accessible using intuitive yet robust business intelligence applications, it serves as an ideal analytic foundation. Current examples include mining the database to identify patients in need of certain interventions prior to a visit (e.g., diabetics with a HgbA1c > 7 for at least two of the past four quarters), identifying “open loops” (e.g., all patients with a positive PSA test without a visit to a urologist, an intervention or a subsequently normal PSA within 3 months) and identifying correlations in treatment and outcomes (e.g., association of patients in payor class x who fail to fill a prescription for disease y). We anticipate that CDIS will serve as the primary historical data feed for our PHM tools, for performance trending and as a near real-time source for data back into our EHR. It is also likely to be a resource to retain and effectively use many different types of derivative and reference data (e.g., libraries of clinical rules, data capture tools, community resources and other geographic information based data sources, insurance costs, co-pays, formularies, etc).

In any clinical business, physical space is expensive and costs (e.g., maintenance, utilities) are likely to increase over time. For this reason, many businesses adopt metrics that are designed to encourage optimal use of space (e.g. revenue per square foot). In healthcare settings, clinic space itself is frequently used in an inefficient manner. For example, large waiting areas serve as a holding space to ensure that exam rooms are efficiently used, but the waiting area itself is not simultaneously used as a “working” area. It largely represents an inefficient use of space. Converting waiting areas to exam or working areas, where privacy is assured increases the amount of active clinical space and provides the means to put consumers to “work” shortly after they arrive in a clinic (e.g., educational or data gathering kiosks in the exam room, group therapy rooms). More generally, largely confining care encounters to the traditional clinic space limits consumer access both because it takes time to get to a clinic and transportation costs are increasingly material to the consumer. PHM should be devised to ensure ready access to care, where the mode of access is tailored to the level of consumer need, risks, complexity, etc (Table 1). Leveraging a diversity of access channels (e.g., retail clinics, guided email exchange, remote online encounters, phone calls, remote monitoring, telemedicine, etc.) can both dramatically increase timely access and reduce the total cost of an encounter.

Objectives 3 and 4: Supporting the Consumer

Individuals utilize health services for varied reasons: because they want to maintain optimal health, because they are aware of or concerned they may have a health problem, or because they seek resolution for specific acute and/or urgent problems. When consumers have a defined health problem, and even as they seek to maintain health, they generally want assurance that their chosen course of action is reasonable/feasible or, in some cases, that “doing nothing” is a sensible approach to care. Meeting these needs requires bi-directional communication between consumers and their various decision support resources; specifically, consumers want to know what health and/or lifestyle behaviors can benefit them, what the associated cost will be, and what, if any, evidence exists to support the comparative value of available interventions (e.g., probability of a positive outcomes, risks, uncertainties, etc). If action is taken, the consumer needs to know (in real-time, to the degree possible) how well the intervention is working; if it isn’t working, consumers need to know whether a change in strategy is required. At the same time, consumers will seek to provide information and feedback to their clinician (e.g., risk tolerance, preferences for specific interventions or therapeutic modalities, symptom reporting, affordability factors, etc) so that the development and modification of the care plan reconciles what the consumer should do, optimizing the intersection with what they want and are likely to do. Our goal for PHM is to meet these needs by systematically eliciting information from consumers and by facilitating their ability to assume substantial decision responsibility and control. Some consumers will make seemingly irrational choices (e.g., “I do not want to do anything about my diabetic risk of limb loss”). Importantly, this kind of choice is crucial to communicate to the provider as a key point for discussion, and one that would remain “hidden” and result in deleterious “non compliance” if not explicitly identified and addressed. A growing literature indicates that value added processes already enable patients to successfully contribute subjective and objective information, preferences, and responses to therapies and to manage treatments themselves (11, 12).

Obtaining Data and Information

The method by which patient-specific information is obtained on a patient has not changed substantially over the past century, even though office visit time and reimbursements have become increasingly constrained. Most verbal interactions between providers and patients occur in a somewhat unstructured manner. Information gathering, documentation, and interpretation are unsystematic and unnecessarily consume precious time from providers. Fortunately, much of the data currently obtained through conversation can be collected from the consumer before the clinician encounter via automated interactions (e.g. exam room-based touch screen enabled screening tools). By so doing, the breadth, depth, quality and utility of those data can be substantially improved, the efficiency of care processes can be increased, clinicians will be better informed, and the reimbursement value of a visit can be increased. Importantly, use of such a process provides the means to re-purpose data (e.g., provider guidance, tailoring intervention options to patient features and preferences, patient education, informed decision making, etc) in ways that give the patient an active voice in the care process and a means to guide their own decision-making.

Over the last 30 years, researchers have developed myriad valid patient-completed questionnaires intended to facilitate administrative management and clinical care decision-making. While potentially useful, existing questionnaires have not been widely used in traditional paper-based practice settings because the workflow is problematic (e.g., provider must interact with the patient while reading the questionnaire), the administrative task associated with choosing questionnaires and interpreting responses (e.g., scoring) is burdensome, the impact on the sequence and timing of existing work flows is difficult and their use in care planning/surveillance is not standardized.

We view use of patient reported data capture tools as a cornerstone to virtually ever other process we envision to support consumers and providers. Beginning in 2003, we experimented with different approaches (i.e., digital pen, scan form, pen tab, touch screen) and workflows (i.e., web portal, waiting area, exam room, etc.) to integrate patient data capture and use of such data in real time as a routine part of the care process. This work continues with a comprehensive strategy to capture and interactively use patient-reported data on diagnostics, medication reconciliation, review of systems, preferences solicitation, outcomes monitoring, etc, to: 1) optimize the settings, timing and reliability of patient-reported data capture for clinical purposes; 2) optimize the accuracy, completeness, specificity and utility of patient-reported data for clinical and administrative purposes; 3) minimize the workload burden - for patients and staff - associated with obtaining data; and 4) fully inform the provider. Our long term strategy will rely on the use of a Patient Reported Data Acquisition and Management System (PARDAMS), that integrates the following: 1) a rules engine that decides which questionnaire module should be used to collect data from a patient before (e.g., screening), during (e.g. diagnostics), or after (e.g., outcomes) a given encounter; 2) a library of well validated questionnaires, with an increasing emphasis on sophisticated use of branching logic and on the use of dynamic questionnaires for assessing outcomes; and 3) a rules engine that intersects the information gained via the questionnaire along with data in our other systems and tools (e.g. data warehouse, decision support engine) to recommend some form of action.

Consumer Choice

When a consumer seeks care, they usually face a complex process over which they have relatively little perceived control. The patient is often a passive recipient of information, where comprehension may vary substantially and is rarely verified. We believe that consumer control, consumer activation, and improved efficiency and quality of care can be achieved through the use of sophisticated automated and semi-automated consumer guides. We do not think it is relevant to ask “do patients want to assume control?” Rather, the more sensible question is “how can systems, processes, and tools be designed to motivate the consumer to be in control?” To some degree, the lack of consumer control is a failure of the delivery system. In this section, we first consider factors that will influence the level of consumer control that is sensible and then specifically consider tools that will be useful in this process.

Table 1. Relation between how much decision control and intervention control the consumer can assume in relation to the risk from intervention, quality of evidence, complexity of the decision process or intervention
RISK	EVIDENCE	DECISION COMPLEXITY			INTERVENTION COMPLEXITY
^Note:Includes mid-level providers, dieticians, social workers for decision support and nurses and other support staff to facilitate logistics and administrative processing
RISK	EVIDENCE	LOW	MODERATE	HIGH	LOW	MODERATE	HIGH
LOW	LOW	Strong patient role	Moderate patient role	Equal role	Strong patient role	Moderate patient role	Equal role
	MOD.	Strong patient role	Strong patient role	Moderate patient role	Strong patient role	Strong patient role	Moderate patient role
	HIGH	Strong patient role	Strong patient role	Strong patient role	Strong patient role	Strong patient role	Strong patient role
MODERATE (MOD)	LOW	Moderate patient role	Equal role	Moderate physician role	Moderate physician role	Moderate physician role	Physician & Other Support ^Note
	MOD.	Strong patient role	Moderate patient role	Equal role	Strong physician role	Strong physician role	Physician & Other Support ^Note
	HIGH	Strong patient role	Strong patient role	Moderate patient role	Strong patient role	Moderate physician role	Physician & Other Support ^Note
HIGH	LOW	Strong physician role	Strong physician role	Physician & Other Support ^Note	Strong physician role	Strong physician role	Physician & Other Support ^Note
	MOD.	Moderate patient role	Equal role	Moderate physician role	Strong physician role	Strong physician role	Physician & Other Support^Note
	HIGH	Strong patient role	Moderate patient role	Equal role	Strong physician role	Strong physician role	Physician & Other Support ^Note

The following factors are likely to be important in directly or indirectly influencing the appropriate level of control chosen by the consumer: 1) level of risk (mortality, side effects, etc.);

2) strength of the research-based evidence, 3) complexity of the decision, and 4) complexity of the recommended intervention (Table 1). To some degree, these same factors will dictate features of the system required to support control. Where evidence about what to do is robust and understandable (e.g., management of hypertension, hyperlipidemia, etc), risks are low, and within the limits of common sense, we believe that as much control as possible should be shifted to the consumer (Table 1). Where the risk of confusion and of making the “wrong decision” increase (e.g., as decision complexity increases), consumer-oriented decision support tools may become increasingly important and useful. In those cases, consumer guidance can be used to simultaneously educate and facilitate thoughtful discussion and decision making.

We envision and are already testing a number of tools relevant to this process. Together, these tools should eventually provide the means for comprehensive guidance (e.g., age appropriate guide regarding near-term needs and longer-term needs for middle age, older age, and near death phases of life) and health management planning/evaluation. We specifically consider pre/post-encounter summaries and interactive preference-based guides (e.g., computerized tools that guide consumers through choices while educating them about risks, benefits, costs, etc). Encounters should be accompanied by a pre-encounter summary that verifies the encounter’s purpose and expectations as well as a post-encounter summary that includes the visit purpose, key information provided by the consumer, the agreed-upon care plan, the established follow-up goals, the clinician orders (e.g., any follow-up visits) and educational information (including references to any other sources of information). Pre/post-visit guides should be routinely voiced at an appropriate education level and in a patient-preferred form (e.g., print, email, fax, etc). Whenever possible, consumers should be able to use interactive tools that serve to both achieve pre-visit work and to guide a preference-based decision process. For example, guidance could be offered on risks, benefits, uncertainties, and costs of any decision consumers need to make. Accurate representation of facts in a preferred presentation form will be essential, as will more nuanced information about the relative certainty of those “facts”. Moreover, where an intervention is deemed necessary, comparative provider performance metrics should be readily accessible and visually intuitive. Such a decision support process can be designed to systematically guide consumers, respond to specific information they provide, obtain feedback on comprehension, and effectively communicate with a provider the consumer’s decisions and uncertainties.

The intention behind evolving consumer guidance clearly is not to eliminate clinicians, but rather to allow clinicians to be more focused on the important tasks that are less likely to be done in today’s environment (e.g., planning and overseeing comprehensive care plans, discussing difficult trade-off decisions) and to optimize the full use of the skills and training of each member of the care team (e.g., nurses, pharmacists, PAs, NPs, and other mid-levels). In the future, providers may offer the greatest value to consumers where the risks (e.g., of preventable mortality, excess utilization, adverse event, excess cost, etc.) for deciding what to do are moderate to high and where the evidence to support these decisions is low to moderate or where the provider has access to comprehensive information on what the consumer wants. We recognize that there are areas where the evidence about treatment options is unclear or inconsistent, as well as, instances for which the evidence will fail to account for clinical and/or patient-specific nuances. Moreover, certain consumers are less likely to be comfortable in this model. Finally, we do not assume that every consumer will achieve full adoption of the PHM model or that the consumer is ultimately responsible for all decision-making; rather, that well-designed PHM tools can play an important role in determining when physician or other provider involvement is crucial, when it is optional and when self-care is sufficient. We expect the relative involvement of consumers and various providers will vary substantially by different types of health problems (Table 1) as will the value proposition to payors and employers.

Objective 5: Supporting the Clinician

PHM will not be possible unless there is a compelling value proposition that fosters clinician adoption. In our view, this means that the PHM process must offer a competitive edge to clinicians by improving their efficiency, consumer satisfaction, clinical outcomes and/or profitability. We fully realize that this is a lofty goal. We consider examples of processes and tools currently under development or being considered at Geisinger that are relevant to objective 5.

Problem List Management: Physicians regularly collect, review, analyze, and synthesize patient data to diagnose, treat, and provide longitudinal care for patients. The concept of the problem-oriented medical record (POMR) grew from a need to effectively arrange and display patient medical information to enable other medical practitioners to appreciate the thoroughness of the data-gathering process and to follow the logic of the resulting diagnostic conclusions and treatment recommendations. The problem list is important to PHM because it is intimately linked to patient data capture, continuous care protocols, automated decision support, etc. In an electronic environment, the problem list offers new opportunities (e.g., when diagnoses are captured electronically, automated, tailored decision support is possible). However, coding practices and naming conventions vary widely, to the point that data in the EHR is often unreliable. The Problem List is also a shared resource in an EHR environment, introducing a new set of challenges and barriers to reliability, credibility, and utility. Based on experience within our own system, we believe that tools that encompass automated protocols for perpetual management of Problem List entries will be critical to address: a) Outdated Information (e.g., diagnoses that are old and/or no longer apply); b) Wrong Information (e.g., diagnoses that are inaccurate); c) Suboptimal Coding (e.g., diagnoses that are correct but lack appropriate specificity); d) Redundancy (i.e., multiple diagnoses that describe the same condition/problem); e) Inconsistent Terminology (i.e., naming conventions that, although perhaps standardized like the ICD schema, do not correspond well with terminology used by clinicians or consumers); f) Disorder (i.e., non-prioritized sequencing and/or illogical grouping that inhibits the user from easily/reliably visualizing and considering the full complement of actionable diagnoses); and g) Missing Information (e.g., diagnoses associated with post-encounter and/or 3^rd-party test results).

Risk Assessment and Stratification

Quantitative risk calculations provide the means for more focused and actionable feedback to consumers and clinicians and offer a means to stratify patients by risk and related care management options. We are currently testing the use of risk calculators for diabetes related macro/micro-vascular risks and for cardiovascular risk in general as a means to communicate to consumers and clinicians. These and other such tools make use of EHR and patient reported data. We will use risk assessment tools for shared decision making processes, to provide clinicians with expert clinical guidance, and sometimes as a global outcome where a number of quantifiable factors mediate the endpoint of interest (e.g., macro/micro-vascular risk in patients with diabetes).

Personalized Prediction Models

Population-level longitudinal EHR data is useful in developing prediction models that are sensitive and specific to patient subgroups that differ by history, treatment status and genetic profile. The development, validation, and real time use of such models is highly relevant to PHM. For example, we recently used EHR data to develop and validate a model for predicting diagnosis of heart failure among primary care patients. A robust model was developed that detects heart failure, on average, 15 months before it is usually diagnosed. We believe that incorporation of this risk model in to the usual care process will facilitate early detection of heart failure and provide the means to influence the natural history of the disease in patients who would otherwise be detected at a later time.

Visual Display of Clinical Information

As the management requirement of increasingly complex costly treatment regimens (e.g., biologic medications) for patients with multiple chronic diseases has multiplied, it has become increasingly difficult for clinicians to find, aggregate and confidently visualize all of the clinical information that is pertinent to a given encounter. In recent work at Geisinger, rheumatologists estimated that even with a fully-functional EHR, it would take an average of 15 minutes to fully review patient data to ensure that a treatment decision was optimal; on average, physicians have 2 to 3 minutes. The EHR makes data can be readily more readily accessible, but not necessarily in an efficient, intuitive and integrated manner. To address this gap, we are developing a web-based dashboard designed to display temporal profiles of patient reported functional status and other questionnaire-based measures, lab data relevant to toxicity, EHR data on current treatments, and other clinically relevant items. The dashboard, which will be accessed via a hyperlink within the EHR, also includes structured text fields to be completed by the nurse and physician. Draft progress notes and a patient after-visit summary will be automatically assembled through the interaction with the dashboard and imported into the EHR after the physician closes the hyperlink. Currently, the display features of the dashboard are not possible to create within the EHR. We believe this work is relevant to other medical specialties, as it provides a means to instantaneously display diverse data to facilitate efficient clinical decision making, without the imposition of EHR vendor-specific constraints on the use of and display of data.

Clinical Decision Support (CDS)

The development and promulgation of clinical guidelines, an activity which emerged in the early 1990s, has been a prolific enterprise focused on codifying disease-specific clinical knowledge intended to, in part, promote the delivery of evidence-based care. Adoption of guidelines in practice, however, has not been successful(1), especially in primary care and for the elderly (13). This is not surprising, as it is impractical to assume that clinicians can gather, winnow, and synthesize the ever-expanding and often-conflicting body of available evidence and research or that they can ever be up-to-date through CME or other modes of education.

Translating knowledge to practice (i.e., reliable efficient role-optimized operational work flows) must address a number of important realities. First, for primary care physicians, the rate of newly-generated relevant clinical guidelines far exceeds the time available to understand and assimilate them.. CME alone is not a solution. Second, guidelines are often stakeholder centric: disease-specific guidelines are generally developed by specialists, whereas primary care providers may treat the majority of cases. Translation to primary care is also hampered by the “siloed” nature of guidelines; whereas guidelines are developed for a specific condition, there are few or no guidelines that address appropriate treatments for patients with more than one condition. Furthermore, if all applicable guidelines were applied to a patient, the resulting treatment regimen would likely include multiple medications with high complexity, a risk for drug interactions, the potential for adverse drug events, and an unsustainable treatment burden (13). Third, knowledge of what to do is only one part of what is required to make the best decision on behalf of a patient. Other necessary steps – including accessing relevant patient data (a potentially time consuming process) and retrieving and interpreting the right knowledge in light of data – mediate the ultimate decision. Based on recent experience, we believe that two types of tools will be essential to bring actionable and timely knowledge to the point of care: 1) tools that identify, extract, and evaluate patient data in real time; and 2) tools that translate output from rules processes into readily interpretable actionable advice. While it would be ideal if EHRs or PHRs fulfilled these functions, our experience over the past three years indicates that this is not likely to be the case anytime soon, if at all. Ideally, clinician CDS should facilitate making optimized decisions that are based on detailed knowledge about the patient, their preferences and the best available applicable evidence

Objective 6: Shared Decision Making

Without a business sensible process, we do not expect providers to systematically engage in shared decision making (SDM) care processes with patients on a routine basis. Practically, SDM is too time consuming and, without incremental reimbursement, not cost effective. We do not envision that research on ways to foster such behavior on the part of patients or providers will be fruitful unless tools are developed to make such interactions cost effective. On the other hand, we believe that patient-completed data capture tools and interactive consumer guides provide the foundation for related tools that efficiently inform the provider of the patient’s choices and naturally foster SDM as part of the process. Patient completed questionnaires combined with risk calculators and SDM tools can be used to increase the level of productivity during an encounter without increasing the workload. We are investing in the development and use of such tools for management of CVD risk and risk of vascular events in diabetics. Our objectives in the use of these tools are: 1) educate patients about risk factors; 2) engage patients in choosing interventions to manage risk factors, including an option to do nothing; 3) understand the intervention(s) that offer the greatest benefit; and 4) seamlessly inform the provider during the encounter of the patient’s choices, or lack thereof, so that the care plan incorporates this information.

Objective 7: Continuous Management

Continuous management is logistically complex and, accordingly, subject to failure. Effective continuous management requires systematic periodic consumer-clinician interactions, data evaluation, and PHM plan changes. In practice, the first 6 PHM objectives lead to the establishment of an agreed-upon goal or set of goals for managing health problems and related risk factors. By definition, goals – which can be reflected as either subjective (e.g. quality of life) or objective (e.g. LDL) measures – need to be modifiable and actionable. The continuous management process involves a schedule of consumer-clinician interactions designed to review current status (e.g., progress relative to agreed upon goals) which, in turn, requires perpetual updating of the treatment plan. A discussion of current status can involve a consideration of satisfaction with treatment, the need to modify goals and/or modifications to the treatment plan itself. The continuous management process also involves safety monitoring and re-evaluation of risk.

Our "e-Rheumatology" project illustrates this how a continuous care process might work in the future. In one panel of a dashboard, trend lines of pain and functioning scores (based on patient self reported data) are displayed along with other relevant data (e.g., treatments prescribed, toxicity measures). In the same panel, goals for specific measures can be documented, based on patient and physician discussion. Documentation of the agreed upon goal is also automatically copied to other panels (e.g., after visit summary, draft progress notes) along with other relevant data, to ensure continuity and accuracy of communication. Lastly, the next visit is scheduled in relation to the urgency for care and/or changes in patient goals/treatment plan.

PHM IMPLEMENTATION AT GEISINGER

In this section, we describe more fully our plan for implementing PHM at Geisinger over the next five years. We are increasingly adopting a view that PHM will require that many new processes be developed external to the EHR. Hard coding processes and tools within the EHR is largely incremental and framed by a process improvement model. The disadvantage of this approach is that solutions are highly constrained. Moreover, maintaining software libraries of idiosyncratic functions becomes increasingly demanding and burdensome, especially when version changes in the EHR software occur. In this section, we focus on an alternative approach that is based on the development of tools and processes independent of the EHR. These tools and processes can be designed to interact with the EHR, an interoperable data warehouse, or other tools and processes, but the approach to development is not constrained by the need to modify the EHR itself.

The processes and tools that we have discussed in this paper vary substantially in their stage of development. A number of the tools (e.g., touch screen questionnaires, visual displays of data, interactive tools for patients, CDS) are currently being developed and tested in early pilot work, while others are still at the conceptual stage. We expect that, as we evolve PHM over time, we will identify other important elements, processes, and tools. For example, we recently concluded that a uniquely defined test environment – developed in conjunction with, but distinct from, the corporate IT environment – is essential to accelerate and diversify the discovery and development of novel IT solutions. The need for system level policies and procedures to ensure information security and compliance with regulations, while important to the process of care delivery, is a barrier to research on and innovation in the use of IT in health care. These restrictions can dramatically limit efforts to experiment with novel IT solutions. In response, Geisinger has recently developed administrative and IT solutions to ensure system level security, while providing the flexibility needed for unrestricted ability to build and test prototype applications.

Geisinger’s implementation of PHM has and will continue to involve deliberate change to existing care models, motivated by an interest to create increasingly efficient, higher quality, and more effective care. Where relevant, migrating towards information-rich virtual and semi-virtual encounters is an important part of our implementation plan. Evolving from a traditional to consumer-centric care model is a stepwise process that requires careful monitoring of what works to manage safety concerns, to facilitate adoption, and to align stakeholder interests. We characterize five stages of activity relevant to our ongoing internal implementation of PHM (Table 2). Each stage is differentiated by its primary objective, its scale, its primary outcomes, and its relevance to the business of delivering care.

The steps in our implementation plan represent a research and development-like model in which there is a progression from proof-of-concept to marketable product. Tactically, the process for any stage of work invariably involves a consideration of the business case (i.e., the potential for a meaningful ROI), management of stakeholder engagement, content creation, workflow re-design, translation of content into actionable specifications, implementation, and evaluation. Each of these steps is quite complex and common to any process redesign. Our focus in this section is on the broader stages of work, not the tactical steps common to any redesign effort.

Progression through the above stages requires that the process and/or tool meet certain requirements. Stage 1a is focused on proof of concept. Proof of concept has necessarily been a dominant activity over the past two years and will continue to be for the foreseeable future. Expectedly, many more concepts are tested in Stage 1a than succeed and progress to subsequent stages. Although nearly all new product development efforts are ultimately motivated by the potential for a meaningful return on investment (ROI), our Stage 1a and 1b development and testing efforts are not focused on either measuring or achieving a positive ROI. Rather, they are focused on rapid feedback and learning while developing a functional model for use in daily care (Stage 1a) and developing a process and/or tools that are deemed acceptable by patients and providers (Stage 1b). For both first and second stage activities, testing is usually limited to one or a few clinics and can range from a new system of care (e.g., medical home) to testing of new tools (e.g., patient data capture tool, visual display tool, clinical decision support tool) or processes. We distinguish proof of concept, which is largely focused on developing a workable model, from initial validation of proof of value (Stage 2). In our experience, there are two reasons why it is not usually sensible to simultaneously evaluate functional proof of concept and proof of value. First, simply developing and testing new processes is complex and consumes enormous resources, as the new process has to be seamlessly integrated with the existing workflows and practice settings. Second, because there are usually so many lessons learned in the first stage, the prototype version used in the second stage (i.e., proof of value) may differ substantially from first stage testing. The value proposition may also change, making the first stage value proposition irrelevant. Second stage work is worth pursuing when feedback from all stakeholders (i.e., patients, providers, clinic staff, etc) indicates that the model is workable and offers potential value.

Table 2. Stages in development and deployment of new care processes
Stage	Primary Objective	Scale	Outcomes	Business Considerations
Note on Outcomes: A key outcome across all stages is the extent to which consumer interactions are increasingly semi-virtual to virtual
1a	Proof of concept for prototype	One to two clinic patient populations	Functionality of process, tools, patient satisfaction	Theory of ROI and associated metrics
1b	Proof of value for prototype	One to two clinic patient populations	Patient acceptability and satisfaction, early quality and efficiency of care, efficiency of care, acceptability by providers	Evidence of potential for ROI
2	Proof of value with the scalable prototype	Multiple clinics populations, using quasi-experimental design	Quality of care, use of care, continuity of care, efficiency of care	Documented ROI
3	System level (single focus) implementation to increase customer growth and competitive advantage	System level implementation within a domain, IT ownership of processes, and re-organization of clinic space	Seamless integration, documentation of increased productivity for the same provider effort, improvement in patient outcomes and reduced need for care. Reduced cost to consumer per unit gain in outcome. Reliable use and process performance across diverse settings	Sustainable increase in net revenues through greater RVUs/time and documentation of work done Sustainable PMPM savings
4	Performance improvement and diversification of processes/tools to other clinical domains and settings	System level	Patient data capture, CDS, education and care planning/oversight happen as part of routine care	Continued increase in net revenues, decrease in PMPM. Collaborative work with payors
5	Export capabilities to other systems and markets	Market Level	External market purchase of consulting, processes, tools, etc	New ventures, partnerships or licensing activities

Stages 3 and 4 represent a fundamental shift in emphasis and responsibility. Initiating Stage 3 work only occurs when executive leadership is convinced that proof of value is established, that safety issues have been fully addressed, and that translation of the prototype to a system level operation is feasible and desirable. This critical step represents an important transition to relying on external funding (e.g., research) and modest amounts of internal funding to a more significant system level investment in infrastructure. Moreover, executive support also means an important shift in IT emphasis and control from research and innovations testing in a secure environment to control by the system level IT department for broader implementation, management and ongoing support. Finally, we note that stage 5 represents a very different type of shift in emphasis and control from one that is internally focused to one that explores commercialization opportunities in the commercial market. Stage 5 work involves the collaboration among venture capital experts, system level IT leaders, and other experts to develop a more generalized solution for diverse environments and to evaluate market opportunities.

In our recent experience, activity relevant to any one stage creates feedback loops to other stages. Expectedly, the accumulation of lessons learned creates a valuable organizational asset. Moreover, tools developed for different projects are combined to create new processes. Over time, we expect an acceleration of new developments that leverage earlier stage 1 and 2 work. Based on work over the past two years, it is clear that experience in developing and testing new processes and tools in one setting, lead to accelerated application of proven prototypes in other areas, reducing the need for Stage 1 testing and possibly Stage 2 testing of a similar process or tool in a different clinical area. We consider two specific examples of this type of evolution in research and development.

We are creating a second generation semi-automated cardiovascular disease (CVD) risk management system for primary care that identifies data needs, quantifies patient risk, engages patients at elevated risk in a decision process, and informs the clinician of consumer choice and expert advice (automated decision support). This system is actually the product of work over the past three years focused on testing patient data capture tools (i.e., Stage 1a) using various technologies in a variety of settings, a cardiovascular risk calculator that will soon be used in primary care, and an early version of an interactive decision support/preference elicitation tool for consumers. This work on CVD risk management has provided the foundation for parallel developments of a diabetes disease management process for primary care. This represents an example of leveraging previous work to bypass one stage (i.e., 1a) of work, moving directly to the subsequent stage (i.e., 1b). The diabetes project leverages many of the developments for the CVD project and, in our view, represents what is likely to unfold in a number of areas.
Development of a rheumatology practice system that captures outcomes data from patients during each visit, provides an integrated and intuitive on-demand display of all patient data (i.e., labs, clinical, functioning, pain, etc), and automatically drafts structured progress notes and a patient after-visit summary. Again, development of this process is based on several years of work, borrowing from experience in the development of other tools. The visual display tool is being developed to function in a web environment but to also interact with our EHR. We expect to apply this same type of process (i.e., patient data capture and interactive visual display tool) to other specialty care areas where information needs are complex (e.g., oncology), and time is too limited to carefully review all data during a given encounter. Moreover, development of tools in a web environment is deliberate, as it provides the foundation for possible use in other settings outside of Geisinger whether or not an EHR is available.

Space does not allow us to fully describe our expectations over the next five calendar years (2009 – 2013) for all clinical care settings (i.e., primary care, specialty care, surgery, inpatient care, home care, end of life care). Instead, Table 3 offers a specific profile of how each PHM objective will be addressed over the next five years in primary care. This profile represents our limited view based on recent experience and will necessarily evolve with growth in experience and lessons learned. While representative, the table itself is incomplete, as not all developments currently underway are represented and others are likely to be added. As previously noted, PHM related work has been under way at Geisinger over the past three years. Specifically, we have been engaged in the ongoing development and implementation of an enterprise-wide data warehouse. While this warehouse is not currently interoperable for real time use, it provides the foundation for future implementation of this type of process. Future work will involve incorporation of existing databases that are not currently accessible in real time and, more generally, the incorporation of other types of “data” resources, including questionnaire protocols, highly tailored clinical rules and related text content, libraries of community resources that are relevant to supporting consumers where they reside. Virtual and semi-virtual encounters will depend on the evolution of access channels in traditional settings as well as at home or under other conditions (e.g., from work). We have gained extensive experience in developing patient data capture tools and developing workflows to support their use in practice. Over the next five years, the creation and implementation of PARDAMS will dominate our work in this arena and foster parallel evolution in implementing protocols for consumer choice processes, provider CDS, and shared decision making. Operationally, there is a strong interdependence among the processes relevant to these four objectives.

We note that the processes and tools that we develop for future application will have to be nuanced to the features of the population of interest. Tools will differ depending on the clinical setting. We expect that sophisticated interactive visual dashboards will be more important in specialty care than in primary care. On the other hand, automated real time actionable CDS that is highly nuanced to the encounter and a given consumer will be fundamentally important to PHM in primary care. The focus of interest in development of tools and processes will, of course, also differ by the health care setting, as well by other broadly defined categories of health problems (e.g., pediatric ENT, end-of-life care) and diseases (e.g., oncology) and conditions. For inpatient care, we expect a strong focus on creating reliable, consistent, and integrated means of managing care transitions. In primary care, the process for chronic episodic conditions (e.g., migraine, depression, asthma, low back pain, GERD, bladder control, etc) will involve integration of sophisticated questionnaires, consumer preference menus, CDS, and periodic asynchronous evaluation of outcomes status. As described above, the process and tools differs to some degree for managing chronic progressive disorders. With experience, we expect that a market focus will play an increasingly important role in identifying new opportunities that offer a meaningful return on investment in new tools and processes.

CONCLUSIONS

In 2007, healthcare expenditures in the US were approximately 2.3 trillion dollars (2). Despite the enormity of this outlay, the collective return on our national investment is unclear: costs continue to rise, millions remain uninsured, and the quality of care remains suboptimal. Our vision of a healthcare system that places the patient – acting as a consumer, but supported by intelligent tools and a receptive care delivery system – at the center of the care process is the primary requirement of this necessary cultural shift to Personal Health Management.

One can reasonably argue that our vision for PHM fails to consider the complexities of the healthcare system. However, consistent with the eventual disproof that the decision making required for non-healthcare interactions is too complex to be enabled by interactive computer-mediated encounters (e.g., social match-making, financial portfolio management, travel scheduling, etc.), we predict that the PHM model is inevitable.

A primary potential benefit of the PHM approach is that as more consumers become more engaged, there are a large number of health-enhancing behaviors that are likely to be adopted that would never have a chance in a passive “what comes, will come” attitude and approach that primarily exists today. Such innovation is sorely needed as we seek to enable individual consumers while simultaneously having a positive sustainable population-level impact on the chronic disease epidemic afflicting our nation.

Table 3. Implementation expectations at Geisinger over the next five years for Primary Care Practices
PHM OBJECTIVE	PHM Stage (St) of implementation and focus of work by care setting and calendar year
	2009		2010		2011		2012		2013
	St	Focus of Work	St	Focus of Work	St	Focus of Work	St	Focus of Work	St	Focus of Work
Interoperable Data Warehouse	3	Phase I implementation of data critical to day to day operations	3	Addition of non-EHR clinical data captured in other devices	3 & 4	Addition of questionnaires, clinical rules, and CDS databases for most common conditions	3 & 4	Expansion of questionnaires, clinical rules, and CDS databases for other conditions and addition of GIS based community resources	3 & 4	Expansion of questionnaires, clinical rules, and CDS databases for other conditions
Access Channels	3 & 4	MyGeisinger web portal for record review, Rx orders, scheduling, communication, etc. with provider	1A	Home based encounters with patient using an online interactive tool and phone consultation. Expand use of computer touch screens in clinics	1B	Home based encounters with patient using an online interactive tool and phone consultation. Re-organize waiting areas to patient working areas	2	Home based encounters with patient using an online interactive tool and synchronous or asynchronous interactions with provider	3	Home based encounters with patient using an online interactive tool and synchronous or asynchronous interactions with provider
Consumer Data	1B	Test proof of value of questionnaires for selected common conditions and utilities (e.g., medication reconciliation)	2	Develop a scalable model with expansion of types and uses of questionnaires for in-clinic patient data capture. Continue ROI evaluation	3	System level implementation of patient data capture processor for in-clinic data capture and remote data capture	4	Expand library of questionnaires and access channels	4 & 5	Expand library of questionnaires and access channels. Commercialize questionnaire product
Consumer Choice	1A	Continue pilot testing and re-engineering preference based care tools for common conditions	1B	Evaluate proof of value of preference based care tool for patient outcomes and clinical ROI.	2	Develop, deploy, and evaluate scalable version of preference based care tool	3	Deploy system level tool for routine use of preference based care tool	4 & 5	Expand libraries of tool interfaces for multiple conditions and expand use to external market
Provider Support	1A	CDS modules to provide real time expert guidance at the point of care for CVD, diabetes, headache	1B	CDS modules to provide real time expert guidance at the point of care for CVD, diabetes, headache. Integrate visual display tool with CDS	2	Testing of prototype system level tool that manages libraries of expert knowledge for multiple conditions and related visual displays. Test automated problem list manager.	3	Implement system level tool that manages libraries of expert knowledge for multiple conditions per prototype in 2011	4	Continued expansion of libraries of expert knowledge for multiple conditions
Shared Decision Making	1A	Proof of principles for linking alert in EHR to web application that displays patient preferences and decisions	1B	Proof of value for prototype model that links an alert in EHR to web applications that displays patient preferences and decisions	2	Integration of consumer data, consumer choice, and provider support to improve SDM process	3	System level integration of consumer data, consumer choice, and provider support to improve SDM process	4 & 5	System level expansion of conditions covered through integration of consumer data, consumer choice, and provider support to improve SDM process. Evaluate market opportunity.
Continuous Management					2	Test prototype models for semi-automated and automated continuous care management in several clinical areas	3	System level version of model for semi-automated and automated continuous care management for use in several clinical areas	4	System level version of model for semi-automated and automated continuous care management for use in an expanded number of clinical areas

References

McGlynn EA, Asch SM, Adams J, Keesey J, Hicks J, DeCristofaro A, et al. The quality of health care delivered to adults in the united states. N Engl J Med. 2003 Jun 26;348(26):2635-45.
National health expenditure projections 2007-2017 [homepage on the Internet]. Centers for Medicare & Medicaid Services. Available from: http://www.cms.hhs.gov/NationalHealthExpendData/Downloads/proj2007.pdf.
Social Security and Medicare Boards of Trustees. A SUMMARY OF THE 2008 ANNUAL REPORTS. 2008.
Call KT, Davern M, Blewett LA. Estimates of health insurance coverage: Comparing state surveys with the current population survey. Health Aff (Millwood). 2007 Jan-Feb;26(1):269-78.
Robinson JC. Slouching toward value-based health care. Health Aff (Millwood). 2008 Jan-Feb;27(1):11-2.
Robinson JC. Reinvention of health insurance in the consumer era. JAMA. 2004 Apr 21;291(15):1880-6.
Buntin MB, Damberg C, Haviland A, Kapur K, Lurie N, McDevitt R, et al. Consumer-directed health care: Early evidence about effects on cost and quality. Health Aff (Millwood). 2006 Nov-Dec;25(6):w516-30.
DiCenzo J, Fronstin P. Lessons from the evolution of 401(k) retirement plans for increased consumerism in health care: An application of behavioral research. EBRI Issue Brief. 2008 Aug;(320)(320):1, 3-26.
Institute of M, Committee on Improving the Patient,Record, Steen EB, Dick RS. The computer-based patient record an essential technology for health care. Washington, D.C: National Academy Press; 1991.
Institute of M, Committee on Quality of Health Care in,America. Crossing the quality chasm: A new health system for the 21st century. Washington, D.C: National Academy Press; 2001.
Fitzmaurice DA, Murray ET, McCahon D, Holder R, Raftery JP, Hussain S, et al. Self management of oral anticoagulation: Randomised trial. BMJ. 2005 Nov 5;331(7524):1057.
Heneghan C, Alonso-Coello P, Garcia-Alamino JM, Perera R, Meats E, Glasziou P. Self-monitoring of oral anticoagulation: A systematic review and meta-analysis. Lancet. 2006 Feb 4;367(9508):404-11.
Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and quality of care for older patients with multiple comorbid diseases: Implications for pay for performance. JAMA. 2005 Aug 10;294(6):716-24.

Return to Table of Contents

Role of Professional and Medical
Specialty Societies in the
Era of Personalized Health Practice

Eric C. Faulker, MPH
Senior Director, RTI Health Solutions
Director, Genomics Biotech Institute, National Association of Managed Care Physicians

William L. Roper, MD, MPH
Dean, School of Medicine
Vice Chancellor for Medical Affairs
University of North Carolina at Chapel Hill

US Cost and Quality Challenges: The Context for Personalized Health Care

Similar to the precarious positioning of Odysseus’ ship between the dangers of Scylla and Charydbis,[1] it is well documented that health care is currently caught between two pressures: demand for improved health outcomes and quality and constraints on the spending required to deliver on these improvements.[2],[3],[4] Successful cost reduction strategies may sacrifice quality. Alternatively, strategies to increase quality may result in excessive system costs. Striking the appropriate balance between these two forces is likely be the best means to overcome the challenges facing our United States (US) health system.

Growth in US health spending hovers at 7% per year, double-digit increases in annual insurance premiums have occurred in many recent years, and overall health expenditures are anticipated to double in under a decade.[5],[6] These trends are viewed by many health policy and decision makers as unsustainable long term.[7],[8] Such trends in health spending have increasingly forced payers, providers, employers, and patients to explore less costly means for achieving high-quality care and maintaining sustainable health care delivery models.

A variety of factors influence the increase in health care spending in the US. Key factors include but are not limited to the aging population, increasing prevalence of chronic diseases, inappropriate use of health services and stakeholder incentives, inefficient health delivery systems, focus on treating sickness versus promoting wellness, and the expansion of emerging health technologies. These factors are complex, intimately intertwined and when taken in the context of our highly fragmented health delivery system, make health reform in the US a daunting prospect.

The US Congressional Budget Office (CBO), Institute of Medicine (IOM) and other influential groups have cited the availability of innovative and breakthrough health care technologies and lack of clarity surrounding their use as a significant driver of increased health care spending.[9],[10] On the other hand, new health technologies can often enable sophisticated, effective, and increasingly personalized care.[11],[12] Some technologies, such as molecular diagnostics and targeted treatments promise to alter paradigms of care and optimize individual health outcomes, pending real world validation in clinical practice.[13] Balancing affordability of new health technologies against innovation and opportunities for care enhancement will prove challenging in an era of tightening budgets and increased health services utilization constraints.

Just as new health technologies may improve health quality and distribution of costs, evolving health information technology (HIT) and decision support systems[14] promise efficiencies in health services delivery. However, adoption of information systems in health care has lagged behind other industries due to lack of information standards, limited systems interoperability or information sharing, insufficient financial incentives and other factors.[15] For example, the retail industry (e.g., large department store chains such as Walmart or Target), financial services industry (e.g., banking and investments), and shipping industry (e.g., Federal Express and United Parcel Service) gain significant operational efficiencies via reliance on highly sophisticated information systems. Likewise, clinical decision support systems must evolve considerably to offer solutions relevant to day-to-day decision needs of care providers. In the near future, mandatory and voluntary government and commercial initiatives focused on clinical best practices, quality improvement, and health care transparency will increase the desire for application of HIT to improve provider operations and compliance with expanding data collection and reporting requirements.[16]

Personalizing and better targeting care practices based upon individual needs and health indicators is one possible solution to help defray growing cost and quality challenges. On September 19, 2007, the US Department of Health and Human Services (HHS) Secretary Michael O. Leavitt unveiled the forward looking report entitled Personalized Health Care: Opportunities, Pathways, Resources.[17] This “early reconnoitering” report lays a conceptual roadmap for harnessing our rapidly expanding biomedical knowledge base and health information systems to enable increasingly personalized health care.

The concept of personalized health care is very broad, integrating our growing knowledge of genetics and biomarkers and their role in treatment selection with HIT, principles of evidence-based practice, and health quality and performance improvement approaches. In essence, personalized health care would combine the best available information from a variety of sources in an actionable manner so that physicians and patients can make appropriate health care decisions and enhance use of individual patient data in health practice.[18] Despite the intuitive appeal of personalized health care, successful implementation will not occur effectively without deliberative forethought and the collaborative engagement of many health stakeholders.

The focus of this paper is to explore what would need to happen for medical and scientific specialty societies (professional societies) to assume a lead role in enabling delivery of personalized health care that leverages knowledge of individual variability.[19] Understanding drivers and incentives of relevant stakeholders will also be important to comprehensive strategies for personalized health care (see Appendix A). More broadly, this paper considers issues relevant to the collaboration among key stakeholders around personalized health care issues, as well as business and operational implications of these factors for professional societies.

Balancing Standardized vs. Personalized Health Care

In some respects, personalized health care approaches may run contrary to or even disrupt health care paradigms centered on standardizing clinical practice and delivery. Further, existing policies, business incentives, and decision drivers, such as those outlined below, may hinder implementation of personalized health care solutions in the US. The following section highlights some basic challenges of balancing standardized versus personalized health care; this overview provides a foundation for discussion of how professional societies may play a leadership role in implementing personalized health practices.

Considerations for evidence-based practice and policy: Evidence-based medicine (EBM) emphasizes using the best available clinical (and often economic) evidence to inform the appropriate use of new health technologies. Evidence-based approaches attempt to link risks and benefits of the intervention to patient outcomes and quality-of-life improvements using a process called health technology assessment (HTA).[20] HTA synthesizes evidence from existing studies to characterize the value of a diagnostic or treatment in the context of clinical practice.

As knowledge linking individual patient information to disease risk and treatment outcomes grows in the coming years, evidence-based policies and guidelines geared largely toward standardizing care approaches for broader patient populations will have to adapt in tandem. Likewise, approaches to evidence assessment (e.g., HTA practices, data modeling) and decision support will also become increasingly necessary to address an expanding evidence-base that incorporates personalized health care information. While there is no doubt that personalized health care will be beneficial in many ways, we are only beginning to realize the implications that knowledge of individual variability confers.

Personalization of health care will vary depending upon the potential for individualization of treatment. For example, our ability to leverage biomarker-related patient management may differ by health condition (e.g., allergic rhinitis, diabetes, psychological conditions and cancer) and other factors (e.g., correlation of genotype to phenotype, ability to intervene based on biomarker information, involvement of single versus multiple genes). While evidence-based practice has always taken into account individual variability and subpopulation effects, it is clear that the magnitude and frequency of new clinically relevant information based on genomics-related knowledge will challenge our existing processes for translating knowledge into practice.

Movement towards personalized health care raises several broader considerations for evidence-based practice and policy, including but not limited to the following.[21],[22],[23]

Under which scenarios are evidence-based standardized approaches or personalized approaches most beneficial, practical, and cost-effective? What are our limitations for personalizing health care?
How does expansion of personalized health information influence prioritization of topics for HTA given limited funding for such endeavors? Should additional focus in this area be facilitated and in what ways?
How should HTA processes change in regard to the timing of evaluations, definitions of ideal or acceptable evidence (e.g., evidence from models or longitudinal databases versus randomized controlled trials), and resultant recommendations for use?
How will personalized health care models influence sponsor evidence requirements for securing market clearance and reimbursement? Will this diminish or enhance incentives for innovation and availability of new health technologies?
What are the implications of increasingly complex treatment scenarios and decision steps for clinical guideline development and maintenance?
How will personalized health information be communicated in a manner useful to policy makers, payers, providers, and patients?
How and in what ways will complex individual health information be practically implemented into existing patient management and health decision frameworks?
Which stakeholders will pay for the complex and costly data collection and analysis that may be applicable to personalized health practice? Who will distinguish “got to have” vs. “nice to know” evidence requirements?
How much emphasis on personalization in health practice is too much (i.e., where the costs of personalized treatment outweigh the value of standardization and population-based approaches)?

Inefficient integration of evidence-based medicine and personalized health care practices may:

Inappropriately preclude access to beneficial health technologies
Create overburdensome evidence development requirements
Involve unnecessary health data collection and reporting requirements, and
Challenge existing health care business models.[24],[25]

On the other hand, appropriate personalized health policies and practices offer the potential to avoid adverse health outcomes associated with unclear treatment scenarios, increase the precision of care management, and redirect expenditures towards health quality and efficiency gains.[26] Addressing relevant questions and sketching viable knowledge translation and health delivery frameworks will be essential to fully realize the promise of personalized health care and adapt existing approaches, as appropriate, to balance standardized and personalized care.

Considerations for health quality and performance management: Health quality measures, just like clinical practice guidelines, are evidence-based and focus on characterizing standards of clinical practice and patient care. Such quality measures are often very specific (e.g., measurement of a diabetic patient’s HbA1c every 6 months) with years of evidence that support the measure as a standard of care. At baseline, health quality measures must be measurable and used in instances where the desired change in health delivery is achievable .

Health quality measures are increasingly used in evolving pay-for-performance (P4P) programs. Quality-measure-driven P4P programs provide financial incentives to hospitals or individual physicians for providing health care services as defined under performance management contracts and often involve “dashboards” of quality measures that characterize various aspects of provider performance. In fact, quality measures are increasingly based on clinical practice guidelines, providing additional incentives for hospitals and physicians to follow established standard care practices and extending the influence of evidence-based practice in all aspects of care.

The Centers for Medicare and Medicaid Services (CMS), commercial managed care organizations (MCO), and large employers/employer coalitions have experimented extensively with quality management and P4P approaches for almost a decade. Although results in the US health system have been mixed and best practices are still evolving, quality improvement and P4P programs are here to stay.[27],[28],[29] In general, these programs are intended to support good clinical and operational approaches and increase the consistent provision of accepted practices in areas with notable and actionable inefficiencies.[30] In the near future, quality and information reporting requirements under CMS and other programs such as the Reporting Hospital Quality Data for Annual Payment Update (RHQDAPU) and the Physician Quality Reporting Initiative (PQRI) will significantly expand collection of quality-related data and accelerate evolution of P4P.[31]

At present, the implications of personalized health care for quality measurement and P4P paradigms is uncertain and has not been well studied. As these programs evolve from broader foci (e.g., inpatient quality, outpatient quality) towards practice- or disease-specific quality measures, there is greater potential for emphasis on individual variability. However, current clinical practice guidelines often lack the specificity for development of quality measures or decision support systems (even without introduction of individual variability). On one hand, system incentives should not be structured in a manner that diminishes individualized health approaches, targeted technology applications, and innovation. On the other hand, integration of personalized health practice should not occur in a manner that rejects the value of standardization and creates unnecessary administrative and financial burdens for health stakeholders.

Considerations for business and operational efficiency: In the short term, as the tools for personalized health care evolve, processes for integration will be subject to uncertainty as we gain familiarity and confidence in applying individual health information. Health decision makers must also weigh applicability of current policies and practices where decisions may be made on a variable scenario-by-scenario basis.[32] Business integration challenges will center on both the cost and efficiency of care as health delivery continues to shift towards preventive and individualized care.

The expansion of biologics and targeted therapies is illustrative of the opportunities and challenges associated with transition to personalized health practice. In 2005 there were approximately 350 biologics in phase III trials or undergoing FDA review, and over 2,000 others are in early development.[33] A recent study of Blue Cross Blue Shield plans reported that spending on specialty pharma products has risen almost 35% between 2002 and 2003, and these products are estimated to represent 25% of all outpatient pharmacy spending by 2008.[34],[35],[36] As the cost to patients of some specialty pharma products approaches or exceeds $10,000 per month, overall affordability and access are key considerations for US health care, despite the potential value of such treatments.[37] While some of these products may markedly improve mortality and quality of life through targeted treatment, others will only offer marginal benefits—because of these scenarios, value assessment is important for informed treatment utilization and sustained access.

Likewise, emerging molecular diagnostics (e.g., gene/protein expression or array-based diagnostics, multi-biomarker panels, gene sequencing tests) show great promise for increasing the effectiveness and individualization of care. Diagnostics may also enable avoidance of certain downstream costs and patient adverse events/complications by informing more sophisticated early decision making and intervention/treatment strategies. Despite this promise, US MCOs have voiced concern because some new molecular diagnostics are priced significantly higher than predecessor diagnostics (generally priced in the 10s to 100s of dollars), with price ranges approaching $4,000 per test.[38],[39] As a result and with a host of similar and costly tests in development, payers and federally supported efforts such as Evaluation of Genomic Applications in Practice and Prevention (EGAPP) are moving to develop HTA processes, decision criteria, and management practices that consider the unique attributes of diagnostics and implications for personalized health practice.[40],[41],[42]

New health technologies have catalyzed a host of cost and utilization management strategies relevant to payers and providers, including but not limited to the approaches presented in Table 1. It will be important to consider “fit” of personalized health practices within this complex system of cost controls and the policies guiding their use.

Table 1. Cost Containment Strategies for New Health Technologies
Coverage	Utilization	Health Plan Rules	Cost Sharing
Adapted from: Hoadley J. Cost containment strategies for prescription drugs. Kaiser Family Foundation report, March 2006 and Cost-containment Strategies for Prescription Drugs. Kaiser Family Foundation. 2005; Morrow T. Biopharmaceuticals: the view of the future BIO 2005 breakout session.
Benefit exclusions Denials not subject to appeal Noncoverage because the product is deemed investigational or experimental Conditional coverage Value-based purchasing Data mining to refine coverage Exclusion of certain drugs/drug classes from coverage	Specialty drug limits Disease limits/disease management models Coverage for label use only Retrospective utilization review Health care transparency and data reporting Provider profiling Dispensing limits	Step therapy Prior authorization Mandatory generic substitution Pay for performance Closed formulary	Coinsurance Copay models New formulary tier strategies Deductibles and out of pocket maximum payments Lifetime maximum payments Reference pricing Risk sharing and rebates

Increasing health data collection and reporting requirements of public and commercial payers also place financial and operational pressures on hospitals, physician practices, and other providers.[43],[44],[45] These efforts, intended to improve health quality and cost control, include pilot and other programs covering multiple care settings and include an increasing array of information related to quality of care (e.g., longitudinal health databases and patient registries), provider performance, and use of health services and technologies. While national provider adoption of electronic health records, data capture systems, and information technology remains low (~23%-27%), broader adoption would be a necessary prerequisite for personalized health care.[46]

At present, while data on individual patient variability (e.g., diagnostic test information) is included in some data reporting initiatives, there is not yet a comprehensive data reporting approach centered on personalized medicine or personalized health care. Recently introduced bills (S.3822, the Genomics and Personalized Medicine Act of 2006 and H.R.1321, the Medicare Advanced Laboratory Diagnostics Act of 2007) would incorporate data reporting elements, but are unclear regarding incentives for providers, diagnostics manufacturers, reference laboratories, and other stakeholders that would defray the costs associated with data reporting, database maintenance, and data access or analysis.

Integration of HIT certainly holds great potential for improving health care efficiency, quality, and cost, but must also be balanced against the practical business and operational impacts on key health stakeholders. Integration of HIT is one important component of the ability to offer personalized health care, but well-developed provider organizations driven by appropriate system incentives and underpinned by organizational supports and systems also will be necessary to support personalized health care across the diverse array of provider organizations in the US.

The aforementioned considerations are a modest sample of the policy, business, and translational issues associated with integration of personalized health practices with established and standardized health delivery models. They are, however, illustrative of the complex challenges facing those involved in health care reform and in efforts supporting the transition to personalized health care. The remainder of this paper will consider elements necessary for professional societies to play a leadership role in evolution of personalized health practice, given the pressures and implementation issues presented.

How Professional Societies Can Contribute to Elements Important to Personalized Health Care

There are hundreds of scientific and clinical specialty and other societies (professional societies) in the US that may be relevant to advancement of personalized health care. Each professional society has its own mission and vision, unique focus, and range of service offerings that are relevant to members and external stakeholders. Professional societies offer great value to the health care field by serving a myriad of functions, including but not limited to continuing medical education; development of clinical practice guidelines; informing health policies and practice standards; refining research standards, decision tools and business practices; and serving as a venue for health stakeholder collaboration and communication.

For purposes of this paper, societies relevant to the support and advancement of personalized health care would fall into the general categories identified in Table 2.

Table 2: Types of Professional and Medical Specialty Societies Relevant to Personalized Health Care
Category	Example Organizations
Medical specialty societies	American College of Cardiology (ACC) American College of Rheumatology (ACR) American Psychiatric Association (APA) American Society of Clinical Oncology (ASCO)
Medical professional and health management societies	Academy of Managed Care Pharmacy (AMCP) America’s Health Insurance Plans (AHIP) American Medical Association (AMA) National Association of Managed Care Physicians (NAMCP)
Medical organization associations	National Comprehensive Cancer Network (NCCN) Association of American Medical Colleges (AAMC)
Disease-focused associations	American Diabetes Association (ADA) American College of Medical Genetics (ACMG) American Heart Association (AHA)
Scientific and clinical professional associations	American Association of Clinical Chemistry (AACC) American Association for the Advancement of Science (AAAS)
Life sciences industry associations	Advanced Medical Technology Association (AdvaMed) American Clinical Laboratory Association (ACLA) Biotechnology Industry Organization (BIO) Pharmaceutical Research and Manufacturers of America (PhRMA)
Health care quality and efficiency organizations	American Health Information Management Association (AHIMA) National Committee for Quality Assurance (NCQA) National Quality Forum (NQF)
Special interest consortia	Genetic Alliance National Patient Advocate Foundation (NPAF) Personalized Medicine Coalition (PMC)
Health research professional organizations	Academy Health Institute of Medicine (IOM) International Society of Pharmacoeconomics and Outcomes Research (ISPOR) Health Technology Assessment International (HTAi)

While the majority of professional societies are nonprofit organizations, it is important to note that these organizations are service-oriented businesses. They certainly help to advance science and health practice, but also protect the interest of their members. As such, professional society success and viability depend on development of offerings valuable to member’s education, decision making, and business operations. Development of niche-oriented or unique areas of emphasis that are sustainable in relation to competing offerings of other professional offerings and other stakeholders (e.g., conference coordinating businesses, government and commercial agencies) is also common.

As with any other business, professional societies also have financial, staff, and other limitations that influence the scope and nature of the activities that they can feasibly engage in. In other words, no single professional society has the capacity to “be all things to all people” and play a role in all activities relevant to personalized health care. Despite the potential interest in or importance of any particular topic or effort, society activities must be identified and prioritized based on alignment with mission, key objectives, capabilities, and member needs.

Understanding the specific areas of focus of a relevant professional society is important to identifying the role that it may play in advancing health services or policy changes. Is the organization message-oriented (e.g., in the context of policy making), content-oriented (e.g., development of clinical practice guidelines and standards), or both? Key activities relevant to personalized health care that professional societies currently engage in include, but are not limited to the following activities. Many organizations will engage in more than one, but not all of these activities.

1. Clinician education, training, and certification:

Professional societies have historically played a fundamental role in offering services that support continuing medical education. Education and training can be delivered in a variety of forms including a peer-reviewed journal operated by the society, newsletters, Webinars, and white papers on key clinical topics. Professional conferences and workshops are also important venues for learning about the latest trends in technology, health services delivery and management, and their implications for future business practices.

Many societies also offer formal training and certification necessary to maintain current licensure for clinical professionals. Others, such as the American College of Cardiology and American Society of Clinical Oncology develop and offer interactive educational portals and resources for physicians that cover a range of topics aimed at keeping practicing clinicians current on clinical, business, and policy issues important to their practice.

2. Clinician decision support and information management:

Within the current “age of information overload,” the efficiency of accessing and managing information key to clinical decision making is important to appropriate health care decision making. Although existing systems may range from broad information interfaces to very specific applications such as drug dosing (e.g., safe dosing the anticoagulant warfarin for prevention of thrombosis and embolism) or targeted therapy selection, most decision support systems are not generally maintained by professional societies.[47] Expansion of complex molecular diagnostics and need to improve treatment use and outcomes will escalate the need for such systems as adjuncts to standard clinical practice in the near future.[48] As these systems continue to develop, professional societies will contribute to ensure appropriate alignment and currency for clinical practice.[49]

As use of genetic and biomarker-based information is increasingly implemented in clinical practice, physician decision support systems that incorporate evidence-based practices and decision steps will need to expand or evolve to accommodate information on individuality.[50] It will also be important to understand how and to what extent health care processes and decision tools that emphasize standardization (e.g., clinical practice guidelines, clinical pathways, quality programs) should incorporate individual patient information to balance potential improved outcomes of individualized care with the costs of this approach.

3. Patient education and decision support:

Some professional societies emphasize patient education and informed clinical decision making. This mission includes providing basic education on disease pathology and outcomes, diagnosis, and treatment alternatives, implications for patient subpopulations, and other resources helpful to the patient. The American Diabetes Association (ADA), American Heart Association (AHA), American Cancer Society (ACS), and many others have highly diversified offerings targeted to the individual patient. Others, such as the National Patient Advocate Foundation (NPAF), collaborate with a variety of health stakeholders to ensure that the patient’s perspective is appropriately included and represented in clinical practice change and health care reform initiatives.

Given the rapid expansion of available diagnostics and treatments, including those related to gene-based and personalized medicine, and the resulting maze of complex choices, patient-oriented information for informed decision making is important. Further, as cost shifting places greater responsibility for health services payment for on individual patients, evidence characterizing the benefits, risks, and value of health services is essential to informed decision making. While basic patient decision support systems have evolved, significant room is available for additional support from professional societies and other health stakeholders, particularly as interoperable health information technologies mature.

4. Health outcomes research and new health technology evaluation:

One of the most important roles that professional societies play is in education, communication, and evaluation of new medical evidence supporting diagnosis, treatment, and health services delivery. This role includes not only vetting the findings of clinical studies, but also providing input on methodologies for study design, good clinical and laboratory practices, and evidence review. In general, such contributions include but are not limited to the following areas:

Clinical study design and implementation support: Most medical specialty societies do not contribute to clinical study design and implementation support to the same extent as life science manufacturers and government and academic researchers. However, professional society membership does enable members from various stakeholders groups to seek and obtain input from peers on these issues, including methods for incorporation of genes or biomarkers into clinical studies.

Other societies, such as the Institute for Pharmacogenomics and Outcomes Research (ISPOR) and Health Technology Assessment International (HTAi) emphasize development of good methodological practices for developing and evaluating evidence characterizing the value of new health technologies and practices. Organizations with a methodological focus may play a significant role in developing criteria, clinical and data modeling research approaches that help fill existing gaps in study approaches and evaluations for diagnostics, drug-diagnostics combinations, and database and registry evaluation that are relevant to personalized health care.
Horizon scanning and HTA: Evaluation of emerging health technologies is also an area where medical professional societies provide valuable input relevant to health practice. The conferences hosted by professional societies are one of the first places, prior to publication in peer-reviewed clinical and scientific journals, where ongoing and novel research findings are presented to clinical peers and other stakeholders for consideration. Such conferences are also where new health technologies may be introduced to the broader clinical community, and adoption, uptake, and diffusion considerations may be discussed. This exposure to health interventions enables payers, policy makers, and other health stakeholders to scan the horizon for potential clinical, financial, and other impacts.[51]

While the process of clinical practice guideline development involves systematic assessment of evidence, many of these efforts are broad and do not focus on specific health technologies. Payers and external HTA bodies are most often the “first line” of health-related organizations to evaluate the clinical (and sometimes economic) value of individual new health technologies and considerations for appropriate use. However, the HTA process often involves clinical experts or key opinion leaders that are members of professional societies. Additionally, health technology evaluators may consult with medical specialty societies during the HTA process. Given the complexity of some emerging diagnostics and treatments (e.g., antibody-based biologics, cellular, and gene therapies), increased involvement of professional societies, manufacturers, and other stakeholders may be necessary to ensure informed technology adoption and use decisions.
Clinical practice guidelines and practice standards: Professional societies play a pivotal role in developing and maintaining clinical practice guidelines and informing practice standards or providing clinical pathways for evidence-based care provision. The National Guideline Clearinghouse, a comprehensive database of clinical practice guidelines maintained by the Agency for Healthcare Research and Quality (AHRQ), lists over 2,200 guidelines, the majority of which were developed and are updated by US and international medical specialty societies. Some of these guidelines are specific to a particular treatment indication and/or scenario. Other guidelines, such as those published by the American College of Rheumatology, National Comprehensive Cancer Network, and American Heart Association are comprehensive resources covering multiple indications or disease areas.

Development of clinical practice guidelines is often a significant and costly undertaking for professional societies. The process often involves systematic review and evaluation of the best available evidence, as well as multiple rounds of review and revision by members, appointed clinical advisory boards, and external stakeholders. Incorporating relevant and actionable information on individual variation into clinical guidelines will introduce greater complexity and decision steps versus current standardized methods of practice. Ultimately these guidelines may be adapted into health quality and performance programs.

At present, the circumstances where genomic and biomarker-based information should be included in clinical practice guidelines generally occurs in a nonstandardized case-by-case basis (as warranted) and is relatively limited compared to what may be the case in 5 to 7 years. While clinical practice guidelines focus on standardization, integration of individualized information in the most appropriate format to support broad applications in general practice is somewhat uncertain (e.g., inclusion in clinical practice guidelines, physician reminder programs, decision support systems).
Database and data clearinghouse standards and support: As new health data becomes more readily available and accessible, databases and patient data registries that contain extensive demographic, clinical, utilization, and other information will be a key foundation element of personalized health care. While professional societies do not currently play a significant direct role in developing and maintaining clinical databases (this is largely done by government, payers, and providers), current examples that have been informed by professional societies include claims databases, electronic health record databases, clinical trial data registries, and health quality or performance data sets. In general, databases specifically relevant to utilization of genetic and biomarker-based tests are limited in number and design.[52] These often longitudinal and information rich resources may also be used in clinical and health services research that will augment and in some ways transcend results of conventional clinical research.

Databases and registries can help link use of particular interventions to long-term effectiveness and safety outcomes, enable “real world” evaluation of health technologies, and provide population-level data necessary to appropriately refine health practices in the face of new knowledge. While a myriad of such databases already exist, at present most are not yet sufficiently interoperable to handle the complex applications anticipated for personalized health care. It is also important to note that these databases require strong organizational supports and funding to establish and maintain, as well as substantive input from clinical, statistical, HIT, and other experts to ensure appropriate functionality and usefulness. Such constraints are currently significant limiting factors.

As databases and registries increasingly include genetic and biomarker information and develop interoperability, issues such as scope and anticipated use, relevant expertise, and funding sources will influence the degree and rate of database development. In the absence of proper stakeholder incentives, fully interoperable databases and information systems may not mature for some time.

5. Health quality and pay-for-performance standards:

Development of health quality measures is another activity that some societies engage in. Health quality measures, like clinical practice guidelines, are evidence-based and focused on characterizing standards of clinical practice and patient care. As previously discussed, these measures are often used in provider and physician performance management programs, including P4P programs that tie financial incentives to performance.

While some societies such as the National Quality Forum (NQF), National Committee on Quality Assurance (NCQA), and the Leapfrog Group focus on health quality evaluation and measure development, medical specialty societies may also contribute by translating elements of evidence-based clinical practice guidelines into health quality measures useful in P4P approaches. The extent to which information on individual variation will be integrated into health quality measures is currently uncertain because these approaches leverage quality and efficiency gains based on standardizing health care delivery.

6. Educate and inform evolving health management practices and operational models:

Although most professional societies focus efforts on clinical aspects of medical education, many also provide education, training, resources, and certification related to business management of provider, payer, and other health-related organizations. For example, the National Association of Managed Care Physicians (NAMCP) conducts medical director training “academies” to teach the business skills that clinically-oriented physicians will need to succeed in provider and payer administrative roles. Where personalized health practices will affect processes (e.g., clinical pathways or guidelines, quality measurement and P4P programs) that have financial and operational implications for professional society members, future member training may include education on the implications of individualized health information on health management practices and provider operations.

7. Stakeholder collaboration, communication, and coordination:

Professional societies currently play an essential role in bringing together key health stakeholders (e.g., payers, providers, employers, manufacturers, policy makers) to advance debate and seek solutions regarding emerging health care issues. Professional societies often have much broader “reach” (versus individual stakeholders) into diverse stakeholder groups that can be utilized to address issues and challenges through workshops, advisory councils, and other initiatives. Some personalized health issues are likely to be sufficiently complex that they will warrant collaboration among professional societies (and other stakeholders) to appropriately address certain education, operational, or health policy issues.

Likewise, since professional societies represent a group of health professionals with similar interests, the collective “voice” of the society is often more influential than individual members or member organizations acting alone. Accordingly, societies may also develop opinion and policy statements, practice standards, decision tools, and business practice recommendations, which could include topics germane to personalized health care. The conferences hosted by professional societies provide virtually unparalleled opportunities for addressing health care issues through open sessions, workshops, collaborative initiatives and even informal dialogue among stakeholders.

8. Rational health policy development that supports viable business models and care delivery practices focused on personalized health care:

The activities of professional societies include not only commercial stakeholders such as payers, providers, health technology manufacturers, and HIT companies, but also public stakeholders in government and policy (e.g., CMS, FDA, AHRQ, and CDC). Professional societies have historically worked on a variety of levels to directly and indirectly inform rational health policy development that supports quality clinical practice and innovation in health delivery. For example, life sciences industry organizations such as the Advanced Medical Technology Association (AdvaMed), the Biotechnology Industry Organization (BIO) and others regularly interact with a medical specialty societies and policy makers to inform thoughtful development of health policies that support appropriate health technology adoption and use on a range of topics particularly relevant to personalized health practice.

While not all professional societies are directly involved in the policy making process, most play a role in education and stimulation of healthy debate and discussion of key health services delivery and management issues. In the emerging era of integration of information on individual variation, broad engagement of professional societies will be critical to development and refinement of sound health policies that integrate personalized health care approaches into standardized and complex policy and delivery scenarios. As personalized health care approaches themselves become accepted as standard over time, professional societies will also be important contributors to implementation and harmonization efforts in the global health care environment.

Although professional societies currently play a role in many activities necessary for successful implementation of personalized health practices, emphasis and participation in particular activities will vary markedly by organization. Accordingly, level of interest and willingness to devote resources to personalized health care initiatives will depend upon the organization’s mission, nature of offerings (e.g., content development, message development, policy processes), availability of funding, and relevancy to members. However, as personalized health practices evolve, it is clear that professional societies are poised to facilitate collaboration among key stakeholders and play a role in development of processes, standards, and business practices that incorporate information on individual variation.

Assumptions Regarding Future Dynamics of Health Care Delivery

To understand the role that professional societies may play in supporting transition to personalized health practices, it is important to consider the implications of health care trends and the future dynamics of health care delivery. For purposes of discussion, we will assume a timeline of 3 to 5 years following the publication of this paper and evaluate the likely state of certain factors, listed below, important to broad implementation of personalized health practices and implications for professional societies.

Factor 1: Gene-based and Other Molecular Tests are Routinely Used in Patient Management

While events such as sequencing the human genome have markedly advanced our scientific knowledge, the reality is that a tremendous amount of additional research will be necessary to understand how and in what ways information on individual variation can be used in routine clinical practice. The process of science and clinical discovery simply takes time, even given the rapid pace of technological innovation and emphasis on accelerating the translation of research into practice. Despite the promise of personalized health care, the convergence of science, medicine, and technology will not occur overnight.[53] In general, it takes up to 20 years to move a new treatment or intervention from research into clinical practice.[54]

At present, while use of diagnostic tests is routine in clinical practice, application of complex molecular diagnostics remains comparatively limited for a variety of reasons. These reasons include, but are not limited to physician and patient educational needs, uncertain reimbursement scenarios, and complexity of interpretation. However, as biomarkers are increasingly studied in clinical trials in the coming years, evidence linking diagnostic test information to treatment selection and health services delivery issues will expand in tandem. At present there are approximately 121 drug labels in the US that contain pharmacogenomic information, 69 of which refer to human genomic biomarkers, which is a fair beginning for personalized medicine following publication of a complete draft of the human genome in 2003.[55] Recent efforts, such as the partnership announced in October 2008 between the FDA and Medco (one of the largest pharmacy benefits management organizations), are poised to further accelerate associations between pharmacogenomics and treatment decision making.[56]

Another key challenge will be overcoming educational barriers for use of some complex tests in physician decision making, particularly in the context of general and family practice.[57],[58] To fully integrate personalized health care, it will be important to create an environment where physician ordering and interpretation and patterns of test use linked to treatment selection/utilization are standard practice. It is likely that expanded emphasis on personalized medicine and information management will occur in clinical and health care management training programs in the next 3 to 5 years, and this is already occurring in academic settings that train new health professionals. To expedite this educational process, professional societies can play a key role in creating and supporting medical education and certification programs, training on emerging decision support systems, and promoting a learning and collaborative environment for personalized health care.

Factor 2: US HTA and Reimbursement Infrastructure Sufficiently Enables Personalized Health Care

While 55% to 65% of US medical and pharmacy directors and physician decision makers feel that personalized medicine will be transformative and usher in new paradigms of personalized care delivery, a recent survey conducted by the National Association of Managed Care Physicians (NAMCP) indicates that these gatekeepers and decision makers recognize the following key challenges facing personalized health care:[59]

Limited information linking diagnostic information to treatment decisions and outcomes
Inconsistent definitions of clinical utility for diagnostics
Limitations of current HTA practices (i.e., not sufficiently geared for personalized health scenarios)
Physician and health practitioner understanding and adoption
Limited uptake of electronic health records and information systems for implementing personalized health care approaches

Many of these issues relate to processes for evidence-based practice, HTA, and translation of research into clinical practice. At present, there is significant uncertainty regarding evidentiary requirements and decision criteria for diagnostics, drug-diagnostic combinations, costly biologics and other scenarios relevant to personalized health care, particularly from the perspective of third-party payers and policy makers.[60],[61] Because of this uncertainty, public and commercial payers (e.g., CMS and the Blue Cross Blue Shield Association), government-affiliated groups (e.g., EGAPP), and private organizations (e.g., ECRI Institute, Hayes, Inc.) are beginning to develop approaches to overcome these obstacles, fill existing gaps, and provide information relevant to decision makers.[62] Professional societies can liaise with these stakeholders to ensure that clinical and methodological perspectives and implementation issues are appropriately aligned will “real world” decision making needs.

Recent authoritative reports produced by the Institute of Medicine; the Secretary’s Advisory Committee on Genetics, Health and Society; and AdvaMed have also cited significant insufficiencies in the reimbursement systems associated with molecular diagnostics. Insufficiencies include HTA and coverage processes associated with diagnostics, as well as medical coding and payment approaches that do not keep pace with technological development or do not appropriately reflect the value of tests to patient care and health outcomes.[63],[64],[65] As some of these barriers to innovation and expansion of personalized health practice are addressed over the coming 3 to 5 years, professional societies can play an important role in informing development of rational policies and health delivery practices.

Factor 3: Prevention and Risk Assessment Approaches that Incorporate Genetic Testing are Standard Practice

In large part, incentives in the US health care delivery system are geared to support “sickness care” and not “wellness care” that focuses on early disease identification and prevention. Because of the large volume and costs associated with preventive health efforts, including screening of asymptomatic patients at risk for disease development, the evidentiary threshold for demonstrating value is high and uptake has been historically limited. For example, Medicare statute prevents use of screening and prevention tests, except as amended by Congress. Since the late 1960s, fewer than 20 diagnostic tests have been approved for screening applications, including coverage of staple tests such as cholesterol testing, prostate-specific antigen testing, fecal occult blood testing and diabetic screening.[66] Additionally, as employment longevity has decreased in the US, commercial health plans have historically been reluctant to support preventive testing for beneficiaries that may only remain in the plan for 12 to 24 months in scenarios where disease may occurs years later.

Greater emphasis from a variety of stakeholders and different incentive structures supporting preventive health services will be necessary to fully realize personalized health efforts in the coming years. Professional societies can play a variety of roles in supporting advancement of preventive health services, ranging from providing input on the viability and business implications of preventive health strategies and applicability of emerging technologies to influencing appropriate policies that support “wellness care.” Efforts may also include member education and training on how preventive and disease management programs can incorporate information on individual variation and maintain efficiencies gained by practice standardization.

Factor 4: Electronic Health Records and Decision Support Systems are a Mainstay in Hospital and Multi-physician Practices

As previously discussed, low provider adoption of electronic health records (EHR) and lack of interoperable health information systems will limit our ability to provide personalized health services. Further, the decision support tools that would improve processes for leveraging individual health information are presently in an early stage of development. Despite government, commercial MCOs, and other initiatives that provide incentives for providers to quickly adopt these systems , issues such as perceived benefit/burden tradeoffs associated with this capital investment, implementation concerns, the pace of technology turnover, and lack of standardized approaches will remain substantial barriers to acceptance over the next 3 to 5 years.

Factors that would be necessary to support EHRs and decision support systems adoption and implementation for personalized health practices include:

Systems that help to identify treatment practices beneficial to specific patient subgroups. Development of systems with these capabilities will be strongly influenced by the use of EHRs and decision support systems within the overall market and demand based upon perceived benefits of customization.
Knowledge translation practices based on differential patient outcomes.
Standards for integrating this information into clinical practice guidelines and quality measurement and/or performance management programs at the multi-physician practice level and the individual level.
Federal and other incentives for providers to collect and report data (in a standard format from EHRs) on gene-based and other molecular test information.

The government and private sector must provide strong incentives to support uptake of interoperable health information systems and their evolution as broadly adopted and routine tools to guide care practice. Sound policy and payment incentives that encourage well-developed provider organizations in addition to data reporting requirements that currently provide disincentives for nonparticipation will expedite HIT uptake and use.

As health information capabilities and knowledge networks evolve, professional societies may play a role in developing the content of clinical decision support systems and/or managing population-level data from member organizations if business incentives are appropriate to support these actions. Professional societies must consider either the availability of a suitable customer base willing to pay to access this information or the viability of partnering opportunities with HER vendors or physician practices/health systems.

Factor 5: Provider Education and Certification is Increasingly Tied to Health Care Quality and Best Practices Initiatives

Provider education delivered by both academic educational centers and professional societies has recently increased emphasis on topics such as use of biomarkers in medical decision making, disease prevention and management, implications of genomics and personalized medicine on managed care, and trends in electronic health records and quality/performance management programs. However, education offerings relevant to personalized health care are currently geared towards making the fundamentals of this topic comprehensible to providers, payers, and other health stakeholders. Further, medical certification and licensure requirements for many physicians do not yet include elements of personalized health care, but will likely need to in the future.

As clinical standards and quality measures emerge that incorporate elements of personalized health practice, professional societies can play a strong role in creation of tools that are appropriately aligned with health delivery practices and patient needs. Such tools, if supported by federally-funded initiatives and MCOs, are likely to initially target high cost/high need chronic disease areas (e.g., diabetes, heart disease, cancer).

However, the extent to which patient-specific guidelines and measures incorporating elements of individual variation will be implemented into quality management and health reform efforts is currently uncertain. At present, the majority of clinical guidelines often lack appropriate specificity for development of quality and performance measures, without adding individual variability into the mix.[67],[68] In addition, the ability to which we can incorporate personalized health information into clinical decision support systems is still a nascent area with significant room for development.[69]

Factor 6: Professional Societies Play a Key Role in Evidence Evaluation and Implementation of Knowledge into Clinical Practice

Professional societies have historically played a key role in the translation of new knowledge and technology into clinical practice. It is reasonable to assume that in the future, this role will extend to personalized health practice. However, the pace of innovation and our capacity for generating information outstrips our capacity to translate new knowledge into meaningful health improvements. As such, professional societies will serve to help clinicians and other stakeholders adapt to technological innovation, information management, and new business practices that are the foundation elements of personalized health care.

Education on the expanding range of new diagnostic and treatment technologies will be critical to correct use of personalized health care technologies. Professional societies can aid in the evaluation and introduction of new technologies and clinical practices by serving as an interface between various health stakeholders, including payers, providers, and technology developers. As previously discussed, this mediation will occur through activities such as providing input on methods and standards for health outcomes and comparative effectiveness research, development of clinical practice guidelines and health quality measures, and informing development of practical tools for knowledge management and clinical/business decision making.

As part of the process of knowledge transfer, professional and scientific societies may also play a role in the conceptualization and validation of viable business models for personalized health care. In part, this can be accomplished by hosting conferences, workshops, and focus groups that address issues relevant to personalized health care practice and policy. As business models emerge, society activities will also include training of physicians and other clinical care providers to ensure that health delivery processes and standards appropriately incorporate knowledge of genetics and individual variation.

A Framework for Professional Societies to Play a Role in Enabling Delivery of Personalized Health Care

What Framework is Needed to Address Personalized Health Care?: In considering a framework that HHS might adopt to encourage uptake and implementation of personalized health care practices, it is important to recognize key barriers and then align strategic initiatives to overcome them. The validation, adoption, and diffusion of personalized health care practices may be limited by several barriers, many of which are common to introduction of any new health technologies and/or changes in clinical practice or standards of care (see Table 3).

In the case of personalized health care, failure to overcome any one of these barriers will influence not only the rate and range of stakeholder acceptance, but also holds the potential to forestall integration of some practice applications altogether. For example the level of uptake of health information management systems and business model implications are two factors that would broadly delay personalized health care efforts on the whole. Another key factor is that the concept of personalized health care is sufficiently comprehensive that unless it is broken down into actionable elements, it will be difficult to address and operationalize.

Table 3: Key Barriers to Integrating Personalized Health Care Into US Health Care Offerings
Action Category	Barrier
Break down integration of personalized health care into discrete tasks	Characterizing specific initiatives relevant to personalized health care in a manner where anticipated stakeholder involvement is well defined and focused on clear objectives Leveraging stakeholders (including professional societies) that are instrumental in accomplishing a particular goal or initiative
Develop appropriate evidence and information to meet decision needs	State of the science and ability to implement personalized health care practices in a clinically meaningful manner Clinical trial, registry, and database standards Pilot and demonstration programs Disease-specific focus areas HTA and evidence-based practice methodological limitations (including acceptance of novel clinical trial and data modeling approaches) Diagnostics (including molecular diagnostics) Treatments subject to personalized medicine scenarios Combination product scenarios (involving diagnostics, drugs and devices) Innovative methods focusing on “least burdensome” approaches that provide “need to know” evidence Communication of medical and personalized health information in a cost-shifting and information overload environment Address the information needs of providers, payers, and patients Inefficient processes for translating clinical research into practice (including clinical guidelines and practice standards, quality measures and performance management programs) Clinical guidelines and practice standards Health care quality measures Approaches for vetting decision tools and processes relevant to provider
Define business plans and appropriate operational policies	HTA and evidence-based practice implementation limitations (i.e., standardized processes do not anticipate personalized health care approaches) State of interoperable EHRs and decision support systems and their level of uptake in practice Health information technology and decision support system adoption issues Data collection and reporting standards and policies Data analysis approaches Business model, operational, and policy implications of integrating personalized approaches into processes that seek to gain efficiencies through standardization
Implement plans for personalized health care	Bringing stakeholders together to collaborate on appropriate policies and implementation plans Insufficient incentives for health care stakeholder participation Manufacturer Hospital and clinician Health plan Professional society Patient Implementation costs of personalized health care approaches, including lack of transparent and/or viable public/private partnership models Approach for pulling the various discrete tasks together and monitoring implication and impact of personalized health care Collaboration and stakeholder engagement Oversight and cross-functional alignment Approach for influencing changes to improve quality and efficiency of operational elements (as appropriate)

These action categories (including associated barriers), form a very general framework from which HHS and key health stakeholders can parcel out and address elements critical to implementing personalized health care. While not specific to personalized health care, it should be noted that a variety of US government initiatives geared towards addressing these barriers are already in motion.

For example, the National Institutes of Health (NIH) and the AHRQ have implemented a variety of initiatives aimed at addressing translation of research into practice over the past decade.[70],[71] The Centers for Disease Control and Prevention (CDC) and FDA have also made strides in recent years regarding integration on information on patient variability into technology evaluation and population-health programs. Likewise, in 2004 President George W. Bush outlined a plan to support adoption of interoperable EHRs and issued an Executive Order to create a National Coordinator for Health Information Technology within the Office of the Secretary of HHS to facilitate this plan.[72] These are just a few examples of existing HHS efforts that can be leveraged to explore opportunities for improving personalized health practice.

Understanding how to weave personalized health care into this framework in a manner that is not duplicative of existing efforts is also an important consideration for HHS, but outside of the scope of this white paper. Factors such as strong leadership support, data to support implementation start-up and evaluation, degree of required organizational change, collaboration requirements, sustainability planning, and dissemination infrastructure have played significant roles in the overall rate of adoption and diffusion and would also be relevant to making progress against a framework for personalized health care.[73] By clearly defining objectives and anticipated outcomes, the approaches and relevant stakeholders necessary to advance personalized health care will be more transparent and easier to accomplish and will enable appropriate prioritization among objectives.

Integrating Professional Societies Into a Framework that Supports Personalized Health Care: As strategies for operationalizing personalized health care practices continue to move forward, professional societies will play a pivotal role, both in regard to short-term evaluation and planning, as well as long-term implementation support. Such organizations are unique in their ability to connect key health stakeholders, provide a neutral grounds for healthy debate and discussion, enable educational and health practice tools and solutions, and support “big picture” objectives outside of the capacity of individual member or affiliate organizations.

In regard to engaging professional societies in efforts targeting personalized health care practices, many societies may embrace the promise of personalized health care, but remain uncertain about specific action steps and their implications for members. Because professional societies operate as any other business, the greater the clarity of a proposed engagement, the easier the evaluation of relevance and participation becomes.

As HHS supports key practice and policy efforts in this area, the following business and operational requirements will be key to anticipating the scope and nature of relevant professional society participation:

Alignment with society mission and vision
Perceived relevance to member interests and needs
Perceived relevance to funding organizations that enable key products or offerings
Alignment with specific “deliverable” offerings and tradeoffs necessary for implementation
Implications for competition with other professional societies or stakeholders
Benefits and risks of stakeholder partnerships around key goals and objectives
Funding requirements (for implementation and sustainability)

Similar to the manner in which barriers to personalized medicine may limit adoption and uptake, it will be important for HHS and other stakeholders in the vanguard of personalized health care to anticipate the extent to which specific initiatives will appeal to professional societies. The more closely aligned the desired objective is with these requirements, the greater the likelihood of securing participation.

The Road Ahead: Enabling the Personalized Health Care Environment

Personalized health care is a complex concept involving many aspects of health quality and efficiency improvement. Because the concept is broad and far reaching, it will be challenging to predict and plan for all of the health delivery and systemic implications of increasing integration of individual variability in health practice. While initial steps will likely be addressed on a scenario-by-scenario basis, it will be important to maintain perspective on the implications for health care delivery on the whole as personalized health care unfolds.

Information inputs envisioned for personalized health care appear to be potentially boundless and complex. In an age of information overload, it will be essential to channel knowledge into decision support systems, “smart tools,” and delivery approaches that better inform health decisions and presumably generate better health outcomes. [74] If integrated effectively, these changes in health care delivery may also refocus our current “sickness-based” system on disease prediction and prevention. The most effective models will balance standardization, best practices, and population gains with personalized health care practices, with greater emphasis placed on one or the other as appropriate to the scenario.

It is clear that professional societies have a fundamental role to play in the new era of personalized health care. While some issues and operational processes will lend themselves to personalization more readily that others, professional societies are cognizant of the potential benefits of personalized health care in a US health environment facing serious challenges and hard decisions. Appropriate engagement of professional societies around specific and well-defined personalized health care issues will require complex orchestration and planning on the part of HHS. Nevertheless, weaving professional societies into decision and implementation steps is likely to confer far reaching benefits by mobilizing key stakeholders and establishing a rational and balanced pathway forward.

Successful implementation of personalized health care will rely on the ability of key health stakeholders to work collaboratively towards practical and sustainable health solutions. Policy makers, professional organizations, payers, providers, employers, health technology manufacturers, and patients must all develop a common understanding of the cause and effect of decisions regarding integration of personalized health practices, including implications for particular stakeholders or market segments. In light of escalating health care costs and threats to sustainable provision of health services, the opportunities represented by personalized health care are great, as is the price of failure to collaboratively forge well founded solutions for the road ahead.

Stakeholder Type	Drivers/Incentives	Implications for Personalized Health Care
Sources: The value of diagnostics: innovation, adoption and diffusion into health care. Advanced Medical Technology Association. Jul 2005; Overcoming barriers to electronic health record adoption. Health Care Financial Management Association 2006; Knowing what works in health care: a roadmap for the nation. Institute of Medicine. National Academy of Sciences 2008; and Faulkner E. The road to personalized health care: translating promise into practice. J Manag Care Med 2007;(10)6:25
Medical Professional Societies	Improve care and health delivery practices Focus on organizational mission and vision Provide education and support to members Maintain operational status by developing offerings/services that provide value to members	Implications of personalized health care for medical professional societies will depend upon the organization’s mission, nature of offerings (e.g., content development, message development, policy), and relevancy to members. Medical professional societies can serve as a bridge between stakeholders by providing opportunities for engagement such as annual conferences, working sessions, position and policy statements, Webinars, and other outreach activities.
Provider Organizations	Improve patient health outcomes Serve patients and the community Offer current and appropriate health services Maintain profitable and competitive service offerings Meet quality and P4P milestones (if applicable)	While personalized health care information offers opportunities for improved treatment selection, patient management, and health outcomes, need for education and decision support for physicians remains critical. Personalized health information must be collected and communicated in a format that is readily useful to the physician during a typical patient consultation. As information on genetic and other variability becomes increasingly available (e.g., via guidelines, standards of practice) the need for concise, easy-to use decision support tools will become more pronounced. Personalized health data will also add to administrative data collection and reporting requirements in addition to those associated with claims processing, quality, P4P, health transparency, and other initiatives. Provider adoption of electronic health records will limit the uptake and diffusion of systems and practices that promote personalized health care.
Health Insurance Plans	Administer health plan assets effectively via policy creation, contracting, and processing of claims Ensure the quality and affordability of beneficiary services Ensure beneficiary access to the broadest array of beneficial services Limit access to unproven/unnecessary services and/or health technologies Maintain profitable and competitive service offerings	Health insurance plans will likely embrace aspects of personalized health care that better characterize value of new health technologies for particular patient categories. This knowledge will be implemented through coverage policies, claims review, quality and P4P initiatives and other programs. On the one hand, payer leveraging of personalized health information will support patient access to the right treatment at the right time and dose where evidence of benefit is clear, presumably improving quality and effectiveness. On the other hand, overly aggressive approaches by payers have the potential to prematurely limit patient access to beneficial technologies based on incomplete information on population versus subpopulation safety and effectiveness. Compared to treatments, criteria and processes for evaluating emerging diagnostics is not clear or well defined from the payer perspective. While such processes are likely to emerge in the next 1 to 2 years, uncertainty regarding clinical utility is likely to result in coverage limitations/noncoverage for tests without a clear value proposition. Evidence assessment for diagnostics and treatments may be conducted by different groups within a payer organization. Noncoverage of a test that is directly linked to treatment use (e.g., as recommended in the product label) may limit or preclude patient access to the treatment in some scenarios. Increased uptake in electronic health records and availability of longitudinal databases that contain personalized health information may ultimately enable identification of “real world” trends in treatment response not captured in manufacturer pivotal or postmarketing studies and more refined population-level beneficiary management capabilities.
Diagnostics Manufacturers	Develop innovative health technologies that improve patient care/health outcomes in areas of unmet need Create and maintain market opportunities for pipeline health technologies At the highest volume and price supported by the market That withstand pressures of competition, changing health policies and service delivery trends Increase revenue and meet expectations of external investors/stockholders	Because diagnostics do not have similarly robust profit margins compared to drugs, opportunities for innovation must be balanced carefully against the following factors. Improvement in one or more of these limitations to diagnostic adoption and diffusion would increase manufacturer incentives to develop additional tests and decision support offerings. Increasing evidentiary (e.g., direct evidence of clinical utility) and funding requirements—for more complex and costly studies than have historically been required for demonstrating the value of diagnostics Outdated and uncertain coding and payment structures that do not fully characterize or value tests (which represent 2% to 3% of US health care expenditures but influence over 75% of health decisions) Limited stakeholder support of preventive/predictive applications, particularly screening of asymptomatic patients Additionally, because diagnostics may be developed through either the FDA or Clinical Laboratory Improvement Ammendments (CLIA) mechanisms, nonmanufacturer health stakeholders have questioned the differential evidence on product safety, effectiveness and value resulting from these mechanisms. More consistent evidence requirements would improve stakeholder assessment of diagnostics as new test development increases significantly over the coming years.
Treatment Manufacturers	Develop innovative health technologies that improve patient care/health outcomes in areas of unmet need Create and maintain market opportunities for pipeline health technologies At the highest volume and price supported by the market That withstand pressures of competition, changing health policies and service delivery trends Increase revenue and meet expectations of external investors/stockholders	Drug and biologics manufacturers: While treatment manufacturers initially resisted drug development approaches that would limit treatments to patient subpopulations in favor of the blockbuster approaches, health reform trends, increasing reimbursement hurdles, and the success of forerunner personalized medicine approaches appear to be changing this perspective. Emphasis on personalized health practice remains limited, but is expanding. Obtaining coverage for a limited market is preferable to non-coverage as payers increasingly limit coverage based upon differential evidence of value. The most widely embraced use of genomic and related information is to “salvage” treatments that may fail to be approved for broader use but where benefits to specific patient subgroups are transparent. Proactive approaches (e.g., drug-diagnostic codevelopment) are expanding as manufacturers realize the potential for securing limited markets at viable price points. Increasing propensity for regulator label changes based on the growing knowledge base correlating biomarkers and treatment response requires manufacturers to alter perspectives on personalized health care.
Policy Makers and Regulatory Agencies	Ensure that health services are sufficiently safe and effective Ensure that access to health services are provided in an ethical and efficient manner Improve access to beneficial health technologies and services by supporting R&D and translation of research into clinical practice	Emphasis on personalized health care must be balanced against other competing efforts that seek to improve the quality and effectiveness of health care in the US. However, personalized health care is broadly related to many ongoing health reform efforts and consideration of the implications of integrating information based upon genetic variability would be relevant to many ongoing HHS and public/private partnership efforts.
Patients	Access the broadest selection of quality health services Ensure that health services are affordable in the context of other financial and life requirements Maintain the lowest possible hurdles or obstacles to access	Increased patient cost sharing is a key consideration for personalized health care. Going forward, patient choices will be influenced by a variety of factors, including perception of need, disease severity, payer and provider barriers to care access, financial tradeoffs required to medical care and availability of alternatives. Patient selection of and access to personalized/targeted health services that may be priced higher than existing alternatives, will be determined by health plan design (e.g., copay structures, use of health savings accounts) and affordability of these care options.

The authors would also like to acknowledge William T. McGivney, Chief Executive Officer of the National Comprehensive Cancer Network; William C. Williams III, President, National Association of Managed Care Physicians; Bradford Walters, Chief Medical Officer, RTI International; Samuel L. Warburton Jr., Professor and Chief, Duke Community and Family Medicine, Steve Ubl, President, Advanced Medical Technology Association, Marylin Dix-Smith, President, International Society for Pharmacogeconomics and Outcomes Research, and Janet M. Corrigan, President and CEO, National Quality Forum for their thoughts and contributions.

Return to Table of Contents

The Role of the Academic Medical Center
in Advancing Personalized Health Care

Judd Staples, MBA
Entrepreneur in Residence, Center for Genomic Medicine

Robert Cook-Deegan, MD
Director, Center for Genome Ethics, Law & Policy

Geoffrey S. Ginsburg, MD, PhD
Director, Center for Genomic Medicine

Duke Institute for Genome Sciences & Policy
Duke University

The use of genomic, molecular, and imaging technologies holds the promise of improved medical decision-making and advancement towards personalized healthcare[1]. Yet, despite the vast number of research discoveries based on the genome sciences, relatively few have been translated into medical practice. The pharmaceutical industry has developed only a handful of ‘targeted therapies’ such as trastuzamab, imatinib, and erlotinib that use molecular diagnostic tests to identify patients who are likely to benefit. Diagnostic companies offer a similarly sparse repertoire of new genomics-based molecular diagnostics that can be readily deployed in the course of care. Of the few genetic tests that have been approved by the FDA many are based on genetic variation that has been known for decades[2]. In short, the market has been challenged to move personalized health care from a ‘nice concept’ to a reality in clinical practice.

Each of the stakeholders in healthcare delivery stands to gain from a more comprehensive strategy to implement personalized medicine in health care systems (See Table 1). For example, payors will realize savings from lower use of ineffective drugs, patients will avoid adverse drug reactions, and diagnostic companies will realize higher margins on their tests. Despite these obvious advantages, investment in the clinical development and deployment of personalized medicine discoveries has been modest at best. The comparatively low development activity of the personalized medicine discoveries may be attributed to the fact that, aside from a few exceptional cases, the value created by commercializing personalized medicine cannot efficiently reach those required to invest in its development.

The goal of making genomics clinically relevant and realizing the full potential of personalized healthcare (PHC) can be achieved if market forces and regulatory policies are aligned, such that academia, industry, government, payors, and advocacy groups are

motivated to share information and resources while equitably distributing the resulting increase economic value. Academic Medical Centers (AMCs), as the locus of discovery, validation, and clinical implementation of these new tools, will be a key enabler of this strategy. Specialized centers or institutes within AMCs focused on personalized medicine will serve to foster collaboration, information sharing, and appropriate handoffs among the diverse group of stakeholders. These centers will facilitate the development of consensus evidentiary standards for new technologies to be adopted into clinical practice and define common data standards to prospectively collect and share knowledge of the complex association of biology, health and disease progression. It will be incumbent upon AMCs to play a leadership role in enabling the realization of PHC by developing and implementing new organizational, care delivery, and funding models as well as adopting new intellectual property licensing and conflict of interest policies[3].

Players across Healthcare Value Chain Stand to Benefit from Increasing Prevalence of Personalized Medicine

Player	Benefit
Pharma Cos.	Enhance clinical trials (faster, smaller, higher POS) Diversify Rx through higher efficacy, tolerability and/or safety
Diagnostic Cos.	Introduce specialized tests with higher margins Spearhead drug development initiatives
Regulators	Improve certainty of drug efficacy and avoid side effects Faster approvals
Providers	Deliver differentiated and better care to patients Derive value from patient data
Payors	Reduce high cost of ADRs and ineffective drugs Earlier diagnosis impacts costs, but not always positive
Patients	Improved drug safety and efficacy Personal knowledge of their risks and value of therapy
Physician	Potential to deliver better care to patients New sources of revenue from KOL consulting

Source: McKinsey & Company

The translation of genome based discoveries, novel biomarkers, and predictive models from bench to bedside are fundamental to the development of PHC. A four-phase framework (T1, T2, T3, and T4) has been proposed by Khoury, et al., to describe this “translational continuum” (Table 1)[4]. A successful PHC application will need to traverse discovery to initial (“first in human”) health application (T1), clinical validation to evidence-based guidelines (T2), to general clinical practice (T3), and to population and public health impact (T4). The AMC will play a role in each phase and must partner and/or adapt its organization and policies to advance new health care models in order to achieve the full impact of these innovations.

Table 1
The continuum of translation research in genomic and personalized medicine: types of research
Translation research phase	Notation	Types of research
T1	Discovery to candidate health application	Phases I and II clinical trials; observational studies
T2	Health application to evidence-based practice guidelines	Phase II clinical trials; observational studies; evidence synthesis and guidelines development
T3	Practice guidelines to health practice	Dissemination research; implementation research; diffusion research Phase IV clinical trials
T4	Practice to population health impact	Outcomes research (included many disciplines); population monitoring of morbidity, mortality, benefits and risk

AMCs as sources discovery in personalized medicine (T1)

AMCs are uniquely suited to discover and perform the preliminary development of the next generation of biomarkers. On AMC campuses the physical juxtaposition of academic research, medical education, leading technologies, and clinical care provides an excellent environment for investigators to develop an understanding of the unmet needs of the market, to discover novel solutions, and to validate their efficacy in experimental models and in clinical cohorts and test their utility in ‘real world’ healthcare delivery settings.

Academic researchers, unlike researchers in private industry whose budgets are primarily dictated by expected investment returns, have the freedom to explore new areas of science with less immediate regard to financial return on investment. Moreover, academic research aspires to operate with norms of Mertonian “open science” as opposed to the generally more proprietary model of R&D in industry. Consequently, AMCs can more readily collaborate, build off one another’s discoveries, and foster more “disruptive thinking” that can bring about new technologies and approaches and introduce entirely new capabilities, rather than incremental refinement and improvement on existing techniques.

AMCs: a strong record of innovation. Academic research institutions have contributed many of the discoveries leading to genomic technologies. The basic methods of DNA synthesis were pioneered at the University of Colorado by Marvin Carruthers. The Maxam-Gilbert DNA sequencing method was developed primarily at Harvard, and Sanger-Coulson sequencing at the University of Cambridge. The prototype for automated Sanger sequencing, using the four-color fluorescent method, was pioneered at Caltech. DNA-chip microarray technologies commercialized by Affymetrix and Agilent drew on Stanford research, and the Illumina bead-array technology grew out of analytical chemistry at Tufts University. Some of these methods were patented (four-color DNA sequencing, DNA lithography, bead-arrays); some were not (Sanger-Coulson and Maxam-Gilbert sequencing). Sequencing methods were widely adopted for academic research, but large-scale sequencing depended on instruments developed by Applied Biosystems and other firms, and microarray technologies were developed by many companies. Competition for the new generation of high-throughput DNA sequencing, en route to the $1,000 genome, is intense among several firms and academic researchers.

Academic research centers have also been deeply involved in another set of discoveries directly pertinent to the advance of genomics-guided medicine – the association of an individual’s molecular biology, both static (e.g., DNA sequence, gene copy numbers, and single nucleotide polymorphisms SNPs)) and dynamic (e.g., gene expression, protein and metabolite levels), with clinical phenotypes. Genomics-guided medicine has grown out of the quest for disease-associated genes that accelerated in the 1980s. This revolution began when genetic linkage maps were used to find mutations associated with Mendelian conditions such as Huntington’s disease and cystic fibrosis. It then expanded into Mendelian forms of diseases with multiple causes, such as Alzheimer’s disease, and inherited susceptibility to conditions such as breast and ovarian cancer or colon cancer. Genetic testing, once restricted to a handful of newborn screening tests, has expanded to include hundreds of tests. At the end of August 2008, for example, www.genetests.org listed 595 laboratories testing for 1610 conditions.[5] With the commercialization of efficient discovery platforms for the measurement of dynamic biological parameters such as gene transcription factors, proteins and metabolites in the 1990’s, genomics-guided medicine expanded to include diagnosis and prognosis of non-Mendelian conditions. Hundreds of gene expression, protein, and metabolite “signatures” are under investigation at AMCs and diagnostic companies as potential tools for use in PHC. Arguably, these T1 research efforts are only possible because of the ability of academic investigators to ascertain and bank high quality clinical specimens from patients and to link these to robust clinical phenotypic data and longitudinal follow up and health outcomes (see below).

Cancer research today is a spectacularly promising example of the AMC’s role in shaping the future of genomics-based personalized cancer care. “The Cancer Genome Atlas,” collaboration between the National Cancer Institute and the National Human Genome Research Institute and several AMCs, aims to develop novel tools for the detection and treatment of cancer. This program utilizes technologies such as large-scale genome sequencing (of germ line and somatic DNA) to better understand the molecular basis of a variety of tumors (glioma, non-small cell lung cancers, and ovarian cancer). The overarching goal of this project is to improve capabilities for preventing, diagnosing, and treating cancer at a personalized level. This program and others will result in a paradigm of medical care is based on our ability to match accurate prognosis and proper therapy to the molecular characteristics of the individual and with the individual patient’s tumor. Whole-genome expression data from this effort and other in the AMCs are now being used routinely to identify subtypes of cancer not previously recognized by traditional methods of analysis: profiles and patterns that identify new subclasses of tumors, such as the distinction between acute myeloid leukemia and acute lymphoblastic leukemia[6], or Burkitt’s lymphoma from diffuse B cell lymphomas[7], without prior knowledge of the classes. More recently several genomic signatures that go beyond disease classification have been discovered and validated that predict prognosis and response to therapy for many solid tumors and hematologic malignancies.[8]^,[9] Much of the science that underlies associating genomic data with clinical decisions has and will continue to come from AMCs. For now, these technologies are mainly research tools, but they will surely become relevant to clinical decisions with the proper investment in their development.

Funding of innovation: A changing landscape. AMCs have been the main recipients of grants for health research, and home to most “public domain” research from which further research and practical applications arise. Innovative technologies described above have resulted in part from these funding streams. A survey estimated government and nonprofit genomics research and development (R&D) spending from 2004-2006 at $3 billion annually,[10] in rough parity to a separate survey that estimated private genomics R&D at $3 billion.[11]^,[12] This balance between private and public genomics R&D is a dramatic change from the early 1990s, when private genomics funding was sparse. By 2000, however, the $2 billion R&D expenditures by publicly traded firms wholly or partially devoted to genomics and another $1 billion genomics R&D at established pharmaceutical and biotechnology firms exceeded the $1.8 billion reported in government and nonprofit R&D.[13] Private genomics R&D is a major force now; AMCs are at the point of convergence between government and nonprofit funded genomics R&D and privately funded genomics, although we know of no estimate of private genomics R&D at AMCs. How these funding shifts will affect the balance between innovation in the AMCs and private firms is uncertain; public-private partnerships (see “A Call for Specialized Centers” below) may yield the greatest productivity from these investments, and AMCs will be essential elements of such partnerships.

Role of Intellectual Property in developing personalized medicine at AMCs. Academic institutions own a much larger share of patents relevant to DNA diagnostics and prognostics than in most other areas of technology, because much of the research studying linkages between genomic factors and disease is federally funded through the NIH or other government and nonprofit sources—with a disproportionately large fraction conducted at research institutions associated with medical schools.[14] While AMCs account for somewhat less than 2 percent of patents overall,[15] government funded research institutions accounted for 39 percent of DNA-based patents 1980-1993,[16] a more than ten-fold enrichment of academic patent ownership compared to patents overall. A preliminary analysis of patents licensed by one major diagnostics firm, Athena Diagnostics, showed more than three-quarters of the relevant “gene” patents were owned by academic institutions.[17]

This prominent role of academic research institutions suggests that sometimes AMCs will be patent owners, sometimes they will need to license patents owned by others, often they will be working in conditions of uncertainty about whether their research—and even more so, commercialization strategies—enjoy freedom to operate or will be subject to patent enforcement. This is starkly different from the patent regime long associated with protein and small-molecule therapeutics, where the zone of uncertainty is smaller because only one or a few key patents cover a small class of molecules. However for business plans being developed today, complex patent landscapes portend uncertainty for the future of DNA-based technologies.

The practice of AMCs governing patenting and licensing of genomic technologies, as both users of the inventions and also as patent-owners, is crucial. Academic institutions have emerged as owners of intellectual property for several reasons. The main reason is that the research they do is fully intended to have practical benefits, creating knowledge that enables development of products and services to improve health. Most health research falls squarely in what the late Donald Stokes called, “Pasteur’s Quadrant,” meaning it is both scientifically important and also has foreseeable practical use.[18] It thus often produces results that can be patented because they are novel, useful, and inventive.

Universities, in the past, have patented some inventions, including drugs and vaccines. Thyroid hormone, vitamin D, warfarin, insulin, and antibiotics (although notably not penicillin) were first described in patents owned and administered by academic institutions. However, the level of academic patenting accelerated in the 1980s, mainly because of the science and technology being pursued, but also because the Bayh-Dole Act of 1980 clarified the default rules for ownership of patents.[19] The Bayh-Dole Act increased consistency among federal R&D funding agencies and it codified the emerging practice of having grantee and contractor institutions own patent rights, rather than government retaining ownership of patents arising in federally funded research. Thus the Bayh-Dole created an incentive for academic institutions to seek patents so they could license them.

Academic institutions responded accordingly by getting many more patents, and this effect, as noted above, is particularly pronounced in DNA-based technologies. Commercial biotechnology in general, and genomics in particular, grew up almost entirely in the Bayh-Dole era, with incentives for universities and AMCs to patent inventions arising from research, and giving them control of licensing of the resulting intellectual property. The development of the Affymetrix chip technology, for example, drew upon Stanford research and personnel, entailed several grants directly to the nascent company, and benefited from federally funded research.[20] The development of Illumina technology is also a classic Bayh-Dole story of a research idea at Tufts being developed by a startup firm with exclusive rights to university patents.[21] In both cases, a big part of the first market for the resulting technology was academic health research, so universities were involved in creating the technologies and later benefited from the availability of powerful new instruments developed by startup firms.

Many DNA patents have been exclusively licensed, and many of the uses of those patents were not foreseen at the time the patents issued and licenses were signed. For DNA sequence patents exclusively licensed for the full patent duration of the patent, even if the exclusive rights were restricted to diagnostic use, these prior intellectual property rights could cast a shadow over the development of genome-wide diagnostics, or over the first-generation “personal genomics” services that have recently become possible through companies like Navigenics, 23andMe, deCODEme, SeqWright, and Knome. The degree to which a legacy of existing patents and licenses affects the future of multi-gene tests will depend on: (1) the specific language of patent claims, (2) specific terms under which the patents have been licensed, (3) the outcome of any cases that set precedents in litigation, and (4) decisions about whether and to what degree patent rights are enforced against the new uses.

As DNA patent holders and also users of the technologies, AMCs will be making these choices. It will be a challenge. Patents and their claims are public, but collecting and analyzing all the relevant patents and interpreting how their claims might affect for a multi-gene test is a daunting task fraught with uncertainty. It is made even more difficult because terms of licenses, which are crucial to determining the boundaries of intellectual property, are generally not public unless licensers and licensees choose to make them so. To the degree AMCs contribute to this inefficiency, they may impede the advance of genomic discoveries into medicine.

AMCs role in biobanking and patient registries: important sources of discovery and validation for novel molecular tools for PHC

. As medicine moves toward PHC, molecular analyses of biological samples will provide a critical component of clinical decision-making. Well-annotated biospecimen collections have enabled the recent identification of genes and genetic loci with over 180 publications documenting over 660 SNPs that appear to contribute to susceptibility and survival to over 100 complex diseases[22],[23],[24],[25] (www.genome.gov/26525384) . Indeed, the acceleration of the clinical application of genomic testing and public health planning (T1 through T4) will be greatly influenced by how quickly AMCs can develop and adopt standards and protocols for sample acquisition, storage, and annotation and their integration into the mainstream of patient care. More than 300 million human biospecimens were stored in freezers across AMCs in the United States in 2000, with an estimated 20 million additional specimens being accrued yearly[26]. The NCI estimated that it spends greater than $50 million yearly on banking samples from cancer patients as part of 125 funded research programs and projects.[27]^,15… The pharmaceutical industry is shifting to a clinical trial paradigm requiring that subjects provide samples with the hope of creating a new model for successful clinical development based on biomarkers derived from the analysis of these samples.

Despite this increased recognition of the role of human biospecimens as a critical enabler of genomics-based research and medical care, the state of storage of human biospecimens is largely in disarray.[28] Most AMCs cannot readily access a list of samples stored on institutional premises, the conditions under which they are stored or the subjects who donated them. The current lack of standards and quality control procedures for sample procurement to biological analyses presents a significant challenge to developing studies of statistical and clinical value as well as to guide public health planning and raises issues concerning the appropriate use of these samples donated by human subjects. Working with the NIH, AMCs have made progress in standardizing practice to facilitate knowledge sharing across institutions. In 2004, the NCI initiated the Cancer Bioinformatics Information Grid (caBIG) to standardize data formats for genomic and phenotypic data captured in cancer research and to develop common research tools among more than 50 NCI-designated cancer centers. Specific biomedical research tools under development by caBIG include clinical trial management systems, tissue banks and pathology tools, imaging tools, and a rich collection of integrative cancer research applications.[29]

Centralized biorepositories and standardized patient registries are aligned with the mission of AMCs and health systems to enable and enhance research opportunities as well as to assist in the structure to support health care delivery. Centralization will manage costs, create synergies and economies of scale, reduce liability, maintain high ethical standards, and enable compliance with applicable regulations. Research and funding opportunities will undoubtedly be enhanced through a centralized system that provides timely access to a large numbers of fully annotated samples, thereby minimizing the need to enroll new subjects and collect new specimens for each study. In addition, centralized biorepositories make costs more transparent and allow the AMC-investigator community to carry out its research and clinical mission more efficiently, rather than spend its time managing sample collections. Longitudinal cohort studies rich in epidemiologic data combined with biospecimen banking create unparalleled scientific power. As we discuss below, biospecimen banks are a not only a valuable source for discovery, but in cases where data has been collected over long periods of time, biobanks may allow for the efficient validation of biomarkers for their association with distant clinical endpoints that would be prohibitively expensive to validate prospectively.

Well-annotated biospecimens collections can also be leveraged successfully into academic-industry partnerships whose goal is improved diagnostics and therapeutics development. Merck & Co and Tampa's H. Lee Moffitt Cancer Center & Research Institute have formed a for-profit center, M2GEN, to collect tissues and clinical information of up to 30,000 consented research subjects with the aim of identifying biological differences that might explain variation in response to cancer drugs. The deal, valued at nearly $100M over five years, gives Merck exclusive access to the database for drug discovery purposes. In a second collaboration, this time with a for-profit company medical device company, Merck partnered with Fox Hollow Technologies, Inc. of Redwood, CA. The partnership provided Merck access to Fox Hollow’s collection of atherosclerotic plaques to test cardiovascular biomarkers for use as diagnostics and as tools for drug development. Similarly, BG Medicine – part of the High Risk Plaque (HRP) initiative, an industry consortium – is working with Duke University to identify biomarkers that identify patients at high risk for acute coronary syndromes using blood samples previously collected and stored by Duke’s cardiac catheterization laboratory. The samples are linked to health outcomes data longitudinally through patient care within the Duke University Health System. These examples underscore the fact that research into PHC, both in academia and industry, could be greatly enhanced by more ready access to annotated patient samples to validate and develop new biomarkers.

Biobanks at research consortia funded by the NCI have played a central role in the development of Genomic Health’s Oncotype Dx testing service to predict the risk of recurrence in early stage breast cancer patients. The initial list of candidate genes came from a search of the academic literature, mainly contributed by AMCs. Genomic Health refined its gene list and subsequently conducted two major validating clinical studies of the test entirely on tissues banked by the National Surgical Adjuvant Breast and Bowel Program (NSABP), a cooperative group based at the University of Pittsburgh. The two major cohorts used (B-14 and B-21) were collected in the 1980s. Without access to such tumor banks with “mature” clinical data, the T2 research necessary for clinical adoption would not have been possible in a timely or cost effective manner and investment and subsequent “translation” of the discovery would likely never have occurred. However, by carefully designed studies within the NSABP biobank cohorts, Genomic Health has been able to successfully launch Oncotype Dx and achieve reimbursement from most payers. Based in part on the data from these studies, the American Society of Clinical Oncologists (ASCO) has included Oncotype Dx in it most recent guidelines for the diagnosis and treatment of early stage breast cancer.

Resourcing the clinical validation of the next generation of personalized medicine (T2)

Demonstrating the clinical utility of most the newly discovered genomic or imaging biomarkers through appropriately powered, randomized clinical trials has proven difficult for academic researchers and industry alike. When asked why genomic discoveries are not advanced to practice, stakeholders in PHC cite the lack of both public and private funding for clinical studies to build an evidence base and the challenges of designing and executing studies in which the clinical endpoints are separated from the interventions by many years[30]. Without clear evidentiary standards, investors cannot be certain of the level of funding necessary to achieve regulatory approval and payor acceptance of a new biomarker. In the face of this uncertainty, clinical validation is often left unfunded by the private sector. Indeed, there is a dearth of studies addressing the impact of new personalized medicine tools. In a survey of PubMed articles published between 2001 and 2006 on genomics and genetics in humans, only 2% of 336,169 manuscripts were classified as clinical trials. Of these trials, few were randomized.⁴ Recognizing the need to develop studies that demonstrate the clinical value of genomics to inform clinical decision making and provide value, the Centers for Disease Control (CDC) and the NIH announced at least three RFAs this year to foster these types of studies.[31] The private sector has not been an enthusiastic funder of T2 research in personalized medicine. This is in stark contrast to clinical evidence produced each year funded by private industry to support the introduction of new therapeutics regulated by the FDA under the Pre-Market Approval (PMA) process for drugs, biologics, and devices. However, recently diagnostic development companies and the pharmaceutical industry have begun to, under certain scenarios, invest in personalized medicine and the T2 research studies necessary to drive their clinical adoption and prove their clinical utility.

The Government as sponsor of T2 research. CDC’s Evaluation of Genomic Applications in Practice and Prevention (EGAPP) working group has begun the process of culling from the literature the genetic and genomic tests that have promise to shift the way health care is delivered. The first EGAPP report on the use of pharmacogenetic testing for prescribing tricyclic antidepressants was released in December 2007.[32] One of the areas EGAPP has identified for study is the use of gene expression profiles for prognosis in breast cancer – an area with a clear demand for a novel diagnostic solution. Of the women that receive adjuvant chemotherapy for node negative, estrogen receptor positive breast cancer, approximately 85% receive no clinical benefit over taking tamoxifen alone[33]. Despite the lack of a prospective randomized clinical trial – the gold standard for proving the value of an experimental therapy – oncologists used RNA expression signatures for risk stratification and prognosis in breast cancer for more than 24,000 “treat” vs. “no-treat” decisions in 2007[34]. A prospective cooperative group clinical trial (MINDACT) by the European Organization for Research and Treatment of Cancer aims to measure the effectiveness of a gene expression predictor of breast cancer prognosis to guiding adjuvant chemotherapy when compared to predictions based solely on the traditional clinical parameters for prognoses[35]. An NCI-sponsored study (TailoRX) aims to utilize the Oncotype Dx test to identify low risk breast cancer patients unlikely to benefit from chemotherapy. A similar opportunity now exists to refine prognosis and redirect treatment in early stage lung cancer[36] and a CALGB sponsored clinical trial has been developed to use an expression signature to randomize patients to surgical treatment with or without adjuvant chemotherapy[37]. These are clear examples of T2 research in which AMCs in collaboration with government and industry are developing novel clinical trials infrastructures to evaluate the performance of genomic medicine tools to redefine disease phenotypes and refine therapeutic strategies.

Diagnostic companies as sponsors of T2 research. In previous decades, private diagnostic companies have been reluctant to sponsor or conduct extensive clinical trials to demonstrate the clinical utility of novel assays, genomic or otherwise. This reluctance to invest has been driven primarily by economics. Under the current payor system a diagnostics company is reimbursed fixed fees for any procedures necessary to perform a test. Typically, these fees do not provide sufficient excess margin to justify an investment in extensive clinical validation, let alone patient and physician education or clinical guidelines development. Moreover, reimbursement by insurance companies has not generally been contingent on proving clinical utility in formal trials. Instead, tests had to be deemed “non-investigational”. As a result, most diagnostics on the market today have arrived after floundering in “investigational use” status as evidence and awareness slowly build up over time. Typically, a diagnostic company will develop a commercial version of a new test only once the biomarker has been sufficiently validated and gained acceptance within the clinical community. As exemplified by troponin testing for cardiovascular injury in the setting of chest pain[38], the AMC has traditionally filled this validation role, often performing investigator-initiated trials and conducting the testing using their own, low-volume laboratory developed test (LDT), prior to the availability of a commercial test. This reluctance to invest in the validation of new diagnostics is often amplified when the performance characteristics of the technologies are less established – as is the case with many of the new genomic and multi-analyte platforms – and when the pathway to regulatory approval and ultimate clinical acceptance is less clear.

There is evidence, however, of a new model emerging for the investment in the development of novel personalized health diagnostics. The tests that receive such investment are often linked to expensive therapeutics and so can carry a high economic value. Under this new paradigm, a few private diagnostic companies, including Genomic Health, XDx, and Third Wave Technologies, have made the decision to invest in clinical trials conducted at AMCs, as well as the physician education and clinical guidelines development necessary to bring a novel test into widespread clinical use. This new model requires IP protection for the test, a clear path to a large market, and justification for a value-based price, which circumvents the traditional code-based reimbursement scheme. Only under these conditions can a private company be assured of an appropriate return on their investment.

A recent example of this new approach is the development of the Oncotype Dx testing service to predict the risk of recurrence in early stage breast cancer patients. Genomic Health invested over $100 million in the clinical development and marketing of the test. However, with a price point of $3,460 and an operating margin of over 60%, Genomic Health has a good chance of recouping its investment in the coming years. Genomic Health can justify its relatively high price for Oncotype Dx based on the potential value it brings to patients and their payers. By identifying those patients unlikely to experience a recurrence of their cancer and therefore unlikely to receive any benefit from adjuvant chemotherapy, the test can in theory reduce the amount of money spent on chemotherapy and the management of its complications.

Genomic Health was able to identify an application in which the potential to save healthcare resources was high compared with the cost to demonstrate the clinical utility of the test and engage the patient and physician communities. Also, by securing patent protection for their test, they have been able to limit direct competition. However, very few new personalized health applications will have such attractive economics. Many other genomic discoveries have the potential to have a positive impact on healthcare delivery, but lack a clear path to near-term commercial profitability. The uncertainty surrounding what will be required for clinical validation and to secure approval by regulators and payors, and the lack of clarity in existing patent law to ensure exclusivity in the market discourage investment in all but clear economic winners. Until significant policy changes are implemented to reduce the uncertainty in validation requirements, level of and time to reimbursement, and ability to practice both freely and exclusively with regard to intellectual property, private investment will likely be limited.

Pharmaceutical companies as sponsors of T2 research. The pharmaceutical industry has the potential to be a significant driver of personalized medicine using genomic information to inform drug development, approval, and clinical drug use. At the same time, pharmaceutical firms have long resisted stratification strategies in clinical development and the resulting ‘segmentation’ of markets. For the most part, pharmaceutical developers are utilizing genomic approaches to identify which populations benefit from drugs after they are approved. Drug manufacturers would be wise to undertake such studies prior to approval. The lessons of cetuximab and EGFR mutations - driven by AMC investigator initiated studies to better understand the populations most likely to benefit from these agents - and recent late-stage drug failures have sounded an alarm. Indeed the FDA’s Critical Path Initiative challenges industry to adopt the use of biomarkers throughout drug development[39]. Voluntary Genomic Data Submissions to the FDA that began in 2005 encourage sponsors to incorporate genomics into their development plans[40] heralding that this may be a requirement in the future. The recent addition of genetic testing to the FDA label for warfarin[41] and the recent FDA approval of a microarray based test for the management of breast cancer[42] as well as a test for tumor of unknown primary[43] are clear signals that the regulatory environment will increasingly encourage medical product development based on genomic information. According to a recent survey by McKinsey and Co., biomarker R&D expenditures within

pharmaceutical firms in 2009 were estimated at $5.3 billion, up from $2.2 billion in 2003. This increase is targeted at the development of safety and pharmacokinetic biomarkers, and in so-called “companion diagnostics” – biomarkers that can accurately identify individuals with a high likelihood of response. Since most drugs show activity in only a fraction of patients, an industry-based strategy to use genomics to identify subgroups of patients most likely to benefit from their products in development will bring more personalized therapies to the market and will incorporate genomic testing into the labeling of the drugs ultimately approved.

AMC-initiated T2 research studies

AMC investigators are now designing studies to test the hypothesis that genomics can improve outcomes for existing and standard of care therapies. At the Duke Institute of Genome Sciences & Policy this has been adopted as a strategy for translating genomics into clinical medicine. The IGSP Clinical Genomics Studies Unit (CGSU) has been established with the goal of setting the standard for genome-based clinical trials (www.genomestohealth.org). This unit functions to vet the scientific merit of trials prospectively testing predictive genomic tests, assess technical and practical feasibility, and developing outcomes data to support clinical utility and cost effectiveness. A typical trial design that tests the ability of genetic or genomic information to improve clinical and economic outcomes underway in the CGSU is shown in figure 2 below.

Figure 2. Design of a clinical trial to test the utility of a molecular test to impact standard of care therapy decisions.

Figure 2. Design of a clinical trial

Conflict of Interest in T2 research at AMCs

AMCs often face the vexing issues of conflict of interest that come with their role as neutral arbiters of the evidence surrounding use of medical technologies, both their benefits and their risks. “Opinion leaders” who influence the introduction and adoption of drugs, vaccines, biologics, and devices are typically drawn from prestigious AMCs. Congress is clearly concerned that the flow of money and other incentives for collaboration between academe and industry can also bias the research system in favor of corporate interests. The trade associations for the pharmaceutical, biotechnology and device industries have agreed to a succession of voluntary codes of conduct.[44] The Association of American Medical Colleges has issued several reports that make recommendations for managing both individual and institution conflicts of interest.[45] As the world’s largest single medical research laboratory, the NIH, tightened restrictions on its federal employee researchers in 2002.[46] NIH also reminded its grantees and contractors of the need to have conflict-of-interest policies and its right to audit implementation of such policies in August 2008.[47] The government is also engaged in formal rule-making that could alter the rules. Several states have passed laws limiting gifts to physicians or mandating reporting of gifts over a certain amount (usually $25 or $50); Senators Grassley and Kohl have proposed a federal law mandating reporting of gifts and payments. Conflict of interest was a feature of the national magazine for state legislatures in September 2008.[48] Pennsylvania has funded a counter-detailing initiative to guide use of drugs, and many states have considered bills about direct-to-consumer advertising of medical products. Most of these proposed policy changes are primarily directed at drugs, but the policies are likely to spill over to change the overall system for introducing and adopting all new medical products and services, including genomic technologies.

AMCs as platforms to study implementation of PHC delivery (T3) and outcomes (T4)

The development and validation of clinical delivery models that support PHC is critical to its implementation and adoption. AMCs, for their part, have an opportunity to fundamentally change their approach to physician education, payment and incentive systems, and metrics of quality and efficiency and act as the first-line testing grounds for innovative T3 research. Moreover, by providing a platform with resident expertise in both clinical research and care delivery, AMCs have the opportunity to provide a common platform to all of the stakeholders for the conduct of the implantation, dissemination and health outcomes research necessary to see PHC brought into practice.

Although clinical care is a core mission of AMCs, as Snyderman and Yoedionio have suggested, academic medicine has not yet become engaged in the systematic exploration of more rational models for health care delivery required for personalized and prospective medicine.³ Only a handful of AMCs have developed comprehensive programs enabling prospective approaches to patient care. In 2003, for instance, Duke University initiated Duke Prospective Health (DPH), a personalized care, disease management, and wellness program for its employees. The program, which Duke University physicians helped develop and manage, sought to prevent or detect chronic conditions related to smoking, diet, exercise, and stress by having patients develop and use a Personal Health Plan to ameliorate their individual risk. The program has three main components: a Health Risk Assessment (HRA), Care Management, and Health Coaching:

The HRA is a tool that analyzes lifestyle and habits and helps patients and their providers identify current and potential health issues necessitating attention. Patients use the results of their HRA to develop long-term strategic goals focused on health and wellness.
Care Management is where a care manager serves as the patient’s point of contact and works with the patient to help formulate a Personal Health Plan that meets his or her health needs.

Health Coaching allows patients to work with a coach in a group setting who assists in facilitating the patients to achieve the goals of their Personal Health Plan.

Although the program is relatively new, preliminary analysis on 154 patients suggest that a multi-modality intervention reduced risk of CHD, by increasing exercise and improving weight loss.[49] Duke is now initiating comprehensive PHC programs that use the DPH as a core delivery model in breast cancer, prostate cancer, diabetes, cardiovascular medicine, pharmacogenomics, and family history.

Developing new economic models

The premise of PHC is that by addressing health concerns pre-symptomatically – while interventions are more impactful and cost effective – health systems can improve health and lower the costs of health care. However, under the current economic models, any cost savings may not be realized by the health care providers bold enough to institute these changes. Currently, most payor systems do not reimburse for preventive services, except when Congress explicitly mandates it. Instead, reimbursement in the modern American health care system is driven by procedures and post-symptomatic interventions. Moreover, intensive in-patient procedures typically yield higher margins to the health systems than out-patient health monitoring and non-surgical interventions. PHC models, if successful, would shift patients from high-margin in-patient procedures to low-margin (or, at present, uncovered) out-patient screening and interventions. From the perspective of the healthcare systems’ finance department, PHC is a money-losing proposition. With such misaligned incentives, personalized medicine approaches may not receive as enthusiastic backing as if it were equally profitable as current procedures, and therefore incentives to innovate in this area of health care delivery are lacking.

Similar countervailing financial incentives combined with an overall lack of compelling clinical data on new personalized health tools make it difficult for payors to fully embrace PHC. It may seem financially prudent for a payor to reimburse for a diagnostic test that could identify high-risk individuals in situations where relatively low-cost interventions could prevent expensive surgical procedures in the future. However, identifying those tools can be difficult. We have already examined the relative lack of clinical data on such tools on which payers might be able to make that determination. This uncertainty is compounded by the fact that even if a molecular diagnostic is shown to work, there is no guarantee that healthcare providers and/or their patient will modify their behavior in response to the result – in which case the payer may end up paying for the test and the surgical intervention. Finally, there is the “hazard” of discontinuity of coverage; due to the fact that people shift health coverage plans over their lifetime, a payor that covers a diagnostic screening for an individual will not necessarily receive the benefits of a healthier client in the coming decades. In fact there is an incentive for people to “game the system” by enrolling in relatively expensive plans which cover PHC, then once testing is complete, shift to a relatively less expensive plan for their long-term care. There are a few examples of situations where PHC has been covered by insurance. For example, Aetna and Kaiser will cover genetic counseling services under many of

their plans. Aetna has even instituted a phone counseling service for its members.[50] This may be financially motivated or for reasons of increasing service levels in a competitive health insurance market.

While all parties ultimately stand to gain from the implementation of PHC, economic incentives present significant barriers to realizing its implementation. It is incumbent upon AMCs to demonstrate leadership in the clinical delivery space by exploring new economic models, and serving as a common forum in which all stakeholders might share data and resources to overcome these barriers and work towards a scenario in which all parties benefit.

Quality, Performance, and PHC

Personalized medicine will continue to meet resistance from individual practitioners unwilling to modify their patient management approach. Clinicians may resist if they feel that their judgment is being superseded by a test result or if they feel the way they have managed patients in the past was adequate. Without the proper systematic incentives in place, adherence to clinical guidelines and adoption of new therapeutics is often lackluster. For example, despite the publication of clinical evidence demonstrating the clinical utility of the use of beta-blockers for patients who were recovering from a myocardial infarction (MI) in 1981, in 1996 – 15 years after the landmark publications– these drugs were only prescribed to 62.5% of patients after an MI[51]. However, once physicians were evaluated based on their adherence to clinical practice guidelines, adoption increased rapidly. The National Committee for Quality Assurance (NCQA) began tracking compliance with certain treatment guidelines, including the use of beta-blockers in MI patients, in 1996 and publically reporting results. By 2006, compliance of beta-blocker administration had improved to over 97%.⁵⁶ The tendency to resist a change in practice holds true for all clinical care models, but will be especially true in the case of PHC. Adoption of new genomics-based tools will require health care providers to become familiar with new technologies and science and require continuing education on awareness on new PHC methods. AMC must make systemic changes in how health care providers are evaluated, compensated and trained if PHC is to be readily implemented and tie the concepts of PHC to quality, safety, and performance.

Health Professional and Public Education

A core mission of the AMC is to train the healthcare workforce. As PHC services and diagnostic tools evolve, AMCs will need to develop training for the workforce that will be required to implement them. In our opinion, this is fundamental to bridging the third translational gap, T3, in the translational continuum. In this regard there is need to break new ground in medical education and to develop a national model for integrating knowledge of new molecular-based technologies in medical practice. The primary care workforce feels woefully unprepared to integrate genomics into regular practice[52]. Consumers are enthusiastic about genetics and are hopeful about their impact but at the

same time they have a low knowledge base about genetics and genetic testing for common diseases. Education of health professionals and the public must be a priority to advance the use of genomics into healthcare. With the rapid advances in genomics research and developing technologies, it will be challenging to keep health professionals informed about the benefits, risks, and limitations of new tools as they become available. In addition, the public and health care workforce will need to understand the appropriate clinical applications of genomic tools -- including their benefits, risks and limitations, and how they may improve clinical management. Direct to consumer genomic testing has only served to greatly intensify the educational needs across the genomic medicine community from the lay public to health care providers to policy makers. Several surveys have documented the below average physician knowledge of genetics[53], but none has assessed knowledge of the newer field of genomics. The importance of education in the application of pharmacogenetics has been described[54], but at present there are no broad initiatives to orchestrate genetics and genomics education of medical professionals, trainees, and the public at large. Basic genomic literacy is a critical need for patients and physicians and communities to engage in genomic research and clinical studies required bring about a change in the care paradigms to support clinical genomics applications.

In 2003, Duke University School of Medicine it revamped its curriculum and set as a goal to “practice personalized health planning for long-range goals”. Although this has not yet happened on a school wide scale, fourth-year electives such as Integrative Medicine and Prospective Health and Health Promotion and Disease Prevention are available. The fundamental concepts underlying the theory and practice of PHC should become central parts of medical education.3 For the basic sciences, this would include teaching concepts of disease evolution from health and the role of predictive biomarkers in this process. Clinical education would include concepts of a medical evaluation comprising health risk prediction, current health status, pathogenesis tracking, pharmacogenetics, health planning, patient motivation, and disease management. To our knowledge, this has not yet occurred to any significant degree among U.S. medical schools or residency training programs. A strategy driven by the AMC community is essential to effect changes in the way the health care providers in medical and nursing students in the USA are trained in PHC.

Patient Centered Medical Home (PCMH)

An important and emerging concept that will facilitate the implementation of PHC is the PCMH or the American College of Physician’s “Advanced Medical Home”. Enabling both T3 and T4 research, the PCMH is a model for delivery of care that is provided by medical practices to strengthen the physician-patient relationship by focusing on delivering coordinated care in a prospective manner – similar to the Duke Prospective Healthcare model above - to patients, focusing more on prevention of disease and promotion of wellness, compared to the current system of focus on episodic care based on illnesses and patient complaints. Genomic medicine should be integrated into each of the principles of the PCMH (Table 2): Personal physician, Physician directed medical

practice, Whole person orientation, Coordination/Integration of Medical Care, Quality and safety, and Access.

The PCMH is an opportunity to improve clinical operations and outcomes for patients in a currently fragmented medical system. The concept of genomic and personalized medicine has synergy with the PCMH in the effort to better define risk of chronic disease for individual patients, and to redefine how health risk is communicated to patients. Ongoing strategies for delivery of genomic or imaging based risk assessment technologies that enable PHC should focus on integration with the PCMH.

Table 2: The Advanced or Patient Centered Medical Home

Health information technology (HIT) and electronic health records (EHRs)

Information technology will be a key component of both health care delivery and the T3 and T4 research that is needed for PHC. Past experience indicates that the new genomic interventions, like any new medical intervention, will remain significantly underutilized for some time without the concurrent introduction of supportive technologies. Moreover, genomic interventions may face even greater barriers to clinical adoption compared to more traditional medical interventions, due to such factors as patient concerns over genetic discrimination, limited clinician familiarity with the science, and the volume and complexity of the data that need to be considered. In recognizing this challenge, Secretary Leavitt announced in March 2007 HIT as a priority to support the achievement of personalized medicine.

The use of electronic medical records (EMRs) as a major component of HIT is expected to substantially improve the quality and efficiency of health care and provide an important vehicle to advance patient-centered personalized care. The use of EMRs in care delivery is expanding rapidly, especially among large integrated health delivery systems. The amount of clinically relevant molecular data and the number of resources devoted to research on genomic medicine are increasing in parallel. While the U.S. health system is fragmented, the large health systems that are adopting EMRs are becoming increasingly integrated, especially in adopting and implementing practice standards. Thus, a significant opportunity exists to incorporate various aspects of genetics, genomics, and predictive tools into the development of these emerging systems to facilitate adoption and clinical decision making.

An integral component of PHC is the application of family history to clinical care. Interest in collecting family history data as a routine part of care delivery is growing, as knowledge advances in linking family history of disease to patient risk. The need for the development of better family history tools has been highlighted by projects at the Centers for Disease Control (CDC) and by the U.S. Surgeon General’s Family History Initiative[55]. However, these efforts have not directly addressed the integration of tools into the real-world scenario of busy physicians and a multiplicity of health record systems, and do not provide an adequate breadth of data capture necessary for research. The need for new tools is apparent; however, no such electronic family history tools have yet been developed, despite the availability of suitable technologies.

An important component of HHS’s vision for the role of HIT in PHC is the use of computer systems to provide clinical decision support (CDS), defined as the act of providing clinicians, patients and other health care stakeholders with pertinent knowledge and/or person-specific information, intelligently filtered or presented at appropriate times, to enhance health and health care. The HHS report entitled Realizing the Promise of Pharmacogenomics: Opportunities and Challenges, the HHS Secretary’s Advisory Committee on Genetics, Health, and Society identified the need for CDS tools as an important to realizing personalized pharmacotherapy.

CDS has been leveraged for many decades now to improve clinical decision making related to traditional medical interventions and when compared to other approaches to improve practice, CDS has generally been shown to be more effective and more likely to result in lasting improvements in clinical practice. However, an interim report from an ongoing RAND study indicates that none of the commercial electronic health record systems currently provide CDS to support genomic medicine. Thus, it appears that the CDS might be an important aspect of the delivery of information for clinical decision-making; however there has been little research or investment in CDS to optimally deliver information to healthcare providers to support practice of PHC.

The nuances in clinical decision-making in PHC already render many care scenarios complex. Access to an EMR and the ongoing codification of medical knowledge (i.e., Clinical Practice Guidelines or CPGs) will be essential to addressing this growing translational gap. CPGs greatly facilitate, but are not sufficient for translating knowledge to practice. Almost 2000 active CPGs exist in the US National Guideline Clearinghouse[56] and an individual CPG may encompass dozens to hundreds of clinical recommendations. These recommendations will rapidly expand in the era of genomic and personalized medicine, and thus codification of knowledge will be essential to increasing its access. EMRs offer a platform to translate codified knowledge into real-time actionable processes. Genomic data will need to be accessed with other patient data located in disparate locations within the EMR and evaluated in relation to a rule set. Real time actionable recommendations and CDS will need to be created and supported by an integrated and intuitive visual display of information.

The EMR also offers an exciting opportunity for population and health outcomes (T4) research. It represents an economically efficient means of obtaining phenotypic data and biosamples for generating genotypic data and for validating discovery data as well as assessing public health impact of these discoveries on long term outcomes. The EMR thus represents a potentially large increase in efficiency for obtaining phenotypic data and can also be an extremely efficient tool for patient recruitment and biosample acquisition. The data federation initiated by the HMO Research Network (HMO RN) offers an example of a venue for the type of outcomes research needed to provide evidence that a PHC strategy will provide value. The network is a consortium of 15 research centers, each affiliated with a non-profit integrated health care delivery system, all of which have or are developing ambulatory care EMR systems. In addition to the development of best practices for research administration for multi-site collaborations, the HMO RN has initiated efforts to establish a Virtual Data Warehouse (VDW), to simplify data sharing among network participants. We encourage partnerships between AMCs and networks such as the HMO RN and mechanisms to fund them such that these data can be obtained expeditiously.

Payors as integrated partners in T3 and T4 research.

As indicated above, by sharing data and coordinating their efforts it may be possible for AMCs/health systems, payors, and diagnostics companies to study the penetration, dissemination and implementation of new personalized health tools and their effect on

health outcomes[57]. This, of course, would even more powerful if common standards of reporting from EHRs were possible across health systems. The NIH’s Clinical and Translational Science Awards program that seeks to fund 60 centers of translational research as a consortium by 2011 may provide the foundation of infrastructure and standards required to begin to address these issues across AMCs as has been done by the HMO RN. This may be an opportunity for any national agenda for PHC to leverage the investment and emerging architectures in that program that span the breadth from the laboratory to the community. With open access to data, scholars and policy makers could determine the factors that affect clinical uptake and the resulting economic and health impact. It would be difficult for any of these parties to make these determinations independently in a reasonable timeframe.

Early examples of these novel partnerships are beginning to emerge. For example, the Mayo Clinic has partnered with Medco to evaluate test results from over 1,000 patients taking Warfarin. In another example, Kaiser Permanente of California partnered with Genomic Health and USC’s Keck Medical School to underwrite a study of the clinical utility of Oncotype DX within the Kaiser Permanente coverage population[58]. In each case, the payor has been willing to sponsor additional clinical research when prior published research indicated both clinical validity and a potential to save costs and the test was already commercially available to test. However, these opportunities are relatively rare as few diagnostic programs have the resources or long history of use to provide the preliminary support.

Figure 3: Specialized Centers for Genomic and Personalized Medicine.

A Call for Specialized Centers as Catalysts Accelerating Genome Information to Medicine

AMCs, while well positioned to discover and develop new tools, lack the resources, infrastructure, and skills to bring new personalized health discoveries into the market place and ultimately into clinical environment. By contrast, diagnostic companies typically have the infrastructure to make tests widely available: high-volume regulatory-compliant labs, sample collection and tracking, regulatory expertise, relationships with payors, marketing and physician education capabilities, but often lack the resources to mount an effective research and development effort to create the “content” for new diagnostic tests. Through intellectual property licensing and sponsored research agreements, academia and industry have shown that they can form synergistic partnerships to advance personalized medicine. However, even with their combined skills and resources, it has proven extremely difficult to navigate a personalized medicine program through the entire “translational continuum”. At the same time, payors are motivated to see effective models of PHC implemented and have the infrastructure and access to longitudinal data to contribute to important research on diffusion and community health impact.

Specialized centers for genomic and personalized medicine in AMCs – perhaps modeled programmatically after the Centers for Excellence in Women’s Health Program at the NIH – can be instrumental in integrating, facilitating and catalyzing the needs of government, academic and industry stakeholders by providing:

access to patients, patient data, and molecular and biological data that drive the development and exploration of genomic information and its link to clinical outcomes
the scientific foundation for novel biomarker discovery for both disease and drug response based on mechanism
an environment for innovation that fuels the development of novel translational strategies
a vehicle for aligning the efficiency and quality metrics of patient care with the goals of personalized medicine
a network for defining, validating and implementing common data standards to facilitate knowledge sharing and accelerate discovery, validation, and monitoring of new PHC tools,
the infrastructure for the types of public-private partnerships required for executing genomic assay guided clinical trials, and finally
a place to engage in a dialogue and research on the key issues challenging the translation of genomics into PHC: education, facilitating clinical trials, regulatory policies, information systems, research on dissemination, and integration into practice.

Currently there are no structured programs in genomic and personalized medicine. Several institutions have made the commitment (Duke University, Vanderbilt, Harvard, Johns Hopkins, University of Utah, Ohio State University) but none has done so with federal support. Moreover, the tasks required, we would argue, are larger than that any single AMC can tackle. To bring about the transformation in health care the genome has promised will require assembling diverse stakeholders focused on the application and translation of genomics with a goal of improving the health of individuals and driving efficiency in health care. These centers will thrive on their interdiscipinarity. Specialized centers housing basic genome science laboratories, clinical researchers, informaticians, clinicians, economists, health policy makers and in partnership with industry (pharmaceutical and diagnostic companies), and with health systems that will enable the scientific output of the genome to cross the chasm between bench and bedside. A series of Centers that focus on specific aspects of the challenges that learn and participate with one another would, in our opinion, be a major step forward in developing and enabling the continuum of strategies required for the fullest impact of genomic and other relevant information on PHC.

[1] Willard HF, Angrist M, Ginsburg GS. Genomic medicine: genetic variation and its impact on the future of health care. Philos Trans R Soc Lond B Biol Sci. 2005 Aug 29;360(1460):1543-50.

[2] (http://www.fda.gov/cder/genomics/genomic_biomarkers_table.htm). Accessed September 1, 2008.

[3] Snyderman R, Yoediono Z. Perspective: Prospective health care and the role of academic medicine: lead, follow, or get out of the way. Acad Med. 2008 Aug;83(8):707-14.

[4] Koury M, Gwinn M, Yoon P, et al, “The continuum of translation research in genomic medicine: how can we accelerate the appropriate integration of human genome discoveries into health care and disease prevention?” Gen in Med 2007; 9:665-674.

[5] Genetests.org (www.genetests.org), accessed 31 August 2008. GeneTests.org is an online service developed by Roberta Pagon of the University of Washington, Seattle, that mushroomed from an initial effort to catalog the availability of genetic tests in the late 1990s and has become a major online resource. Described in RA Pagon. GeneTests: an online genetic information resource for health care providers. Journal of the Medical Library Association : JMLA. 2006 Jul;94(3):343-8 (http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1525324).

[6] Golub TR et al. Molecular Classification of the Cancer: Class Discovery and Class Prediction by Gene Expression Monitoring. Science. 1999;286:531 – 37.

[7] Alizadeh AA et al. Distinct Type of Diffuse Large B-Cell Lymphoma Identified by Gene Expression Profiling. Nature. 2000;403:503 – 11.

[8] Ramaswamy S and Golub TR. “DNA Microarrays in Clinical Oncology.” J. Clinical. Oncol. 2002; 20:1932 – 41.

[9] Staudt LM. Molecular Diagnosis of the Hematologic Cancers. N. Engl. J. Med. 2003; 348: 1777 – 85.

[10] Pohlhaus & Cook-Deegan, Genomics Research: World Survey of Public Funding, submitted.

[11] This estimate derives from two recent surveys of companies and of government and nonprofit funding courses. The surveys estimate $2 billion annually from the largest genomics firms and another $1 billion from pharmaceutical and established biotechnology firms

[12] Chandrasekharan, Perin, Wiechers and Cook-Deegan, ch 37 of “genomic medicine” 2-vol set, forthcoming. This 2004 study found 470 firms in 25 countries whose businesses included some aspect of genomics, including 88 firms with publicly traded stock. The largest 15 such public firms spent over $2 billion in 2004 on R&D Established biotechnology and pharmaceutical firms were not included in the survey, but this estimate assumes they continue to expend at least $1 billion annually, in line with the 2000 survey in which they were included.

[13] Carmie Chan, Amber Johnson, David Kauffman, and Robert Cook-Deegan, World Survey of Funding for Genomics Research, Stanford-in-Washington program, funded by Burroughs Wellcome Fund, 1999-2000. Unpublished data.

[14] Fossum D, Painter LS, Eiseman E, Ettedgui E, and Adamson DM. Vital Assets: Federal Investment in Research and Development at the Nation’s Universities and Colleges. Science and Technology Policy Institute, RAND Corporation. 2004. (http://www.rand.org/pubs/monograph_reports/2005/MR1824.pdf).

[15] A report from the US Patent and Trademark noted “For the most recent calendar years, the percent share of utility patents issued to a U.S. college, university, or association of U.S. academic institutions has held steady at just under 2.0% (see the table in Section 1B for percentage shares between 1992 and 2005).” U.S. Patent And Trademark Office, Electronic Information Products Division, Patent Technology Monitoring Team (PTMT), U.S. Colleges And Universities-Utility Patent Grants 1969-2005, a report funded by the National Science Foundation. Available at http://www.uspto.gov/go/taf/univ/doc/doc_info_2005.htm and http://www.uspto.gov/go/taf/univ/univ_toc.htm (accessed 31 August 2008).

[16] Cook-Deegan & McCormack, Science 1998.

[17] Rydholm C, Chandrasekharan S, and Cook-Deegan R. “The impact of University Licensing Practices and DNA-based patents on Genetic Testing access.” Poster presented at Changing Horizons, Association of University Technology Managers, Annual Meeting, San Diego, California, USA, February 28- March 1 2008.

[18] Stokes DE. Pasteur’s Quadrant: Basic Science and Technological Innovation. Washington, D.C. : Brookings Institution Press, 1997.

[19] The Bayh-Dole Act was passed in December 1980 and took effect the following year. It made practices among US agencies funding research more uniform and established default rules favoring ownership of patent rights by grantee and contractor institutions. Starting in the late 1960s, many academic institutions operated under Institutional Patent Agreements that gave them title to resulting inventions. The Bayh-Dole Act extended this to most federally funded research and codified the administrative practice in statute, thus increasing consistency and reducing transaction costs of negotiating agreements institution-by-institution. A dispassionate view of the role of Bayh-Dole is found in Mowery, et al., Ivory Tower and Industrial Innovation: University-Industry Technology Transfer before and after the Bayh-Dole Act, Stanford, CA: Stanford Business Books, 2004.

[20] Lenoir T and Giannella E. The Emergence and Diffusion of DNA Microarray Technology. Journal of Biomedical Discovery and Collaboration. 2006; 1:11.

[21] Parsons DB. Seminal Genomic Technologies: Illumina, Inc. & High-Throughput SNP Genotyping Bead Array Technology: A Case Study. MS diss., Duke University.

[22] Topol EJ, Murray SS, Frazer KA. The genomics gold rush. JAMA. 2007; Jul 11;298(2):218-221.

[23] Grant SF, Thorleifsson G, Reynisdottir I, et. al. Variant of transcription factor 7-like 2 (TCF7L2) gene confers risk of type 2 diabetes. Nat Genet. 2006;38(3):320-323.

[24] Helgadottir A, Manolescu A, Thorleifsson G, et. al. The gene encoding 5-lipoxygenase activating protein confers risk of myocardial infarction and stroke. Nat Genet. 2004;36(3):233-239.

[25] Stefansson H, Sigurdsson E, Steinthorsdottir V, et. al. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet. 2002;71(4):877-892.

[26] Eiseman E, Bloom G, Brower J, Clancy N, Olmsted S. Case studies of existing human tissue repositories: “Best practices” for a biospecimen resource for the genomic and proteomic era. 2003. Santa Monica, CA: RAND Corporation.

[27] Eiseman E, Haga S. A handbook of human tissue sources: A national resource of human tissue samples. 2000. Santa Monica, CA: RAND Corporation

[28] Ginsburg GS, Burke TW, Febbo P. Centralized biorepositories for genetic and genomic research. JAMA. 2008 Mar 19;299(11):1359-61.

[29] (http://cabig.cancer.gov/media/backgrounder.asp) Accessed October 16, 2008.

[30] Deverka PA, Doksum P, Carlson RJ, Integrating molecular medicine into the US health-care system: opportunities, barriers and policy changes, Clin Pharmacol Ther, 2007;82:427-34.

[31] http://www.cdc.gov/od/pgo/funding/GD08-001.htm) Acessed September 14, 2008, (http://grants.nih.gov/grants/guide/rfa-files/RFA-DK-08-004.html) Accessed September 14, 2008.

[32] Berg et. al. Recommendations from the EGAPP Working Group: testing for cytochrome P450 polymorphisms in adults with nonpsychotic depression treated with selective serotonin reuptake inhibitors. Genet Med 2007;9(12):819–825.

[33] Paik S, Shak S. et. al. A multigene assay to predict recurrence of tamoxifen-treated, node-negative breast cancer

N Engl J Med. 2004 Dec 30;351(27):2817-26

[34] Genomic Health Announces Fourth Quarter and Year-End 2007 Financial Results and Business Progress and 2008 Financial Guidance. Redwood City, CA: Genomic Health, Inc., 2008 Feb 5.

[35] Cardoso F, Van't Veer L, et al. Clinical Application of the 70-Gene Profile: The MINDACT Trial. J Clin Oncol, 2007 Feb 10;26(5):729-35

[36] (http://www.cancer.gov/clinicaltrials/digestpage/TAILORx) Accessed September 14, 2008.

[37] Potti A, Mukherjee S, Petersen R, Dressman HK, Bild A, Koontz J, Kratzke R, Watson MA, Kelley M, Ginsburg GS, West M, Harpole DH Jr, Nevins JR. A genomic strategy to refine prognosis in early-stage non-small-cell lung cancer. N Engl J Med. 2006 Aug 10;355(6):570-80.

[38] Newby LK, Goldmann BU, Ohman EM. Troponin: an important prognostic marker and risk-stratification tool in non-ST-segment elevation acute coronary syndromes. J Am Coll Cardiol. 2003 Feb 19;41(4 Suppl S):31S-36S

[39] (http://www.fda.gov/oc/initiatives/criticalpath) Accessed September 14, 2008.

[40] (http://www.fda.gov/cder/genomics/VGDS.htm) Accessed September 14, 2008.

[41] (http://www.fda.gov/bbs/topics/NEWS/2007/NEW01684.html) Accessed September 14, 2008.

[42] (http://www.fda.gov/bbs/topics/NEWS/2007/NEW01555.html) Accessed October 15, 2008.

[43] (http://www.fda.gov/bbs/topics/NEWS/2008/NEW01870.html) Accessed October 15, 2008.

[44] Pharmaceutical Research and Manufacturers of America (PhRMA). Code on Interactions with Healthcare Professionals. Accessed 9 September 2008 at http://www.phrma.org/files/PhRMA%20Marketing%20Code%202008.pdf. Biotechnology Industry Organization. “Protecting Research and Research Participants.” Accessed 9 September 2008 at http://www.bio.org/healthcare/clinical/protect010303.asp. Medical Device Manufacturers Association. “MDMA Guideliens for Interactions with Customers.” Accessed 9 September 2008 at http://www.medicaldevices.org/public/issues/documents/MDMAGuidelines.pdf.

[45] Association of American Medical Colleges. Industry Funding of Medical Education: Report of an AAMC Task Force. June 2008. Accessed 9 September 2008 at https://services.aamc.org/Publications/showfile.cfm?file=version114.pdf&prd_id=232. See also Association of American Medical Colleges. Task Force on Financial Conflicts of Interest in Clinical Research. Protecting Subjects, Preserving Trust, Promoting Progress II: Principles and Recommendations for Oversight of an Institution’s Financial Interests in Human Subjects Research. October 2002. Accessed 9 September 2008 at http://www.aamc.org/research/coi/2002coireport.pdf. Association of American Medical Colleges. Protecting Subjects, Preserving Trust, Promoting Progress – Policy and Guidelines for the Oversight of Individual Financial Interests in Human Subjects Research. December 2001. Accessed 9 September 2008 at http://www.aamc.org/research/coi/firstreport.pdf.

[46] National Institutes of Health. “Financial Conflict of Interest – Objectivity in Research: Institutional Policy Review.” 10 February 2003. Accessed 9 September 2008 at http://grants.nih.gov/grants/guide/notice-files/NOT-OD-03-026.html.

[47] NIH Conflict of interest policy site: http://grants.nih.gov/grants/policy/coi/ (accessed 31 August 2008).

[48] Brand R, Marketing Drugs: Debating the Real Cost, Magazine of the National Conference of State Legislatures, September 2008 available at http://www.ncsl.org/magazine/articles/2008/08slsep08_drugs.htm (accessed 31 August 2008).

[49] Edelman D, Oddone EZ, Liebowitz RS, Yancy WS Jr, Olsen MK, Jeffreys AS, Moon SD, Harris AC, Smith LL, Quillian-Wolever RE, Gaudet TW A multidimensional integrative medicine intervention to improve cardiovascular risk. J Gen Intern Med. 2006 Jul;21(7):728-34.

[50] http://www.aetna.com/members/health_wellness/support_tools/genetic_counseling.html (accessed 9 September 2008).

[51] The state of managed care quality. Washington, D.C.: National Committee for Quality Assurance, 2006.

[52] Ginsburg GS. Genomic Medicine: 'grand challenges' in the translation of genomics to human health.

Eur J Hum Genet. 2008 Aug;16(8):873-4.

[53]Metcalfe, S., Hurworth, R., Newstead, J., and Robins, R., (2002) Needs assessment study of genetics education for general practitioners in Australia. Genetic Medicine. 4, 71-7.

[54] Frueh FW, Gurwitz D. (2004) From pharmacogenetics to personalized medicine: a vital need for educating health professionals and the community. Pharmacogenomics 5,571-79.

[55] Wolpert, CM. Surgeon General launches new public health campaign: the family history initiative. JAAPA 2005;18(1): 20-2

[56] Kozma, CM. The National Guideline Clearinghouse. Manag Care Interface. 2006;19(5): 43, 51

[57] Kupersmith J, et al. Creating a new structure for research on health care effectiveness. J Invest Med, 2005, 53:67-72.

[58] Habel L, Shak S, Jacobs M, et al., A population-based study of tumor gene expression and risk of breast cancer death among lymph node-negative patients, Breast Can. Res., 2006; 8:R25.

Return to Table of Contents

Personalized Medicine and Disruptive Innovation:
Implications for Technology Assessment

Kevin A. Schulman, MD
Ana Valverde Vidal, MBA, CFA
D. Clay Ackerly, MSc

Center for Clinical and Genetic Economics
Duke University School of Medicine

ABSTRACT

The fulfillment of the promise of personalized healthcare will likely require not only technology innovation but the adoption of new business and organizational models to allow for the new technologies to take hold in a disruptive fashion. At the root of the problem lays the question as to whether we have the right public policies and private strategies to allow for innovation to take hold in the healthcare arena. The current paper discusses a framework for considering this question, and proposes potential policy solutions to enable the adoption of technologies to yield improvements in both quality and costs.

The Promise of Personalized Medicine

New technologies offer the potential for revolutionary changes in the practice of medicine, from molecular diagnostic tests that detect disease before symptoms are evident to patient profiling techniques that help predict which patients are most likely to benefit from or be harmed by specific therapies. These approaches and the extensive data they require will need to be supported by a new information architecture. This system has been described as personalized health care—treatments and services targeted to the specific biology of the individual, leading to potentially significant improvements in patient care. Although this vision has been articulated for several years, researchers are slowly gathering the information required to support the adoption of specific technologies that will be the crucial building blocks of the system. Other aspects of this vision are less developed, and the investment theses required to bring new technologies to market remain speculative.

At the policy level, there are recurring questions of the correct approach to innovation in health care. Do we have the right public policies and private strategies in place to foster innovation in the health care system? At the core of these discussions is a question of whether personalized health care will require a new approach to technology assessment and dissemination, one that embraces the tremendous potential of the vision of personalized medicine. What is the role of technology innovation in health care, and what should be the public policy responses to innovation?

Technology and Innovation in Health Care

Technology development and diffusion can offer new opportunities for patient benefit. In assessing the role of technology innovation, one field of scholarship has explored the relationship between technology innovation and organizational innovation. This line of inquiry presents a useful framework for discussions of the broader policy questions related to personalized medicine.

New technology often is accompanied by new business models. In competitive markets, innovation in technology enables new business models to use the advances of the new technology to offer cost or quality advantages to the end user. When successful, these new combinations of technology and business strategy are able to supersede their predecessors.[1] This issue has been examined in detail by Christensen,[1] who assessed the relationship between technology innovation and organizational innovation in the computer disk drive industry.

Christensen’s concept of “disruptive innovation” begins with an assumption that consumer demand for a given technology is normally distributed, with the tails of the consumer preference curve representing high-end users with specific needs and price insensitivity and the low-end users with limited needs and price sensitivity. Firms already in the marketplace offer their products to all users but develop their technologies to meet the needs of the high-end users, their most valued customers. High-end users are thought to be the most profitable consumers and to be the most articulate about their needs. To satisfy the demands of this group, firms improve their technologies over time within the constraints of an existing business model, an approach Christensen termed sustaining innovation. The resulting products and services target the needs of the most lucrative segments of the market.

Christensen’s key observation is that so-called sustaining innovation leads firms to develop products that possess capabilities far beyond the needs of the average consumer. This strategy creates opportunities for new firms to enter the market with new technologies and business models that focus on the more limited needs of average consumers. When successful, these new firms can supplant the existing firms in a process called “disruptive innovation.”

There are many examples of disruptive technologies. One includes digital photography, which was disruptive to photographic film, as initial digital cameras had worse picture quality than traditional film and the computer tools available to edit and share photos were still in their infancy. However, the convenience of digital photography grabbed a new market niche, and soon the quality of digital photography improved and in many ways surpassed film photography. Other examples include minicomputers, which were a disruptive innovation to the mainframe market; personal computers were in turn a disruptive innovation to the minicomputer market. Mobile telephones were disruptive to fixed-line telephones. This dynamic process of firm entry and firm exit from markets offers tremendous potential benefits to consumers over time in terms of reducing costs and improving the quality of products and services available in the marketplace.

Disruptive Innovation in Health Care?

In the health sector, it is difficult to identify examples of truly disruptive technologies. Some have argued that home glucose monitoring, coronary angioplasty, and the nurse practitioner model are examples of disruptive innovations in health care.[1] However, none of these technologies has been able to fully disrupt the market. None has fundamentally changed the system of primary care or fostered the development of new and innovative models of health care delivery.[1] In most cases, these innovations have been unable to develop into the type of disruptive innovations we see in other markets when the new replaces the old.[1] Instead, technology has generally added to existing systems in the manner of sustaining innovation.[1] Physicians have fought the entry of the nurse practitioner model, and payment regulations restrict nurse practitioners to a primary care role.[1] The business model to offer a real-time interface between home glucose monitoring and the physician office has taken years to evolve.[1] Furthermore, angioplasty relies on the same hospital-based model as cardiac surgery; the procedure is simply performed by a cardiologist instead of a cardiac surgeon.[1]

In short, there is a lack of solid examples of disruptive innovation in health care. It is not difficult to discern why this might be the case. The health care industry exists through a relationship between business and government that is different from the computer disc drive industry that Christensen observed. These interactions, in the form of regulations, professional standards, and administrative procedures, create opportunities for incumbents to support the status quo by erecting barriers to market entry. The typical firm bringing a disruptive innovation to market is unable to meet these established rules, since it characteristically offers products or services with a narrower or more limited scope, a different business model and potentially a different customer focus from that of incumbent firms. An unintended consequence of this system is an environment that supports sustaining innovation over disruptive innovation. The health care market does not have the advantage of disruptive innovation to drive cost and quality improvements in the marketplace.

Administrative Barriers to Innovation

We have previously proposed the adoption of the definition of regulatory controls (“regulations”) offered by Berenson[1]. Further, we have adopted this framework for both the public and private regulators in the healthcare market as administrative barriers. Regulation has the effect of developing a set of rules and standards for a market, including rules governing market expansion and a process for firm entry into a market. Regulations in health care include the governance of third party payments in health insurance[1] a medical liability system based on the standard of care, or rules on hospital markets (certificate of need requirements). As we discussed above, new entrants may not meet these administrative criteria or may not be able to navigate this process. As a result, all of these regulations may inhibit entry of new business forms.

Market entry is a dynamic process. Given an equal opportunity, entry will be greater when profit opportunity is greatest and barriers to entry are lowest. Given the high cost of most services in health care and the inherent profitability of the system, the health care market should be an attractive opportunity for firm entry. Also, given the quaternary care model rampant in the marketplace, existing firms have developed capacity that outstrips the needs of most consumers (and have failed to provide the front-end services demanded by most consumers). So the lack of entry should not be ascribed to a lack of interest in the market by investors.

There is a constraint on entry in this simple model—the cost of entry. The cost of entry can be seen as the cost of complying with administrative processes to create a new business model, or the cost of complying with regulatory standards that require entrants to achieve the same form or capabilities as incumbents to enter the market. These requirements can increase the cost of entry to the point where entry is no longer attractive to new firms.[1] Alternatively, these factors may alter the risk of any investment by increasing uncertainty regarding approval of a new business model.

The relative lack of firm entry has consequences throughout the health care marketplace, on both incumbents and new entrants. In the absence of new market entrants (or a viable threat of entrants), organizational innovation of existing firms lags or disappears.[1] This lack of organizational innovation on the part of incumbent firms compounds the cost and quality consequences of firm trajectories comprised of sustaining innovation on the marketplace.

The Policy Maker’s Role

This discussion has emphasized the potential negative consequence of current regulatory and governance practices on the health care marketplace. Clearly, regulations serve an essential role in the healthcare system[1]. However, by establishing a threshold above average consumer performance expectations, regulations may also preclude quality-enhancing, lower-cost innovations from entering the market. What can policy makers do to promote innovation and allow for these new technologies to enter this regulated marketplace? Another way of framing this question is, how should we take account of the negative externalities of a regulatory scheme on the marketplace?

One simplistic framework would suggest we should support adoption of disruptive innovations over sustaining innovations. This approach is clearly supported by the theoretical framework, but it contrasts with current technology adoption models, in which most technologies that reach the market are simply sustaining innovations that “add on” to existing technologies. For example, greater availability of angioplasty is now associated with more revascularization procedures per population among people older than 65 years[1]. Similarly, the availability of more magnetic resonance imaging (MRI) units does not reduce the number of computed tomography (CT) scans performed.[1]

Supporting disruptive over sustaining innovations is not a simple task. Disruptive innovation is not a result of technology innovation; rather, it is a combination of business and technology innovations. It is unclear from an assessment of a technology itself whether the business model is one that offers the potential for disruption. Second, although many products purport to be disruptive innovations, true disruptive innovations can only be identified in arrears when markets have changed as a result of the innovations. Even with these limitations, however, potential pathways forward could emerge.

First, not all types of innovation are, or should be, of equal interest to policy makers. In most markets, sustaining innovations are ones that enter the market continuously. New versions of Microsoft Windows and new models of the Apple iPod come to market with greater capabilities than previous versions at equal or lower prices. From this perspective, the regulatory approach could be one that expects sustaining innovation as a condition of remaining on the marketplace and limits the financial rewards to products or services over time. For example, imagine if we lowered the price for an MRI each year based on an index of computer costs in the broader marketplace.

The treatment of potentially disruptive innovations, however, could be considered quite differently. In the health care environment, disruptive innovations face tremendous uphill battles, with new combinations of technology and business models that have not previously existed. Based on the theory presented to this point, regulators could consider facilitating entry of these firms and technologies as a means of enhancing the price and quality of health care services for consumers. At the same time, regulators should curtail these incentives for firms and products that do not prove to be disruptive. This suggests that broader regulatory reform would accomplish the former goal of allowing access but at the expense of many sustaining innovations benefiting from the new framework. An alternative would be a time-dependent facilitated pathway for market entry that is unique to the regulatory framework we have constructed for health care. For example, policy makers could determine a mechanism to identify technologies with potential to become disruptive and to allow these technologies to enter the market in a disruptive fashion.

Exploring Potential Policy Options

One such mechanism would be the creation of an Office of Personalized Medicine (OPM) charged with reviewing new technology applications to determine if they have the potential to become disruptive. With data and business cases presented by the owners of the technologies, the OPM would assess the ability of an innovation to transform health care delivery and treatment and to eventually lead to improvements in both outcomes and cost. Such a review mechanism would encourage technology owners to think beyond the novel characteristics of their proposals to consider early on other important business and operational features that would eventually determine if an innovation goes beyond being sustaining to become truly disruptive.

For innovations deemed disruptive by the OPM, policy makers could play an important role in providing incentives for these technologies to successfully enter the market. Owners of disruptive innovations could receive vouchers for accelerated review, or the innovations could command a premium in reimbursement negotiations. Regulators could even define a special regulatory pathway for these technologies, with distinct market approval and reimbursement criteria that would more closely align with the characteristics of these technologies. As an alternative, regulators could carve out “safe harbors” for these technologies, giving the owners of such innovations flexibility and time to change the prevailing business models in their sector. Following a model similar to ”coverage with evidence development” (CED) in the Centers for Medicare & Medicaid Services (CMS), innovations considered disruptive could be subject to special reimbursement mechanisms for a given period of time, altering the prevailing incentive system in the market place and enabling the new technology to take hold. For example, under the current encounter-based reimbursement system, health care providers have little incentive to acquire technologies that enhance service but reduce the number of encounters at the clinic, because such innovations would likely result in reduced revenues for the provider. Under a “safe harbor” mechanism, health care providers who use an OPM-labeled disruptive technology to remotely monitor patients would likely be able to bill for the informal communications that such technologies would generate (eg, e-mail consultations, phone conversations).

Example: The Complex Development Path for a Potentially Disruptive Innovator

One clinical application for personalized medicine is targeted therapy for individual patients.[1] The potential implication of this approach is to offer improved safety and efficacy for individual patients[1] and have an immediate economic impact by avoiding therapies with low potential to be efficacious (although one would expect manufacturers to respond to this technology in terms of development and pricing strategies over time).

Currently, the regulatory pathway for development of diagnostic tests for personalized medicine applications is controversial depending on whether the test or the information from the test kit are the product. A company seeking approval for a novel molecular diagnostic test for which the test kit is being marketed requires approval from the Center for Devices and Radiological Health (CDRH) at the Food and Drug Administration (FDA).”[1] CDRH classifies devices into three regulatory classes based on the anticipated use of the technology and the inherent risk. The class assignment determines the requirements for approval as well as the complexity of the marketing approval process (either premarket notification or the more stringent and lengthy premarket approval).

Alternatively, in vitro diagnostic devices can be developed and marketed under the Clinical Laboratory Improvement Act (CLIA) of 1988, which governs “laboratory-developed tests” (ie, tests performed in a single site where the test kit is not marketed; samples can come to the laboratory for this service from anywhere in the country). CLIA establishes three categories of testing on the basis of the complexity of the testing methodology: waived tests, tests of moderate complexity, and tests of high complexity. Laboratories performing moderate- or high-complexity testing must meet requirements for proficiency testing, patient test management, quality control, quality assurance, and personnel. However, CLIA-governed tests do not require FDA approval.

The distinction between these two separate pathways has created a special area of controversy for personalized medicine. Using gene expression technology, scientists have reported an ability to classify patients based on risk of disease recurrence.[1] Although the technology is in its infancy, the FDA has raised concerns about the regulatory pathway for in vitro diagnostic multivariate index assays (IVDMIAs). These devices combine the values of multiple variables using an interpretation function to yield a single, patient-specific result that is intended for use in the diagnosis of disease, or in the cure, treatment or prevention of disease; and they provide a result with a nontransparent derivation that cannot be independently derived or verified by the end user.[1]

Most IVDMIAs in the market are laboratory-developed tests marketed through the CLIA route, that is, tests developed by a single clinical laboratory for use in that laboratory alone. Given the strategy of not marketing the test kits and performing the tests at a single site, these tests did not fall within the scope of lab tests over which the FDA had generally exercised enforcement discretion. Concern over this space has led to a proposal to begin regulation of this market by the FDA, with the issuance by CDRH of draft guidance in July 2007.[1] This regulatory issue has not yet been resolved.

In addition to regulatory approval, companies seeking to enter the market with new molecular diagnostic tests also must work with CMS to obtain reimbursement for their products. This separation of approval and reimbursement results from the different missions assigned to both FDA (approval) and CMS (reimbursement). FDA approval is based on meeting statutorily defined criteria of safety and effectiveness, and literally provides permission to market a product in the US. Implementation of these criteria varies by product category, but serves as a minimum set of criteria for entry into a market. FDA review and approval is not an assessment of value, uniqueness, nor a recommendation for use or funding of a product or technology. CMS review, on the other hand, is based on a statutorily defined standard of “reasonable and necessary” for the treatment of illness or injury. This standard for reimbursement is an assessment of whether a technology should be used in the care of Medicare beneficiaries. It is a relative standard and can be influenced by the existence of an unmet medical need, the existence of comparative therapies and the value of a new technology. In principle, the separation of approval and reimbursement provides an easier entry to the market for a technology (approval), and allows the sale of a product even if there is no reimbursement by CMS for the technology.

The CMS reimbursement process itself is a complex one. The process governs three key issues—coverage, coding, and payment:[1] As we mentioned above, to be covered by Medicare under the Social Security Act, the new technology must be “reasonable and necessary” for the treatment of illness or injury; however, technologies that are predictive may not meet this standard since prevention is not considered medical treatment. Second, as medical claims processing has become automated, assignment of specific codes for new medical technologies has taken on a unique importance in the reimbursement process. If specific codes are not available for a new technology, payment for the technology cannot be differentiated from previous technologies. Finally, payment schemes in Medicare can vary from bundles (inpatient DRG payment) to specific (outpatient laboratory testing). When the technology will be reimbursed separately, a payment rate must be established.

Coverage, coding, and payment decisions are not necessarily made in any particular order, and the decisions can span a 12-month period. To add to the complexity, different components of CMS are responsible for different aspects of these decisions. The Office of Clinical Standards and Quality oversees national quality initiatives and includes the Coverage and Analysis Group and its three divisions, which are responsible for developing national coverage policy. Payment and coding decisions are developed by the Center for Medicare Management, with two groups and ten divisions potentially involved in the process. In addition, there is the possibility that different regional decisions can be made about these issues in the absence of a national decision.

In recent years, CMS has shown an awareness of the need to streamline this process and has taken several steps aimed at improving it. In 2004, the Council on Technology and Innovation (CTI) was established under the Medicare Prescription Drug, Improvement, and Modernization Act to serve as a coordination point for new medical technologies. In August 2008, the CTI published the Innovators’ Guide to Navigating CMS[1] to assist stakeholders in understanding the processes used to determine coverage, coding, and payment. While serving to help technology developers understand the CMS process, the CTI group is not an expedited pathway to market for new technologies.

CMS has launched several demonstration projects to test innovations in reimbursement policy. For example, the CED policy provides an abbreviated pathway to Medicare coverage while still requiring further evaluation of a new technology. At the same time, CMS is working to make coding processes more efficient and has implemented a number of initiatives to reform one of its major coding systems, the Healthcare Common Procedure Coding System (HCPCS), while moving to replace the International Classification of Diseases, Ninth Revision, with the more flexible and clinically relevant Tenth Revision.[1]

A Way Forward for Disruptive Innovations

While these actions are steps in the right direction, a broader approach could help accelerate the access of disruptive innovations to the market. The OPM could play a significant role as a unifying and coordinating agency, acting as the single point of contact through the Department of Health and Human Services for technologies deemed disruptive. The OPM would help expedite the approval process by expertly understanding all the potential pathways involved and by helping the technology navigate the regulatory mesh. In this role, the OPM would act as an ombudsman for disruptive innovations that are seeking market approval. As described above, this process would not be open to all potential innovations but rather to those that, based on the technology characteristics and the proposed business model to implement them, are considered to have disruptive potential.

The consideration of disruptive potential would only be granted for a fixed period of time. If after such period the technology fails to deliver its disruptive promise and its novel business model fails to take hold, the OPM could elect to levy some penalties on the company, either monetary (to payback the competitive advantage gained through early market entry) or other (such as closing the OPM pathway for future innovations from the company for a given period). The intent is to make the penalty significant enough that companies will exercise best efforts to deliver on the disruptive promise of their innovations.

The OPM could build on these changes and work in tandem with the Council on Technology and Innovation, as well as the CMS Office of Research, Development and Information. Close communication between these groups would ensure tight coordination through the regulatory and reimbursement approval processes. The OPM could also work with these groups to expand current initiatives and create new, larger demonstration projects or safe harbors for disruptive innovations. The OPM would also have to follow a strict timeline to ensure a speedy decision about whether a technology will meet the OPM standard.

Much of this policy assessment has focused on the unique role of the Federal government in the health care marketplace. Private health plans often adopt much of their coding infrastructure from Medicare and can select to follow Medicare in coverage decisions. Thus, efforts to adopt these policies by the public sector will have effects on the private sector, as well. Creating transparency in the rationale for OPM decisions and communicating the results of evaluation of technology implementation can also help to shape decisions in the private sector. Separate study of the role of the private sector in fostering disruptive innovation merits further consideration.

Conclusions

Personalized medicine offers the potential for revolutionary change in the practice of medicine. It also provides a unique window into the relationship between new medical technologies, new business models for health care delivery, and the role of government in this unique marketplace. Using personalized medicine as a test of disruptive innovation in health care, we find the need for a different approach to these technologies in order for them to achieve their full potential. Achieving this result, however, is fraught with difficultly, as disruptive innovations are deemed truly disruptive only in arrears. Thus, our approach offers the potential that designations of a technology as potentially disruptive would provide competitive advantages to products or services that may not merit this consideration. A robust framework for continuing assessment (and the potential for penalties on misrepresented technologies) might help protect the integrity of this process. However, the benefits of unlocking the health care market to disruptive innovation seem to be worth the risk.

Return to Table of Contents

Assessing Risk and Return:
Personalized Medicine Development
And New Innovation Paradigm

Frank L. Douglas PhD, MD - Senior Fellow
Lesa Mitchell - Vice President, Advancing Innovation
Ewing Marion Kauffman Foundation

Introduction

In making a credible business case for investors and industry stakeholders to view personalized medicine as a viable business model, we not only must create excitement in the promise of personalized medicine, but also must find viable alternatives in addressing the barriers or risks surrounding the biomedical discovery and development models of today. Some of the risks we identify include IP issues, difficulties in validating targets, ability to rapidly achieve proof of concept, navigating the famed “Valley of Death,” and inefficiencies in the current clinical development process, as well as the need for new industry business models that predict an attractive return on investment. In this paper; however, we limit our discussion to the potential for personalized medicine to create efficiencies in the preclinical and clinical phases of drug innovation and generate economic returns. We also introduce unique industry collaboration mechanisms with nonprofit disease-focused organizations that serve an important role in de-risking aspects of drug discovery and clinical development in their respective disease sectors, as well as bridging early-stage funding needs. These collaborations and de-risking strategies could provide an important model for the further development and growth of the personalized medicine sector.

With respect to definition, we shall use the more general term “stratified medicine,” of which personalized medicine is the individualized member of a spectrum that includes empirical medicine, stratified medicine, and personalized medicine.[1] In the latter two, a biomarker is critical in identifying sub-populations or strata of patients that can benefit from a therapeutic intervention that is related to that biomarker, or develops a therapy that specifically benefits an individual who possesses that biomarker.[1] A biomarker also may identify strata of patients that might be susceptible to side effects from a particular therapy.[1]

Current Challenges in Productivity and Investment Returns

The increasing interest and excitement over the promise of stratified medicine is based on the promise of genomics, proteomics, and metabalomics to enable the researcher to identify genes and gene products that are relevant for disease, and to instruct the creation of the best therapies for patients with the respective diseases or side effect susceptibilities.[1] This comes on the heels of the biopharmaceutical industry struggling to meet the increasing demands on its R&D investments while facing declining levels of productivity and innovation, and loss of revenue due to patent expirations. More than three dozen drugs are losing patent protection between 2007 and 2012, with an anticipated $67 billion loss in sales for the large pharmaceutical companies to generic competition.[1] The industry has responded with pharmaceutical companies increasing R&D spending by 160 percent—from $15 billion to $39 billion from 1995 to 2005—and with similar increases in the biotech industry, with a 150 percent increase—from $8 billion to $20 billion—in R&D spending during the same period. Meanwhile, submissions for regulatory approval of new drugs and therapeutic indications declined from eighty-eight in 1995 to forty-four in 2004.[1] Innovation in the sector also is continuing to decline, with only seventeen new molecular entities (NME) and two biologics approved in 2007, at a cost of $2.5 billion per NMEs approved,[1] which is the lowest innovation-to-productivity level since 1983, when twelve NMEs were approved at a cost of $266 million per NME.[1] (See Figure 1.)

Figure 1: Comparison of biotech and pharmaceutical R&D

Figure 1: A comparison of biotech and pharmaceutical R&D productivity. Source: Parexel’s Pharmaceutical R&D Statistical Sourcebook 2005/2006; Defined Health Analysis. NME, new molecular entity.

The decline in productivity and innovation has increased M&A and partnering activities among large biopharmaceutical companies at a record high in the last few years, with $150 billion generated through M&A transactions in 2006 and $22 billion in partnering deals for the same period.[1] The strategy of focusing on a few drug candidates from their combined pipelines, with a focus on producing several “blockbuster” drugs that will generate at least $1 billion individually in peak annual global sales and be marketable to fifteen million patients or more, has not improved their productivity levels, resulting in increased delays in development time/costs and increasing cancellations of projects at later stages of development.[1] Additionally, increasing regulatory pressures to conduct more lengthy and complex trials has added to the current $1 billion[1] in drug development costs, of which half are attributable to the time value of money—that it takes eight to twelve years to get a drug to market.[1] It is also the case that, even after a drug is marketed, 70 percent of the approved drugs do not meet or only match their R&D costs.[1] Thus, with lower efficacy levels (40 percent to 60 percent) of most blockbuster drugs,[1] as well as some high-profile successes of stratified medicines such as Genentech’s Herceptin and Novartis’ Gleevec, the industry is beginning to realize the deficiencies in the economics of the blockbuster business model, which is one of the drivers of increased interest and investment in the development of stratified medicine.[1]

Early-Stage Funding Challenges in Stratified Medicine Development

The identification of clinical biomarkers or diagnostics linked to gene expression profile of individual or sub-populations of patients is an essential feature of stratified or targeted medicine. This type of research attracts and often is best pursued by small biotech companies. One of the main challenges for these companies lies in the lack of early-stage funding to translate new discoveries into the clinic and, ultimately, to commercialization. With a narrowing access to public capital and venture capitalists increasingly reticent to invest in early-stage technology companies, smaller biotech companies increasingly are engaging in alternative financing mechanisms that often compromise their value in terms of access to future returns.[1]

Various alternative financing mechanisms, including partnering and out-licensing, sale of royalty streams, and Contract Research Organization (CRO) financings, all include investment capital in exchange for future royalty rights or equity shares in the biotech company.[1] Other innovative financing mechanisms do exist, such as collaborative development financing (CDF), where an investor provides capital and clinical expertise in exchange for licensing of a company’s pipeline, while the company maintains the “exclusive right to reacquire the drugs,” at prices determined at the time of the agreement.[1] An example of a CDF arrangement is the 2006 Symphony Capitol and Isis Pharmaceuticals (“Isis”) collaboration,[1] where Isis received $75 million to continue the development of its cholesterol-lowering (Phase II) and diabetes drug products (two in pre-clinical) and agreed to an exclusive purchase option for its products at an “annual rate of return that averages 32 percent and is 27 percent at the end of the anticipated” collaboration period.[1] In 2007, Isis exercised its repurchase option, paying Symphony $131 million. Isis, in turn, executed collaboration agreements with Johnson & Johnson and Genzyme for the three molecules in the contract. These arrangements included upfront fees in the aggregate of $370 million with potential milestone payments of nearly $2 billion.[1] (See Figure 2.)

Figure 2: Alternative financing sources for biotech companies[1]

Most of the alternative financing mechanisms, however, are not necessarily accessible for many early-stage companies, as these companies may not have the types of products that meet the returns desired by larger companies and venture capitalists. A case in point is the lack of investment in orphan drugs or neglected disease areas. Aside from Genzyme, which has been one of the few successful orphan drug-focused companies with three drugs on the market, including a $1 billion-a-year treatment for Gaucher, and Novartis’ Gleevec, a treatment for chronic myeloid leukemia with $2.5 billion in 2006 sales,[1] therapeutic discovery and development for orphan and neglected diseases often have been the bane of nonprofit foundations and patient advocacy organizations, many of whom have increasingly taken on a new role of bridging early-stage funding and development gaps in disease areas where the patient population often is less than 200,000, the FDA definition of orphan drugs.[1]

To uncover mechanisms by which venture capitalists and biopharmaceutical companies—whose measures of success ultimately are captured in their return on investment (ROI)—could be incentivized to participate in developing stratified medicines, we have looked at the various activities of nonprofit foundations. In our view, these foundations, whose ultimate success is in bringing therapeutics and diagnostics to their patients, increasingly are engaged in “de-risking” strategies. In some cases, their target patient populations fall within the orphan disease category. Their strategies, however, not only fill important funding gaps but also have the objective of increasing the probability of success through their support activities.

Venture Philanthropy—Early-Stage Funding/Proof of Concept

Although the nonprofit foundations traditionally provide basic research grants to increase scientific knowledge in their disease sectors, some have since adopted a more investor-like approach—early-stage funding for proof of concept and target validation, as well as project management support and access to their network of scientific experts and research clinics critical in translating discoveries into the clinic.

One example of nonprofit disease organizations that provide early-stage funding for proof of concept and target validation is the Muscular Dystrophy Association (MDA). Through its Translational Research Program (TRP), MDA’s approach is to stratify its patient population based on various sub-sets of the disease, including Duchenne Muscular Dystrophy (DMD), Myotonic Muscular Dystrophy (MMD/DM), Fascioscapulohumeral Muscular Dystrophy (FSHD), Spinal Muscular Atrophy (SMA), Pompe Disease, and ALS, and seek to develop targeted therapies for the sub-patient populations.[1] Of the $32 million in MDA’s 2007 annual R&D budget, $6 million was dedicated to its largest collaboration effort with ALS Therapy Development Institute (ALS-TDI), a nonprofit corporation, and $7 million was dedicated to industry collaborations.[1] Muscular Dystrophy Association’s TRP provides four types of funding mechanisms for the industry—IND Planning Grant, Clinical Research Training Grant, Infrastructure Grant, and Corporate Grant—to catalyze early-stage development leading up to INDs and Phase I/II clinical trials.[1] (See details of collaboration deal examples at Figure 3.)

**Figure 3: Examples of TRP Industry Grants**[1]
Disease Type & Company Grantees	Collaboration Description and Status
DMD/PTC Therapeutics	MDA provided PTC with an initial $1.5 million grant, enabling the company to begin developing PTC124, a medication with the potential to treat a significant portion of patients with DMD. In July 2008, PTC entered into a collaboration deal with Genzyme, where Genzyme will provide $100 million to PTC, with potential additional payment options, and will commercialize PTC124 outside the United States and Canada.
Pompe Disease (acid maltase deficiency)/Myozyme (approved 2006) from Genzyme	MDA provided supplemental funding of $150,000 to cover unreimbursed costs of patients participating in Genzyme’s clinical trials for Myozyme in infantile-onset Pompe disease. In 2007, Genzyme also found Myozyme effective for older children and adults with the disease.
ALS Therapy Development Institute (ALS-TDI)	MDA is collaborating with ALS-TDI to comprehensively characterize disease progression in ALS using animal models of neurodegeneration and ALS clinical samples. MDA committed $6 million annually for three years.

To qualify for the TRP grants, the collaborating company is required to provide matching grants and agree to a collaboration contract that includes royalty-sharing agreements and march-in rights if the projects fail to meet milestone targets. Similar to a majority of the nonprofit disease organizations, MDA neither takes equity positions in the companies with which it collaborates, nor pursues IP ownership.[1]

Another example of nonprofit disease organizations providing early-stage funding to industry includes the Industry Discovery & Development Partnerships (IDDP) Program of the Juvenile Diabetes Research Foundation (JDRF). IDDP’s main focus is to translate scientific discoveries into the clinic and support commercialization of therapeutics to treat type 1 diabetes.[1] Of its $160 million research budget in 2008, $16 million will be dedicated to industry partnerships,[1] which is a marked change. Previously, 100 percent of its research funding went to support basic science and exploratory research within academia.[1] To date, IDDP has fostered twenty-four collaborations with industry, totaling $30 million in IDDP grants.[1] IDDP’s development partnerships are generally two- to three-year contracts, and “are intended to provide support for promising mid-stage research programs (i.e., advancement of a pre-clinical-stage program to clinical trials, or “proof-of-concept” Phase II clinical testing of promising therapeutics.”[1] By funding early-stage testing and validation of research, JDRF’s model of “de-risking” works to make it possible for its industry collaborators to advance their compounds from proof of concept to clinical development, attract additional financing, and eventually secure global licensing and marketing alliances with larger pharmaceutical companies.[1] By funding and providing development support of early trials through IDDP, JDRF also sees this as a way to build evidence in persuading public and private payors to cover these novel technologies.[1] A case in point is IDDP’s collaboration with Tolerx. JDRF provided early-stage, multi-million dollar funding for proof of concept trials in both animal models and early human trials for anti-CD3 antibodies (Otelixizumab) for the treatment of early-stage Type 1 diabetes in collaboration with academic researchers in the United States and Europe.[1] To catalyze further development and commercialization of Otelixizumab, IDDP invested $3.5 million in an equity stake during Tolerx’s latest round of fundraising to conduct Phase II trials.[1] This is the first project where IDDP has taken an equity position in a collaborating biotech company. As of October 2007, Tolerx entered into a strategic alliance deal with GSK to take the antibody through Phase III trials, with a total deal value potential up to $155 million.[1] Figure 4 below also exemplifies the significant commitment IDDP has made to companies to support discovery, development, and commercialization of therapeutics and devices for type 1 diabetes.

Figure 4: IDDP Discovery and Development Pipeline

Venture Philanthropy and Nonprofit Venture Affiliates

Few nonprofit disease organizations have created wholly owned nonprofit venture affiliates to navigate through the challenges of translating early-stage discoveries into the clinic or bridging the “Valley of Death.” These entities serve as catalysts on various scales, not only by providing variable funding options from annual to multi-year commitments averaging from thousands to multi-millions of dollars, but also by providing mechanisms to address the development challenges. These include: providing project management expertise and scientific, clinical, and development networks (in some cases CRO outsourcing networks) that can assist the collaborators. In terms of return on investment, most do not take equity positions in the companies they collaborate with; instead, some deals are royalty-based, in which the organizations get a multiple back if the drug is approved and, in some cases, additional compensation for extraordinary sales results. Additionally, in cases where collaboration programs suspend due to milestone failures, some organizations obtain worldwide rights to develop the products with an agreement to negotiate royalties to the original collaborator once their investment is recouped.

An example of a nonprofit disease organization that has created unique project management and target validation mechanisms is the Multiple Myeloma Research Consortium (MMRC), a supporting organization of the Multiple Myeloma Research Foundation (MMRF). Through a collaborative contractual arrangement with its fifteen research centers,[1] the MMRF’s strategy is to incentivize biopharmaceutical companies to collaborate on the development of new drugs and therapies. The MMRC’s tri-focus on genomics and credentialing of molecular targets, validation of drugs, and its offering of multi-site clinical trial capabilities creates efficiencies that are critical in de-risking early-stage proof of concept and target validation.[1] One of the MMRF’s strategies is to identify genetic complexities of multiple myeloma and to identify molecular targets by analyzing the MMRC’s tissue bank and patient data bank on disease onset and progression, with the goal of personalized medicine development.[1] To assist in the process of validating new targets, the MMRC has created screening tools—including a panel of twelve extensively characterized myeloma cell lines with full genetic and biological characterization—to screen new drug candidates.[1] The MMRC also has funded the Multiple Myeloma Genomics Initiative, investing $8 million in research funding over the past four years to analyze 250 patient tissue samples via gene expression profiling, comparative genomic hybridization and exon re-sequencing.[1] To expedite and create efficiencies in conducting multi-site clinical trials of novel and combination therapies, the MMRC has created uniform contracts, clinical trial agreements, and correlative sciences agreements.[1] (See Figure 5.) To further expedite the process, the MMRC provides supplemental project management to accelerate projects from protocol concept through trial conduct and provides clinical research coordinators for the MMRC members.[1] The MMRF sees its main function as an integrator and facilitator of research and collaboration among biopharma companies with the research centers.[1] Since 2003, the MMRF has helped bring four drugs to market, including Millennium Pharmaceutical’s Velcade in 2003, Celgene Pharmaceutical’s Thalomid^® and Revlimide^® in 2006, and Millennium Pharmaceutical/J&J Pharmaceutical’s Doxil^® in 2007, [1] and has supported more than thirty compounds and combinations in trials or pre-clinical studies to date.[1]

Figure 5: MMRC Clinical Trials. MMRC Trials and the year in which they have opened. A total of 15 trials have initiated in the MMRC since 2005. Abbreviations: R: Relapsed; R/R: Relapsed/Refractory; Rev: Revlimid; Dex: Dexamethasone; Vel: Velcae; IST: Investigator-sponsored trial. Unless marked as IST, all trials are company-sponsored. **Trials expected to open by year-end 2008.

Figure 5: MMRC Trials and the year in which they have opened

From a funding perspective, 93 percent of the MMRF’s annual budget goes to research and related programming.[1] Of these, in 2007, the MMRF earmarked approximately $15 million for R&D, with $2 million allocated for direct funding to biotechs.[1]

One of the leading examples of a nonprofit venture affiliate is the Cystic Fibrosis Foundation Therapeutics, Inc. (CFFT), a wholly owned venture arm of the Cystic Fibrosis Foundation (CFF). CFFT’s focus is to develop stratified medicine based on CF-related genetic mutations, of which there are 1,400 on a single gene.[1] To date, CFFT has successfully identified and is working on the development of therapies that target the basic defect of the disease, as well as those that will provide better options for disease management. Therapies that target the basic defect are based on various genetic mutations, including Delta F508, a genetic mutation present in 90 percent of cystic fibrosis (CF) patients, and G551D, which is present in 10 percent to 30 percent of CF patients.[1] CFFT’s strategy is to invest in early-stage discovery and development. Their funding ranges from $50,000 to $25 million, with an average of $2 million to $4 million per year, with some multi-year commitments averaging $15 million to $20 million.[1] CFFT’s successes in aiding drug discovery are measured in terms of increasing its pipeline, which has grown to more than thirty drug candidates.[1] CFFT administers the collaboration contracts based on milestone successes, with pull-out rights for failures.[1] It also invests in a wide range of technologies, from target identification, novel screening platforms, detection of new chemical compounds, and screening of existing compounds and drugs.[1] In terms of return on investment, CFF does not take equity positions in the companies with which it collaborates; instead, some deals are royalty-based, in which CFF may get a multiple back and/or a percent of revenue if the drug is approved and, in some cases, receives additional compensation for extraordinary sales results.[1] Should the development program suspend due to milestone failures, CFF obtains automatic worldwide rights to develop the product with an agreement to provide some royalties to the original collaborator once CFF’s investment is recouped. [1]

An example of CFFT’s largest industry collaboration to date includes a multi-year collaboration with Vertex Pharmaceuticals, Inc. (Vertex), in which CFFT provided an aggregate of $76 million from 2000-2008[1] to support the development of two compounds (VX-770 and VX-809), which target the functional restoration of the cystic fibrosis transmembrane conductance regulator (CFTR) protein, the protein responsible for the progression of cystic fibrosis. Through this collaboration, Vertex was able to develop VX-770 from discovery to Phase IIa, where it focused on how VX-770 affects CFTR protein function and clinical endpoints in CF patients with genotype G551D (affects approximately 4 percent of the 30,000 CF patient population in the United States), achieving positive interim results in March 2008.[1] See other examples of CFFT’s portfolio in Table 6.

**Table 6: Examples of CFFT investments**
Collaborating Company	Project Description	CFFT Investment
EPIX Pharmaceuticals, Inc.	Use of EPIX proprietary PREDICT technology to create a computerized 3-D model of CFTR protein, using the model to identify sites within Delta F508 mutation of CFTR and search their library of chemical compounds for a small molecule that may work on those sites. In 2007, EPIX discovered a molecule that, in the lab, restores function to Delta F508 CFTR in cells.	$52 million including an original $18 M research award over 3 years and a subsequent discovery and development award over 7 years.
FoldRx Pharmaceuticals, Inc.	Use of a novel screening platform to detect new chemical compounds that could improve the function of misfolded proteins, like the Delta F508 mutation.	$22 million over five years to use its high-throughput screening platform to discover and develop new compounds.
CombinatoRx, Inc.	Screening approximately 2,000 approved drugs individually or in combination for its impact on correcting Delta F508 in the lab.	Commitment up to $13.8 million.
Vertex Pharmaceuticals, Inc.	Development of VX-770, its first CFTR modulator clinical compound, which entered Phase II clinical in 2007. Also developing second compound known as “correctors,” VX-809.	$76 million to date for VX-770 and VX-809.

Venture Philanthropy and Nonprofit Venture Intermediaries

Few large foundations, like the Gates Foundation through its Global Health Program (GHP), utilize independent nonprofit venture intermediaries to finance and manage the discovery and development of innovative therapies for neglected diseases affecting the developing world.[1] GHP’s goal through its venture intermediaries is to accelerate R&D and provide global access to new vaccines, drugs, and other health tools that combat infectious diseases, including malaria, HIV/AIDS, TB, and pneumonia.[1] The venture intermediaries serve “as a virtual pharma company looking for good ideas, progressing them to the point where proof of concept is achieved,”[1] and de-risking projects to the point that big pharma may be incentivized to collaborate in developing the therapies.[1] GHP is involved in the portfolio management of the venture intermediaries, but the intermediary conducts the project management.[1] To date, GHP has committed $6 billion in global health grants to organizations and researchers worldwide, including $200 million to Medicine for Malaria Ventures (MMV) over five years.[1]

The venture intermediaries, often called Product Development Public-Private Partnership (PDPs)[1] entities, operate globally with a focus on providing R&D funding and project management expertise in the neglected disease areas such as Malaria and TB.[1] MMV is one of the nonprofit venture intermediaries that the Gates Foundation and GHP funds.[1] MMV’s role is to facilitate the discovery and development of innovative anti-malarial drug candidates into clinic.[1] MMV does not conduct discovery or development itself but provides financial and project management support requiring milestone achievements and quick termination rights for those who fail to meet milestones.[1] In return for its investments, MMV often seeks IP rights from the discovery and development projects it funds.[1] In projects that it funds through commercialization, MMV will often negotiate for the delivery of drugs to poor developing countries at "no profit, no loss" basis.[1] It also will retain the ability to license to multiple drug manufacturers.[1] In cases where industry partnership fails during the development phase, MMV will either take full ownership of the IP or require an exclusive, worldwide, transferable license that is royalty free in malaria endemic countries.[1]

In 2007, MMV invested more than $37 million in nearly forty projects that include four projects in late-stage Phase III clinical trials and three mini-portfolios with GlaxoSmithKline (GSK) (three projects), the Broad Foundation/Genzyme (five projects), and Novartis Institute for Tropical Diseases (NITD)/Novartis (nine projects).[1] Clinical trials MMA supported in 2007 include: Collaboration with Novartis' submission to Swissmedic for approval of its first ACT (Coartem^® Dispersible); Eurartesim^® (with Sigma-Tau Pharmaceuticals, Inc.), which received orphan drug designation in the U.S. in 2006 and by the EU in 2008; and MMV/Shin-Poong Pharmaceuticals collaboration for Pyramax^®. MMV has a wide platform in its collaboration with Shin-Poong, covering two pivotal trials for Plasmodium falciparum, trials for P. vivax, and also a new formulation specifically for small children.[1]

MMV also has engaged in identifying new targets based on the genome sequence of Plasmodium falciparum, the main cause of human malaria, and has collaborated with Novartis and GSK to screen their collection of compounds that may be able to kill the malaria parasite. Out of more than three million compounds tested, more than 10,000 showed interesting activities at low micromolar concentrations.[1] (See Figure 7.)

**Figure 7: Sample MMV investments in 2007**[1]
Collaborating Company	Project Description	Amount Invested in 2007
MMV/Novartis (Coartem^® Dispersible)	Phase III trial—Development of a pediatric dispersible tablet, Coartem^® Dispersible, containing a fixed-dose combination of artemether and lumefantrine. (ACT)	$1.68 million
MMV/Sigma-Tau Pharmaceuticals, Inc. (Eurartesim^®)	Phase III trial—Fixed-ratio drug combination of dihydroartemisinin and piperaquine, being developed to treat uncomplicated malaria.	$2.85 million
MMV/Shin-Poong Pharmaceuticals, Inc.	Phase III trial—Fixed-dose oral combination of artesunate with pyronaridine. The course of treatment is once a day for three days. Currently carrying out pivotal Phase III studies in Plasmodium falciparum and P. vivax patients to confirm safety and efficacy. A specific pediatric granule formulation also is being tested for safety and efficacy.	$12 million
MMV/GSK mini-portfolio(five projects)	Engaged in five separate projects ranging from 1) development of next-generation pyridones derivative; 2) development of a second-generation macrolide; 3) identification of additional potent falcipains inhibitors; 4) high-throughput screening assay to study the effect of the entire GSK library of compounds on the growth and death of P. falciparum (To date, the majority of the 1.5 million compounds have been screened in a high-throughput assay, and more than 10,000 hits have so far been identified with interesting activity. The goal for 2008 is to complete the screen, characterize the hits, and use chemo-informatic technologies to cluster them.); and 5) discovery program to screen new class of compounds, namely THiQ, that showed promising activity against P. falciparum from its previous Fab1 project.	US $2.2 million
MMV/Broad Institute of MIT and Harvard/ Genzyme mini-portfolio (three projects)	Engaged in three projects: 1) screening of the broad compound collection against whole parasite assays with expansion plans in 2008 to include more compounds from the Genzyme library; 2) identification of natural products for malaria treatment; and 3) use of proteomics technology to identify molecular targets. Targets for one of the natural products have been identified, allowing it to be developed for a molecular-based, high-throughput screening (HTS) assay. Focus is to continue identifying more molecular targets that will not only be essential for parasite growth, but tractable in terms of finding small-molecule inhibitors.	$1.6 million
MMV/NITD/Novartis mini-portfolio (nine projects)	Engaged in nine projects ranging from early-stage research into identifying new targets for liver stages of P. vivax infection, through to optimization of compounds based on artemisinin dimmers. Several projects are moving forward from early-stage hits to lead compounds. One is the chemistry strategy based on successful screening of more than two million compounds from the Novartis compound collection, which led to the selection of more than 6,000 active compounds.	$589,000

As demonstrated above, the nonprofit disease organizations are having an impact on translating early-stage discoveries to development phases, not only by providing funding for proof of concept and target validation but also by providing project management and a ready-made network of scientific and clinical infrastructures to expedite and de-risk the development of novel therapies. These approaches are instructive for developing and funding early-stage development models for the stratified medicine sector, but are only part of the picture in making a business case for stratified medicine. We also must assess the clinical trial development risks and how the nonprofit disease organizations may contribute in de-risking clinical development and its applicability to stratified medicine, which will be discussed in the next segment of this paper.

Risks and Impact on Return: De-risking Clinical Trials

The critical part of assessing potential return on biomedical product development hinges on the assessment of risk factors in terms of clinical development costs, time, and success probabilities to get to market.[1] Although most venture capitalists and biopharmaceutical companies use their own valuation models to assess potential investment returns of biomedical products in development, a baseline industry average provides a snapshot of the development risk factors and possible mitigation strategies to employ through unique collaborative models with nonprofit disease organizations.[1]

Development Risk and Clinical Trial Design

With increasingly complex and chronic diseases as potential targets for new biomedical innovations, the industry is continuing to face decreasing productivity and increasing clinical trial failure rates, adding to the increase in development risks in terms of cost/time.[1] Currently, approximately 80 percent of Phase I trials are expected to fail (i.e., they have a 20 percent chance of successfully making it to market), and 70 percent are expected to fail in Phase II,[1] with expected success rates from Phase III to market between 50 percent and 70 percent. New biologic molecular entities have slightly better success rates than those identified for new chemical entities.[1]

These tools will play a significant role in de-risking the drug development process.[1] Continued advancement in new genomics-based technologies and high throughput screening tools will improve researchers’ abilities to discover reliable clinical biomarkers that can stratify and enable the discovery of the best therapies for patients.[1] For instance, use of clinical biomarkers early in the clinical trial process could help to decrease costs by identifying better responders, thereby reducing trial sample size to demonstrate efficacy and help to exclude patients early using toxicity biomarkers.[1] In addition, stratifying for key biomarkers early in the trial process not only creates the possibility of shortening end-point observation times, but also creates the ability to gather data to improve the compound or alter the trial design altogether early on, allowing for educated data mining to better define the appropriate patient population.[1] Additionally, the collection of DNA information from ongoing clinical studies, with patients’ consent, also offers the possibility to accelerate future research with increased efficiency.[1] Shorter trials with specific results also have the advantage of expedited FDA reviews, as exemplified by FDA’s review and approval of Genentech/Roche’s breast cancer treatment, Herceptin, which took six months,[1] or that of Novartis’ Gleevec, which took three months. It is anticipated that stratifying patients based on clinical biomarkers may reduce the cost of clinical trials by a factor of two to five, as it would help to narrow the test populations and commercialization time from the current ten to twelve years to five years or less.[1]

Time/Cost Correlation

The current industry expectations are the following—in Phase I of the clinical trials, twenty to eighty healthy volunteers are given a new drug compound to test for safety at a cost ranging from $8,000 to $15,000 per patient with an average time period of six months to a year.[1] In Phase II, 100 to 300 patients are given the new drug compound to assess clinical efficacy and dosage levels at a cost ranging from $8,000 to $15,000 per patient, with an average time period of two to three years.[1] In Phase III, 1,000 to 5,000 patients are tested, often in placebo-controlled, randomized, and double-blinded trials for efficacy and overall risk-benefit assessment at a cost of $4,000 to $7,500 per patient. These data sets, however, do not provide a clear picture of the real drivers of time/cost correlation. For instance, key drivers of time delay in clinical trials include difficulties in patient recruitment (this causes 33 percent to 66 percent of time delay) and data management

(8 percent to 14 percent), as well as difficulty in manufacturing and regulatory/ethics approvals,[1] resulting in upwards of 75 percent of all U.S. trials experiencing delays of one to six months or more.[1] With more than 40 percent to 50 percent of per-patient costs attributable to clinical operations, including project management, monitoring, and regulatory and data management,[1] finding ways to mitigate delays and deploying strategies to increase efficiencies in the clinical process will be critical in decreasing risks associated with development costs/time. (See Figure 8.)

Figure 8: Clinical Trial Parameters[1]

Venture Philanthropy—Clinical Trial De-risking Mechanisms

In identifying ways to de-risk the time/cost factors in clinical development, one of the emerging models is industry collaboration with nonprofit foundations which, at varying levels, offer mechanisms to expedite and create efficiencies such as readily accessible patient registries and databases, and a broad network of clinical and investigator sites that offer scientific expertise and support.

Venture Philanthropy and Patient Registry/Database

Patient recruitment in clinical trials, especially for specific disease indications, are extremely time consuming and often difficult, adding tremendously to clinical trial time/costs. One of the important de-risking mechanisms provided by the nonprofit disease organizations is access to their network of patient registries and databases. Although most organizations are at various stages of developing their patient registries, Cystic Fibrosis Foundation (CFF) has created an extensive infrastructure to serve this purpose. For instance, CFF accredits more than 115 cystic fibrosis care centers with ninety-five adult care programs and fifty affiliate programs nationwide,[1] creating one of the largest patient registry databases among U.S. foundations, with information about more than 24,000 CF patients receiving care at one of the CF care centers.[1] CFF’s database includes not only the patient contact information, but detailed information about genotypes, pulmonary function test (PFT) results, pancreatic enzyme uses, length of hospitalizations, home IV use and complications related to CF, which are critical in assessing trends and in clinical trial designs.[1] The MMRC also has developed a patient database consisting of contact information from 165,000 patients and has launched a new initiative called the patient navigator program to identify and match patients with clinical trials.[1]

Venture Philanthropy and Clinical Trial Networks

One of the critical de-risking mechanisms in terms of development time/costs that many of the nonprofit disease organizations offer is their extensive network of clinical trial sites and expert investigators, as well as information about the ongoing trials in their networks. This offers the ability to conduct multi-site trials with expediency, combined knowledge, and access to quality data from the ongoing trials. Such clinical trial networks also provide the ability to scale up quickly in Phase III studies and, in some cases, conduct Phase IV studies.[1] An important aspect about such a network is the nonprofit disease organizations’ collaborative approach to trials, as they often offer centralized review of clinical trial protocols, are able to set common policies to protect patient safety, establish standardized research procedures, share expertise among top researchers, and provide network-wide staff training.[1]

CFF may be one of the leading organizations that, through its Therapeutics Development Network (TDN), offers access to its network of eighteen clinical research centers that specialize in conducting Phase I and II studies for treatment of CF.[1] TDN centralizes and standardizes CF research while providing access to clinical trials data and CF experts through a centralized coordinating center at the Children’s Hospital in Seattle, Washington.[1] To enlarge its network, CFF invested $3 million in 2007 in forty-five new research centers in twenty states nationwide to build an infrastructure to help with patient recruitment and to increase its clinical network.[1] As discussed previously, the MMRF also offers a network of fifteen academic centers that collaborate in conducting multi-site clinical trials.[1]

To increase efficiencies, productivity, and sustainability of conducting clinical trials in developing countries, MMV works with a network of international organizations such as the Malaria Clinical Trials Alliance (MCTA), Malaria Vaccine Initiative (MVI), and the INDEPTH Network.[1] MCTA facilitates site preparation for effective conduct of Good Clinical Practices-compliant trials for malaria vaccines and therapies, while supporting the long-term development and sustainability of clinical trial sites in nine countries across Africa (Mozambique, Tanzania, Malawi, Gabon, Nigeria, Ghana, The Gambia, Kenya, and Senegal).[1] MVV also works with the European & Developing Countries Clinical Trials Partnership (EDCTP) to facilitate Phase II and III clinical trials in HIV/AIDS, malaria, and tuberculosis in sub-Saharan Africa.[1]

Assessment and Recommendation

To reverse the trend of declining productivity and innovation, and embrace the new technological and scientific advances that will allow for safer and more effective treatment of diseases through stratified medicine, industry stakeholders must be open to unique models that could de-risk current drug development processes and increase their combined probabilities of success. Through our discussion, we have identified new collaborative mechanisms with nonprofit disease organizations that can not only help bridge some of the funding gaps in early-stage discovery and development of new technologies, but more importantly, de-risk the clinical process in terms of time and costs.

**FOUNDATIONS “DE-RISKING” PROCESSES**
Foundation	Academic Research Networks	Clinical Centers Networks	Tissue Banks	Patient Registries	Project Management
Cystic Fibrosis	Yes	Yes	No	Yes	No
Multiple Myeloma	Yes	Yes	Yes	Yes	Yes
Myelin Repair	Yes	No	Yes	No	No
Juvenile Diabetes	Yes	Yes	No	No	Yes

In the short term, these mechanisms offer a model for the biopharmaceutical industry in how they can better work with existing nonprofit organizations to capitalize on their offerings. The elements of such a model would include biopharmaceutical companies collaborating with other groups, such as nonprofit foundations, who could establish and manage the programmatic research of networks of academic and investigators from small biotechnology companies, patient registries, and expert clinical centers. In return, large biopharmaceutical companies would provide some funding and commitment to take over the late-stage development of “de-risked” clinical candidates to approval and marketing. There could be several innovative ways to reward the nonprofits for their contribution without violating their mission or 501(c)(3) status.

A critical success factor in stratified medicine is the discovery of the biomarker and/or diagnostic kit. Intellectual property rights can be a potential barrier when there is only one supplier of the diagnostic kit, particularly if that kit has not been approved by regulatory bodies. This presents challenges in reimbursement, as well as potential liability issues if such a kit is used to qualify patients for a drug, and the specificity and sensitivity of the diagnostic test have not been established. This liability exists for both tests of efficacy and susceptibility to side effects. There is, therefore, need to address this downstream issue of potential biomarkers that are discovered in the NIH and other Biomarker consortia.

In summary, this paper focuses on ways to address two of the issues—return on investment and probability of success—that are barriers to the adoption of stratified medicine by large biopharmaceutical companies. The various activities of some foundations serve to identify the relevant patient subgroups and generate data to better qualify potential drug candidates. We call these “de-risking” activities, which not only fill gaps in funding, but improve the probability of success of the drug discovery and development effort. These diseases also are excellent examples where subgroups of patients might be discovered and stratified, and prospective health care—anticipation, prevention, intervention—as described by Dr. Ralph Snyderman, could be pursued on a more rational basis. Thus collaboration between large biopharmaceutical companies and disease foundations provides an interesting model within which several aspects of the development and implementation of prospective health care and stratified medicine might be assessed for technological and economic feasibility.

However, a broader challenge remains in the ability to scale these de-risking mechanisms to a larger set of disease sectors, and on the question of who will bear the cost of creating the necessary infrastructures. One possibility is the U.S. government; as such efforts would be consistent with both the FDA’s Critical Path Initiative and the NIH Road Map. We would argue that both the FDA and NIH, under these two initiatives, could encourage the collaborative model suggested in this paper, either by disease category, such as cancer, where there is a known familial or genetic predisposition for the disease. In addition, two areas need to be urgently evaluated or assessed: the barriers that present intellectual property rights pose to adoption of the collaborative model, and the financial value of the varying de-risking strategies that we have discussed. These are the questions we pose today in opening the discussion on how we can make a business case for the growth and adoption of stratified or personalized medicine in the near future.

*With special thanks to Lauren Choi, Counsel, Buchanan, Ingersoll & Rooney, for research and editorial assistance.

Return to Table of Contents

Investing in Personalized Health Care Innovation

M. Kathleen Behrens
General Partner
RS& Co. Venture Partners, IV, L.P.

Risa Stack
Partner,
Kleiner Perkins Caufield & Byers

Bruce Quinn
Senior Policy Specialist

Foley Hoag

Kelly Slone
Director,
Medical Industry Group
National Venture Capital Association

Background and Purpose

Over the last decade, a series of key research developments in the fields of genetics and medicine have enabled the possibility of tailoring treatments for patients based upon the molecular basis of disease and/or the individual’s ability to respond to a specific treatment. This possibility has given rise to the emerging field of personalized medicine.

Personalized medicine represents a major leap in the evolution of healthcare because it enables care providers to deliver the right treatment to the right patient at the right time. This ability will not only lead to improved health outcomes and better qualities of life both during and after illness, but may also help lower costs through greater efficiency of treatment.

Much in the same way that it helped create the biotechnology industry through its investments in Genentech, the venture capital industry has played a critical role in driving the development of personalized medicine by helping to translate multiple breakthroughs in molecular medicine technology into marketable products. Venture-funded companies like Genomic Health, Inc., Monogram Biosciences and Adeza Biomedical Corp. have already made an impact on patient health. The next generation of companies is expanding into new therapeutic areas, some of which are utilizing novel technology platforms. Venture capital will likely remain the primary source of financing for young innovators in this space due to the extraordinary risk associated with investing in healthcare technologies.

Despite its enormous promise, personalized medicine faces a number of barriers at what has become a critical point in its development. Some of these stem from current regulatory policies and the uncertain reimbursement outlook for new technologies. Others have resulted from the recent turmoil in the capital markets. For each of these variables, even minute fluctuations and adjustments can alter the risk profile for even the most promising technology. Working in concert, they can price risk beyond levels acceptable even to venture capitalists – effectively stunting the development of emerging technologies and undercutting the incentive for future innovation.

Within these contexts, the purpose of this paper is to:

Explain the role that venture capital plays in innovation across industries.
Illustrate the general process of venture investing.
Outline the challenges and risks specific to healthcare investing.
Articulate venture capital’s vision for advancing personalized medicine within this context.
Examine current and potential business models within personalized medicine.
Discuss personalized medicine’s implications for healthcare delivery.
Identify current barriers to the development of personalized medicine.

Venture Capital: A Key Force in Innovation

The venture capital industry drives innovation by turning ideas and advances in basic science into marketable products and services that improve people’s lives. They do this by identifying the most promising advances and then guiding their commercial development with expertise and funding.

Typically, venture capital concentrates on funding innovations that threaten to replace – or render obsolete – established products and services in a given marketplace. Venture capitalists use their expertise to find such “disruptive” technologies and evaluate which ones have the most market potential. Only the most promising advances get funded, and the venture capitalist typically takes an active role in guiding further development.

In this sense, the venture capital industry creates and maintains a de facto research and development pipeline for a wide variety of technology and knowledge-based industries. This role has become critical in recent years, as many public companies have slashed R&D budgets as a way to ease the quarterly scramble to meet earnings estimates. In some cases, venture capital has created entirely new industries.

The early-stage risk associated with disruptive innovations often precludes financing from traditional sources such as banks and public equity. Without someone to step in and assume this risk, many promising advances would have no capital for further development. By providing funding and expert guidance during this critical period, venture capitalists ensure that the most promising technologies have the best chance of making it to market – where they can make the greatest possible impact on quality of life.

By driving innovation in these ways, venture capital investment has fueled the development of many high-tech industries in the United States. These include biotechnology, medical devices, network security, on-line retailing, RFID and Web-based services. In fact, venture capital has helped to build innovative powerhouse companies such as Genentech, Microsoft, Google, Apple, Starbucks, Staples, eBay and FedEx.

Venture Capital Funds and the Investment Process

The first venture capital firms date back to the 1930s and 1940s.[1] Up through the 1970s, by which time venture capital had established itself as an industry and a profession, investors at these firms worked as “generalists” – investing in companies and ideas across various industries. The relatively small size of the industry permitted this lack of specialization. In 1980, for example, the National Venture Capital Association tracked only three sectors: life sciences, information technology and industrial/energy.

Beginning in the 1980s, the skill set required for venture investing grew along with the industry. Successful venture capitalists needed the industry experience and insight to determine which innovations offered the most promise, domain business expertise to advise entrepreneurs and technologists in building portfolio companies, and the experience to manage the multi-stage investment process for those companies. As a result, venture investors (and often their funds) began to concentrate their efforts on a few sectors in which the partners had the most experience and that offered the most opportunities to innovate – most commonly software, industrial/energy, biotechnology, medical devices and diagnostics, and media and entertainment. Today, an individual venture capitalist typically specializes in one or two related sectors -- e.g. biotechnology and medical devices.

The venture capital process

Venture capital firms raise funds from investors – most commonly large institutions such as corporations, foundations and public entity pension plans, but also from individual investors. The partners in a venture firm also invest their own money in their funds. Venture fund managers are compensated through annual fees associated with managing fund investments, as well as a percentage (carried interest) of the profits derived from successful investments. The latter are offset by losses associated with unsuccessful investments.

Venture capitalists generate profits and losses from the funds they raise by making equity investments in a range of portfolio companies. The average timeframe for this process is typically seven to 10 years. Generally, venture capital is used to purchase an equity stake in a series of rounds of investments in each individual company. Because venture capitalists tend to make investments in young companies, those companies often do not have products or services to generate cash flow from operations. As a result, they are not sufficiently creditworthy to take on debt. For this reason, venture capital is known for taking the highest risk in the spectrum of stages (with the exception of the “angel” stage) in which capital can be invested in building and running companies. Venture investors do not typically loan funds to their portfolio companies unless it is provided on a short term basis and eventually gets converted into equity.

As is the case with most investing, venture funds require risk diversification—especially considering the fact that they invest in one of the highest-risk stages of investing and often must wait 10 years or longer to realize their returns. To mitigate this risk, some venture firms raise separate funds dedicated to specific industries, while others specialize by investing in a single industry with multiple segments that can contribute some risk diversification.

Most young companies raise capital in a series of investments. Venture investors can participate in some or all of these financings. Young companies need only small amounts of capital (relatively speaking) when starting out – perhaps to develop a “proof of principal” or to reach a benchmark demonstrating measurable progress with a product or service. This enables them to raise capital in a series of later stages, at which points they can achieve higher valuations than in earlier rounds and can sell less equity – i.e. experience less “dilution” from the price at which they sell/accept new capital. This also allows the venture investor to invest more capital in the company, but in a way that demonstrates a series of increasing valuations associated with favorable progress. More detail about this process can be found in Appendices A and B.

After purchasing equity in a portfolio company and nurturing it for many years, a venture capitalist generates a return for the investors in the fund by selling that equity. There are multiple ways to generate liquidity for these equity investments. Most commonly, the company sells equity in the public market – enabling the venture investor to sell all or some of its stock to public market investors – or the company is sold to another firm. In the latter case, the investors receive either stock or cash upon the sale of the investment, thus providing either immediate or eventual liquidity. Again, this typically happens between seven and 10 years after the initial investment.

Venture Capital Investing in Healthcare^[1]

Venture investors generally assume significant technology development risks; however, healthcare presents some unique and additional challenges. These largely relate to the added complexity of long product development timeframes (often associated with clinical trials), government regulation and significant capital requirements, as well as the complexity of reimbursement associated with the healthcare payor system. These factors add up to larger capital requirements on the part of venture capitalists (and other stakeholders) and an investment time horizon that stretches to 15 years or longer. All of these increase risk.

Despite these complexities and the additional patience required, the venture capital industry invested approximately $9 billion, or 30 percent of its total, in companies in the biotechnology (including pharmaceuticals and research tools companies), medical devices and healthcare services (including healthcare information technology) sectors in 2007.

Healthcare investing by the venture capital community for many years has been concentrated in the three areas mentioned above. During this period, the degree of specialization required for successful investing in each has increased.[1] In large part, devices and biotechnology investments have been tied to new technological developments and related venture expertise, while services have met the needs of evolving healthcare delivery with a different set of skills and experience. Approximately 98 percent of the venture capital invested in healthcare (as measured by aggregate invested capital and number of investments) has been devoted to biotechnology and medical devices.

Venture capital and biotechnology

The venture capital industry played a critical role in creating the biotechnology industry during the 1970s and 1980s. During this period, researchers achieved a number of key advances in the fields of gene sequencing and expression technology, recombinant DNA technology and monoclonal antibody technology. (Not coincidentally, all of these provide the foundation for personalized medicine today.) At the time, however, there existed no process for translating these advances into commercial products. In these years, for example, when Cetus Corp. developed polymerase chain reaction (PCR) and Genentech began cloning insulin, established pharmaceutical companies simply weren’t inclined to take the risks involved in funding these advances. In both cases, venture capital stepped in to provide funding and management – helping both companies to advance their product development and creating a modern blueprint for building successful companies from innovations in medical technologies. Venture capital played a similar role during this time with biotech pioneers Amgen, Chiron, Biogen Idec and Genetics Institute, LLC.

Today, business models for biotechnology companies developing biopharmaceutical products continue to rely upon significant capital from both the venture industry and the public market. The capital is needed to sustain long product development cycles required for research and clinical studies, as well as manufacturing and product launch. Estimates vary for the cost of individual therapeutic products due to amortization of product failure costs along with successes; nonetheless cumulative individual product expense estimates range from $500 million to nearly $1 billion.

More so than in other industries, the risks associated with the increased costs and extended timeframes in the biotechnology sector preclude traditional sources of financing in the early stages of development. At the industry level, venture capitalists step in to not only fill the funding gap during the early stages, but also vet companies’ scientific platforms and assess their commercial viability. They also lend management expertise and strategic vision to the companies in which they choose to invest.

Venture Capital Facilitates a New Vision for Diagnostics and Personalized Patient Treatments

As described earlier, an important role of the venture capitalist is to help facilitate change within an industry. In the healthcare industry, the venture capitalist acts as an advocate not only for the entrepreneur but also for the patient by helping to drive advances in patient care – as demonstrated by the commercialization of advances in DNA research into biologic drugs (described in the previous section) and other treatments. Thanks to continued research and investment, these advances have in turn spawned personalized medicine, which uses these technologies to improve patient outcomes and potentially reduce costs in the long term.

Vision for transforming role and value of molecular diagnostics

Historically, diagnostic products have provided incremental and supplemental information to physicians and patients in managing care largely because they have contributed additional pieces of information to assemble into an overall assessment of disease and treatment.

New molecular diagnostics (or, personalized medicine) have the ability to raise the importance and value of the information derived from testing. Newer technologies enable the collection of data and analyses at a scale that was not previously possible – providing new insights about patients and diseases that can inform patient care and treatment. Also, new-generation tests may provide critical information for patient management, which in some circumstances may be as or more important than the value of existing therapies.

Technologies contributing to personalized medicine progress

The foundation of personalized medicine lies within our efforts to better understand the biology of disease at the genetic and protein levels. Three technologies at the center of this effort are gene sequencing, gene expression and proteomics. Gene sequencing, made possible initially by Cetus’ development of PCR, enables researchers to clone DNA and thus amplify genetic material. Gene expression technology allows researchers to identify genes in patients that indicate the presence of, or an increased susceptibility to, a given disease. In addition, it also helps illuminate the development and growth of cancerous tumors. Proteomics examines the molecular biology of diseases, enabling researchers to identify individual proteins associated with specific disease states and susceptibilities.

All of these technologies – either in development or application – have been informed and/or advanced by the Human Genome Project, which generated large quantities of genetic and genomic information and helped enable the acceleration of the sequencing process. Researchers have then studied this information in great deal to better understand the link of genetics to diseases; included in these efforts are large scale studies of both patients and agents of disease.

Personalized medicine: Understanding the patient and understanding the disease

At the most basic level, personalized medicine has two goals: understanding the molecular nature of a disease and understanding how an individual will respond to therapy.

One of the best-known examples of the former is Herceptin^®, a drug developed and marketed by Genentech, one of the first biotechnology companies to emerge from the venture industry. The recognition of the role of HER2 over-expression in breast cancer patients has aided in patient selection for treatment with Herceptin^®. An important advance in understanding the patient’s response to therapy is the ability to assess thiopurine (a group of chemotherapeutic agents) drug metabolism by measuring thiopurine methyltransferase (TPMT) activity in an individual patient. This advance has enabled physicians to identify which cancer patients are likely to suffer adverse effects from the treatment. However, the opportunity to bring applications of more recent technological developments to bear in personalizing healthcare is driving venture capital investors to start new companies in this field today.

The business of personalized medicine

Venture capitalists play a key role in building personalized medicine companies. They work with entrepreneurs to craft the business strategy, recruit the management team and often catalyze the key relationships necessary in building the business. While there is no formula for success, following a thoughtful process that addresses key issues increases the likelihood of success. Below are key early considerations in building a personalized medicine business:

Clinical situation in which a physician needs more information to make an important treatment decision (for example, administer a life-saving therapy or device).
Attractive target market from a business perspective; the ideal opportunity involves a large market coupled with a key clinical decision requiring a potentially expensive treatment (for example, placement of an implantable cardioverter-defibrillator (ICD) or administration of a biologic drug).
Patient samples and technologies that can be used to address the clinical need. Often initial studies need to be performed to determine if together the samples and technology enable the development of the diagnostic.
Management team with the appropriate scientific, clinical and regulatory expertise.

While the above factors may seem relatively straightforward, all components must come together to build a successful business. One factor of particular importance is the availability of well-annotated clinical samples, which makes development more practical. Such samples can reduce the risk of getting to clinical trials, the cost of development and the duration of development. Consider this example: Genomic Health’s Oncotype DX^® was developed using tissue archives that include data on each patient’s five- or 10-year outcome. Had this archive not been available, development would have required new tissue samples and a waiting period of five to 10 years to track the clinical course of the patient.

Personalized medicine is reality: current products

As described above, the goal of personalized healthcare or medicine is to tailor treatments for patients based upon their individual medication responses and the molecular basis of disease. This practice is evolving by addressing groups of patients that can be categorized by having similar susceptibilities and responses to therapy—in effect stratifying them by risk and response. While it is possible to envision a very broad range of future applications from these types of assessments, below are some examples of innovative personalized medicine companies funded by NVCA members. While the disease area may be different, these companies all have a common goal, providing answers to key clinical questions enabling better patient management.

Early identification for individuals at risk for a disease. Genome scans identify genes which are associated with risk of disease, such as diabetes and osteoporosis. Currently, tests are available to consumers that identify these genes in individual carriers; two companies that market these tests are Navigenics and 23andMe. Understanding the genetic risk enables individuals to make lifestyle changes that may reduce that risk. Other tests detect patterns of markers in the blood that are early indications of potential disease. Tethys Bioscience has developed a test that quantifies the future risk of diabetes in patients who are at risk, but do not yet have clinical diabetes.

Prognosis. Two recent examples of venture-backed companies with such products include XDx, which markets the non-invasive Allomap^® test for heart transplant patient management to measure acute cellular rejection, and Genomic Health, Inc., which sells the Oncotype DX^® test for breast cancer patients that assists in predicting the possible recurrence of disease in those with early-stage, invasive disease.

Response to therapy. The recent introduction of the new Trophile^TM assay, by Monogram Biosciences, identifies which HIV patients are most likely to benefit from Selzentry^TM. Another company, ARCA Biopharma, Inc., has identified subpopulations of heart failure patients presumed to have greater therapeutic benefit from bucindolol, a new beta-blocker, based upon certain genetic subtypes. Individuals needing treatment for heart failure would be given a genetic test to see if they qualify for the new treatment. These types of applications are important – especially in those cases where a patient can be spared the significant side-effects of a given medication that actually offers a low likelihood of response. They are also very useful in rescuing otherwise failed drugs in clinical trials now and perhaps in the future, when safety problems arise.

Given that there is no shortage of clinical situations in which physicians could benefit from more information, many opportunities for the development of future products exist. The next generation of personalized medicine companies will continue to expand product offerings in oncology and cardiology, infectious disease and woman’s health. Young companies are also tackling new frontiers, such as autoimmune disease, neurodegenerative disease and psychiatric disease.

Future business models

Potential Model 1: Avoiding adverse effects. A future healthcare system could use electronic health records to identify patients with adverse effects, enroll patients (both with and without the adverse event) in research studies and screen for genes or biomarkers associated with the adverse event. This is currently not practical, nor is it likely be in the very-near future – given that all patients would need to be tracked for all adverse events and all samples would have to be data-mined for genomic or proteomic correlations. However, today there is no system at all in place to call out the most likely targets for adverse effects research or to signal where or how payors will pay for cost-effective diagnostics that protect patients from prescriptions that will hurt them. Based on the risk and investment principles described so far, an optimal combination of public investment could be balanced with rewards for specific product development in the private marketplace. A model example would be the basic research that took place in warfarin pharmacogenomics, which discovered several genes playing a key role in warfarin metabolism. A number of private companies are now competing to market increasingly rapid tests for these genes.

Potential Model 2: Safety alert or early intervention systems. Modern electronic health record technology could incorporate existing knowledge of adverse event reactions and drug/drug interactions in an “interactive health record” that incorporates patient specific information, data-based best practices, and laboratory test results in real time. This could provide optimal treatment pathways and ongoing appropriate tests to anticipate and reduce adverse events and to ensure the optimal treatment of the patient. One venture capital-based firm, Proventys, Inc., is currently developing important tools in this space. New business models will develop the best marketplace strategies for this category of personalized health technology, such as interfaces with pharma management systems, physician offices and payors. This space will be accelerated by policy initiatives such as adoption of ICD-10 disease codes, and electronic health records that facilitate rapid (ideally, immediate) transmittal of key information between the billing entities such as physician offices, hospitals and specialty laboratories. The Harvard Medical School/Partners Healthcare Center for Genetics and Genomics (HPCGG) has emphasized that the current relationships among these entities are many-to-many systems, making the information technology problems impossible to solve. The simplest solution, which does not exist today, would be a central data repository which can be accessed with appropriate confidentiality protections and permissions to result in the development of much more sophisticated solutions.

Personalized medicine and implications for the healthcare delivery system.

As many of the preceding examples suggest, personalized medicine has the potential to transform the way care is delivered to patients across a full spectrum of conditions. To date, the healthcare system has caught only a small glimpse of the clinical and economic outcomes personalized medicine can yield. How soon these benefits can be fully realized depends on how quickly and effectively the healthcare industry can overcome the challenges inherent in harmonizing the interests of its multiple stakeholders. Clinicians, payors, manufacturers and health information technology firms alike will have to play meaningful roles in enabling innovations to fit into the context of the marketplace and achieve acceptance on a large scale.

The following provides a brief overview of challenges that key stakeholders face as the growth of personalized medicine accelerates in the coming years:

Clinicians: At the center of decision making. Through advances in personalized medicine, clinicians will be empowered to more precisely diagnose a patient’s condition and select the safest and most efficacious treatment based upon the patient’s unique clinical profile. However, the adoption of new technologies will pose a considerable challenge in the context of today’s busy medical practice. Among the challenges are (i) keeping pace with the proliferation of personalized medicine technologies (and the vendors providing them); (ii) identifying which patients are appropriate for the various technologies being introduced to the market; (iii) interpreting molecular diagnostic test results in the broader clinical context for their patients; (iv) processing the sheer volume and complexity of data to make personalized clinical decisions; (v) reviewing and understanding the scientific evidence supporting the new technologies prior to relying upon them in routine practice; (vi) understanding the various payor coverage determinations to ensure appropriate reimbursement; and (vii) implementing the necessary operational processes for handling biological samples and working with various personalized medicine vendors.

Payors: Driving value-based approaches. As current medical management strategies such as disease management mature, payors are seeking innovative solutions to help reduce variability in care and control the rise of medical costs. It is in the best interest of payors to support personalized approaches to care where better understanding of patients’ individual profiles (including their risks of an adverse event or potential response to a given therapeutic path) will guide better clinical decisions Payors can play an important role in shaping the emerging market for personalized medicine by (i) identifying the clinical areas of greatest unmet need through population-level medical claims data analysis; (ii) setting clear requirements for the technology validation necessary to secure reimbursement coverage; and (iii) supporting the appropriate utilization of emerging technologies through the implementation of novel, value-based reimbursement models.

Diagnostic and biopharmaceutical manufacturers: innovators and educators. Personalized medicine represents a significant paradigm shift for both the diagnostics and biopharmaceutical industries. Biopharmaceutical manufacturers must reassess the fundamentals of their business as they contemplate the attendant shift from discovering the next blockbuster drug to unlocking the enormous value of targeted therapeutics that serve more distinct and segmented populations of patients. Diagnostic manufacturers must effectively demonstrate the increased value of their technologies as they play a more central role in the personalization of care. These manufacturers will help accelerate the acceptance of their own innovations by (i) effectively validating their technologies to support both payor reimbursement and clinical adoption; (ii) educating clinicians in novel ways with sound scientific support to ensure the appropriate utilization of these new technologies; and (iii) investing in the ongoing innovation necessary to establish a sustainable personalized medicine market.

Health information technology (HIT) vendors: Vital enablers. For personalized medicine to evolve from the current discrete instances of esoteric testing and targeted therapeutics to a more sustainable and widely accepted approach to care, a foundational system of information technology is required. HIT vendors have a unique opportunity to provide the dynamic, point-of-care decision support necessary to support the broad adoption of personalized medicine. These vendors must (i) develop more robust information solutions focused on delivering high-value decision support that empowers clinicians; (ii) make data more accessible and actionable to the care team within their current workflow; (iii) establish effective approaches to health information exchange to allow for a comprehensive view of a patient’s medical history; and (iv) work with clinical data in novel ways to spur innovation while ensuring patient privacy and data security.

Barriers to Personalized Medicine Innovation

In prior sections, this paper illustrates the considerable extent to which innovation relies on the ability of entrepreneurs and technologists to develop products from advances in research and commercialize these products. This section focuses on the complexity of this task in the field of personalized medicine and the barriers to success that currently exist.

As described earlier, some of these barriers exist inherently in healthcare investing. The longer time horizon and increased capital commitments necessitated by complex product development and clinical trials provide two such barriers. Market-driven fluctuations in the availability of capital provide a third. While these economic barriers are simply part and parcel of the process, a number of policy issues –specifically with regard to regulation and reimbursement – also hinder the development of personalized medicine. These barriers are formidable and urgent, yet also within the government and the industry’s power to mitigate – if the two groups work together now to remove them.

Laboratory medicine deals with a complex transaction system

Successful company development in personalized medicine involves simultaneously balancing:

Extremely sophisticated technologies, such as gene chip microarrays.
The complex nature of medical knowledge (clinical trial environment; human subjects regulations; the need for evidence-based medicine).
The complex decision-maker marketplace (e.g. physicians; standards of practice; society guidelines; accepting innovation versus the tried-and-true in medical technology).
The complex payor marketplace (coverage decisions by insurers as well as the underlying coding/reimbursement system).

A laboratory medicine test faces this transaction system:

Figure 1

Figure 1: A laboratory medicine test transaction system

Figure 1 shows arrows for just one payment pathway (in this case, a government payor). Other arrows would connect different payors to the laboratory. Note that the personalized medicine lab ultimately receives money originating from one of four sources (represented by rectangles): patient self-pay, from taxpayers via a government payor, or from an individual or an employer via private insurance.

The diagram risks understating the complexity of the information transactions involved.[1] A national lab – even a one-test startup lab – must deal with hundreds of private and government payors. Each payor must make coverage and pricing decisions (which can involve complex technology assessments) about the test and each payor (at least in principle) needs to know something about the patient’s condition at the time of the test.

The arcane complexities of the insurer coding and pricing systems for laboratory tests have been well-documented.[1] During the early investment stages, the venture capital firm must project five to 10 years in the future how physicians, patients, hospitals, private payors, government payors and the self-pay patients will respond to the test, as well as the test’s likely position in the marketplace relative to both existing alternatives and alternatives likely to be introduced in coming years. All of these factors must be tracked during, or informed by, optimally planned and staged investments which lead to the most efficient reduction in risk as the project develops and investments increase. (As shown earlier, with the reduction in risk, the projected value of the company increases substantially, which in turn makes a new investment round possible.)

Venture capital investors must evaluate the likely stances of third-party payors closely – a necessity that is very specific to the medical technologies sector. Generally, payors are concerned about two issues: 1) the overutilization of diagnostic tests and treatments and 2) the absolute costs of these tests. Thus, the challenges for venture capitalists regarding this new generation of molecular diagnostic tests stem from the fact that the development process is costly, as like theraputics, they often involve the productization of new technologies and large clinical trials. These factors in turn drive up prices for patients and payors.

However, entrepreneurs and developers of these technologies are willing to risk the concerns of payors because the results generated by them provide information of much higher value for patient care. Therefore, despite the initial higher cost, these diagnostic tests will ultimately lead to more cost-effective patient management for payors.

Specific reimbursement considerations

As discussed previously, the current products that are most strongly associated with personalized health care are molecular diagnostics. Today, far more is known about the molecular heterogeneity of major diseases, including cancer, than was known even 10 years ago. Research has made it clear that targeted and more effective medical treatments will often be unattainable unless physicians have precise molecular information about each patient’s disease. That is, there will be no “magic bullet” chemotherapy for “colon cancer” across all patients, but there may be a very effective treatment for those patients whose colon cancer expresses Gene X.

In many ways, these tests seem to be the easiest to integrate into the existing care delivery system. If Chemotherapy Drug X is effective when tumors express Gene X, then we can test those patients and prescribe the right drug to the right patient at the right time.

Although the integration of these tests into clinical care would seem like a fairly straightforward process, companies and investors have found two key factors providing barriers to innovation. These are 1) level of evidence for payor coverage and 2) legacy pricing systems.

Diagnostic tests: level of evidence for payor coverage. Payors are most experienced at performing technology assessments for drugs and for other therapeutic interventions (e.g. ultrasound for kidney stones.) Diagnostic tests present several difficulties for payors. First, few payor guidelines for technology assessment contain the same level of sophistication and granularity as the research that led to and supports the technology being assessed. [1] Some guidelines don’t even recognize the difference between diagnostic and therapeutic applications. Second, there are few guidelines on the degree to which clinical benefit can be extrapolated from test accuracy or retrospective studies, or whose extrapolation is credible and why. For example, imagine a researcher studies Gene X in an archive representing 100 colon cancer patients treated with drug XYZ. Only the 20 with Gene X responded, and responded well; the other 80 did not respond at all and quickly died. Should a randomized trial be conducted, where 80 percent of entrants will be treated with XYZ despite having Gene X? Can we assume Test X is necessary and impacts clinical care greatly? What if the numbers were less clear cut? The lack of consensus guidance leaves both innovators and payors with a great deal of uncertainty in how to evaluate diagnostic tests for coverage.

Diagnostic tests: reimbursement. Most traditional diagnostic tests long ago became commodities (such as a serum glucose test or a thyroid hormone test). Most payors pay fixed and inexpensive test prices related to Medicare’s laboratory fee schedule, which was established in 1984. Since then, many stakeholders have asserted that the reimbursement environment for novel diagnostics is much more challenging than the environment for other medical devices or drugs.[1] In the U.S. payor system, new drugs are assigned specific codes for insurance claims and paid at market prices that are set relative to alternative drugs. The payor system for diagnostic tests has developed in a different and less consistent manner. Diagnostic tests are usually described by generic codes (e.g. microbiology antibody test) and paid at a fixed rate (say, $30). In the case of one novel molecular medicine test (the Oncotype DX^® test), however, private insurers and Medicare have paid near list price – several thousand dollars in this case – for the test. High levels of uncertainty regarding “value-based pricing” of molecular diagnostics pose serious difficulties in the venture capital investment model because such uncertainty inverts the assumption of progressive risk reduction (the notion that a venture becomes less risky as it matures) outlined in prior sections. For example, in the case of a novel molecular test, the uncertainty over how Medicare will price it resides at the far end of the development and investment pathway; this uncertainty remains constant during progressively larger staged investments. Prices that converge on marginal costs are characteristic of mature and highly competitive markets, but make innovation impossible.[1]

Changes, such as published guidance, which make coverage and reimbursement more predictable will reduce the overall level of risk for innovators and thus encourage innovation in ways that are otherwise costless to the system.

Specific diagnostic oversight issues

New molecular diagnostic tests primarily fall into classes of products defined by the Food and Drug Administration (FDA) as in vitro diagnostic (IVD) tests or in vitro diagnostic multivariate index assays (IVDMIAs). The former are often less complex than the latter, though precise regulatory definitions are still being actively reviewed as new products reach laboratories and the FDA.

Historically, oversight responsibility for in vitro diagnostic products has resided in both the FDA and the Center for Medicare and Medicaid Services (CMS). FDA approves the safety, efficacy and manufacture of IVDs under its authority to regulate medical devices, while CMS oversees compliance with performance standards for laboratories under the Clinical Laboratory Improvement Amendments of 1988 (CLIA). For many years, the FDA has exercised “enforcement discretion” for emerging diagnostic tests largely performed by specialized laboratories. However, in July 2007, the FDA published its most recent guidance on the topic, signaling the agency’s intent to actively review all current and new IVDMIAs that have not already been voluntarily submitted for review. This guidance has sparked significant dialogue within different industry groups, as well as between industry and the FDA.

The FDA’s regulation of IVDMIAs will follow the device evaluation path--where substantial equivalence to an existing, or “predicate”, device is established for the new test, or a more extensive, pre-market approval (PMA) path may be required. A number of new molecular diagnostic tests have been developed as Laboratory Developed Tests (LDTs, or “home brew” tests) for exclusive use in CLIA-certified laboratories and have not been reviewed by the FDA. The July IVDMIA guidance, when implemented, will require these tests and others under development to undergo more extensive regulatory evaluation, in some cases requiring clinical data that may or may not have been developed prior to their use as LDTs.

There are number of outstanding issues associated with new IVDMIA development and the FDA’s plans to actively review this class of tests. The outcome of this review could significantly influence innovation by the venture community. These include, but are not limited to:

Developing a clearer understanding of what constitutes an IVDMIA (i.e. the definition).
Identifying what types of data will be required for favorable claims reviews (i.e. retrospective versus prospective clinical trials).
Understanding the evaluation process for new algorithms integrated into test design and analysis.
Clarifying the roles for CMS and the FDA in dual regulation of this class of devices.
Providing a reasonable transition time for tests currently marketed as LDTs, as well as those under development.

As is the case with many new technologies, patient safety concerns must be balanced with regulatory processes. This must be done without defeating innovation, however.

This discussion has been devoted to the oversight of new molecular diagnostic tests largely for two reasons. First, the vast majority of venture-backed companies that are focused on personalizing healthcare are developing these types of products. Second, the issue is timely for the venture community as it weighs ongoing investment in companies developing new tests. However, regulatory decisions are also evolving for pharmaceutical product development--ones that anticipate incorporating new molecular technologies. For example, the FDA is discussing plans to require the collection of DNA samples from all patients participating in clinical trials, so that such material can be accessed in the future if drug-related safety issues arise. Further guidance from the FDA would also be constructive in updating the preliminary regulatory path for co-developed diagnostic and therapeutic products.

Conclusion

Thanks to continued federal funding and the extraordinary promise for improving health outcomes, advances in the fields that drive personalized medicine will continue. Demand for treatments and therapies based on these advances will also grow as people begin to understand aspects of their personal health in unprecedented detail and look to take greater control over that health. Given these realities, the question becomes: Will the industry be able to meet this demand by bringing advances in personalized medicine to the marketplace?

The fundamental process for bringing innovations in this sector to market probably won’t change. Federal and academic research will continue to move the science of personalized medicine forward. Innovation will continue to spring from small companies – as opposed to large institutions and corporations – because of the freedom and creativity they encourage. Venture capital will continue to step in during the critical early and middle stages to assume the risks inherent in building these companies.

Unfortunately, venture capital can only take personalized medicine and the innovative companies that drive it so far before the acute regulatory and reimbursement barriers discussed in the previous section begin to hinder development. The consequences of this inefficiency are significant – given personalized medicine’s potential for dramatically improving both the efficacy and efficiency of healthcare delivery. Together, these elements could play a major role in broader healthcare reform in the U.S. by reducing costs and enabling greater individual participation in health outcomes. Without a joint effort by government and industry players to remove or ease existing barriers, however, personalized medicine may never achieve its full potential.

APPENDIX A

Guide to venture capital investing rounds and terminology

Most young companies raise money in discrete stages. This practice enables the owners and their venture investors to raise funds at increasingly higher levels of valuation as the company’s assets grow and its risk profile improves. (See Appendix B for step-by-step details.)

The earliest round of financing is typically called either a “seed”, “first”, or “Series A” (denotes the legal name for the category of stock and investor) round. In most cases, it represents the first time that a company raises funds and usually garners a small amount of capital (i.e. between one and-several million dollars) from only one or a couple of investors. Funds raised during this round may contribute to product development and market research. Other uses include building a management team and developing a business plan if the initial steps are successful. This is a pre-marketing stage.

Subsequent rounds are called “follow-on” rounds – typically named “Series B, C, D” and so on. These rounds generally draw down larger amounts of capital from an increasing number of investors as the company’s needs grow. Series B capital often funds additional product development, product launch and initial marketing efforts. Once a company is producing and shipping its product and has growing accounts receivables and inventories, Series C capital may provide funds for an initial expansion. Beyond this point, the company may engage in additional rounds, or even begin to take on some debt and possibly sell equity to public market investors. Such late stage rounds are commonly called “mezzanine” rounds.

While many venture funds invest in both the early and follow-on rounds, some also specialize in the stage at which it makes its investments. For example, the unusual expertise and operational experience required for creating companies from scratch have given rise to funds specializing in seed round investment. Similarly, mezzanine rounds often call for funds that focus on taking companies public and/or selling them to other companies. Such funds are commonly affiliated with public market investor funds. Other funds, known as “crossover” funds, may specialize in investing in both the late private rounds of investment as well as the public market (although most venture funds in the healthcare space reserve the ability to make investments in their portfolio companies in both private and public rounds of investment).

APPENDIX B

Venture financing: How portfolio companies generate and preserve equity through multiple financing rounds

The venture capital financing process begins when venture capital investors and the founding entrepreneur(s) of Company A negotiate a valuation that takes into account the company’s technology, experience and other assets, as well as the risks it entails. At the first stage of financing, the company has a much higher risk of failure than success and will require significant additional capital to develop its products. In this example, Company A has been valued at $10 million (most start-ups are valued below this amount, but it is a useful number for demonstration purposes).

Next, the founders of Company A raise capital by selling 40 percent of the company’s equity to “first round”, or Series A investors (i.e. venture capitalists). The company now holds $4 million in cash, with 40 percent of the firm’s equity held by venture capitalists and 60 percent owned by founders and employees.

Assume that Company A is successful in further product development and that the likelihood of success gradually increases (along with a commensurate reduction in risk of failure). In Year Three, the company demonstrates measurable progress and seeks additional capital at a “higher valuation”. For this second round (or Series B), the owners find new investors who believe that the pace of product development, competitive advantages and markets sizes for planned products, discounting new risk and return analyses for timing of liquidity and return on investment, value the company at $50 million. In this case, the 40 percent purchased earlier by Series A investors is now worth $20 million. One-half of the remaining $30 million of founder and employee value is sold to the Series B investors for $15 million. At this point the series A investors continue to own 40 percent of the company, the Series B investors own 30 percent of the company (1/2 x 60 percent), and the original founders and employees own 30 percent of the company (100 percent - 40 percent - 30 percent). The company has $15 million in new capital and has invested the original $4 million in product development.

At this time, typically, 70 percent of the board members will be represented by outside investors. They will look for the optimal exit strategy, such as taking the company public or selling to a larger firm. But in this example, at Series B, the company still has a limited chance of success and a reasonable chance that it will fail – in which scenario the $19 million that was raised will be lost.

Return to Table of Contents

Patients’ and Consumers’ Interests and Perspectives in Personalized Healthcare

Greg Simon
Margaret Anderson
Cecilia Arradaza
Kate Blenner
Kathi Hanna
Kristin Schneeman
FasterCures

I. INTRODUCTION

“Progress in [personalized medicine] will characterize medicine
in the 21st century and extend life span much like the use of
antibiotics did in the 20th century.”
-- Gerald Levey[1], Provost and Dean, University of California,
Los Angeles School of Medicine, FasterCures Board member

The 20^th century witnessed the greatest expansion of life expectancy in the history of humankind. The challenge for the 21^st century is to not only extend the length, but to also improve the quality of life by preventing and defeating deadly and debilitating diseases. Across the spectrum - from basic science to clinical research to health services research - the impressive advances of recent decades in the biomedical, physical, computational, and behavioral and social sciences present unprecedented opportunities to improve human health and quality of life. Capitalizing on this reality will usher in an era of personalized medicine and solidify its place at the frontier of medical science.

The ultimate value of personalized medicine will be to improve treatment options for patients and prevent the onset of disease in the first place. But to realize these important gains, we need to transform our current research and healthcare systems from the outdated model of the last century to an integrated, information-based, high-quality, health-sustaining model that will extend life expectancy and improve the quality of life for generations to come.

To achieve this transformation the new system must focus on patients. How personal is personalized healthcare and what do consumers think about the advent of this era?

Embedded within each patient is the information – family history, medical records, lifestyle, biological samples, etc. – that is crucial to understanding, treating, and preventing disease. Patients need to be empowered by accurate information and armed with a clear understanding of the opportunities to:

participate in research and clinical trials;
donate biological material such as tissue and blood samples; and
advocate to have interoperable electronic health records (EHRs) to aid care and research.
As a contribution to U.S. Department of Health and Human Services Secretary Michael O. Leavitt’s Personalized Healthcare Initiative, FasterCures submits this white paper summarizing the perspective of patients and consumers, the prime constituency in the discovery of personalized medicine advances and the ultimate beneficiaries.

To paint a complete picture and accurately represent the numerous patient perspectives on personalized healthcare, FasterCures conducted a qualitative research survey of disease research organizations, patient advocates, and patients to gauge understanding, awareness, and expectations of personalized healthcare and elucidate the issues that affect millions of Americans.

II. The Path to Personalized Medicine: Patient Involvement

“Success is when everyone can learn which methods and treatments work,
and which don’t, in days instead of decades.”

-- Carol Diamond and Clay Shirky[83]

In 1799, explorers unearthed in Egypt a stone slab – the Rosetta Stone – bearing parallel inscriptions in Greek, Egyptian hieroglyphic, and demotic characters, which made it possible to decipher the written language of the ancient Egyptians and the stories that it told about the people and their culture. Each of us is, in a sense, a Rosetta Stone, for within us is the information necessary to unlock the relationship of genetics, proteomics, behavior, nutrition, and environment to the emergence and, ultimately, the management of disease.

The three "languages" of our Rosetta Stone are medical records; biological material such as tissue, blood, and DNA; and our biology as observed in clinical research. By participating in clinical research – trials to test potential new therapies as well as epidemiological, observational, or natural history studies – and by providing tissue samples, blood, or medical histories, patients can provide critical information and resources, without which the search for cures and advancements in personalized medicine could slow to a halt.

Many respondents to our survey felt that the greatest payoff to personalized healthcare will come from leveraging the patient’s role in these critical areas:

Biological specimens. It is important that patients understand the key role that biospecimens play in medical research, and how critical they are to future research discoveries. To understand the connections between genes, proteins, and the environment, sophisticated comparisons must be conducted. These comparisons cannot be done by hand or by eye, or patient by patient.

It is interesting to note that the importance of tissue sample collection was generally not mentioned by our survey respondents. Some pointed out that patients can be uncomfortable with the notion of donating their tissue, and the time to educate patients about tissue donations for research is not at the moment a consent form is being signed for diagnosis or clinical care. Patients and patient groups must be brought into the process as partners in helping to ensure that the patient community understands how biobanks work, and the role they play in the clinical research infrastructure. FasterCures has a website devoted to this topic www.biobankcentral.org.

Clinical trials. Clinical trials are the only way of evaluating whether new diagnostics, drugs, experimental medical devices, and surgical techniques actually work. These trials are dependent upon patient involvement. The FasterCures Patients Helping Doctors (PHD) Program facilitates the understanding of the critical role patients play in research, with the ultimate goal of increasing patient participation in this process. We have found that there are many reasons for the lack of patient participation including:
- patients not having enough information about clinical research,
- physicians not having enough information and not informing their patients about the possibility of enrolling in a clinical trial, and
- patients and doctors having misconceptions about clinical trials.[84]

Respondents to our survey outlined how highly motivated their patients are to participate in clinical trials. For example, in the National Institutes of Health (NIH)-sponsored Alzheimer’s Disease Neuroimaging Initiative (ADNI) trial, the enrollment has exceeded the study program director’s expectations despite some of the painful medical procedures trial participants are undergoing.[85]

Overall, patients who enter trials see it as part of their larger role of advancing science. One respondent said, “Within the cancer community, there is a profound altruistic feeling. They want to help by participating in trials, and the data shows that when they do, they feel positively about the experience.” Survey respondents did feel we need to incentivize more participation in clinical trials; otherwise, it will be hard to move personalized medicine forward.

Electronic Health Records (EHRs). The promise for personalized medicine offered by integrated EHRs is immense. EHRs will go a long way to solving the information gap that often exists as patients travel from one provider’s office to another. EHRs will also provide much-needed ways to aggregate data about treatment and outcomes for research and offer unprecedented opportunities to speed up the quest for cures. As patients wait for better therapies and eventual cures however, EHRs will help to manage some of the chaos created by complex individual co-morbid conditions.

Enabling research use of information collected in the patient care process could significantly accelerate medical research. EHRs and clinical databases and warehouses can make the work of specialists in one discipline widely accessible to specialists in many disciplines. EHR systems could speed data acquisition and searching, allow mass computing and sampling, and provide the research community access to a broader and more diverse patient population. Improvements made in EHR systems in response to research needs will ultimately serve clinical care needs as well.

III. The Personalized Medicine Landscape: What Do Patients and Consumers Think?

“We must remember that the true foundation of this progress is public trust. It is not enough merely to develop the knowledge and information that will make personalized healthcare possible. In addition to developing the information, we must use it correctly.”

-- Michael O. Leavitt, Secretary of U.S. Department of Health and Human Services[86]

It would be inaccurate to say there is only one patient community. There are hundreds, perhaps thousands of them, each defined by different experiences as their members manage disease from diagnosis through treatment and possibly cure. Patient awareness and understanding of personalized medicine and healthcare has begun, but it will be an ongoing process that will vary and evolve based on the disease.

The national discussion about personalized medicine has mostly occurred at the 30,000 foot level and has yet to comprehensively engage and permeate the broad array of patient communities with its myriad concerns.

Methodology

In order to understand the key role of patients in driving the adoption of personalized healthcare approaches, FasterCures conducted a qualitative research survey of disease research organizations, patient advocates, and patients to determine understanding, awareness, and expectations of personalized healthcare. For the survey, we reached out to senior executives of 10 groups in the FasterCures Redstone Acceleration & Innovation Network (TRAIN). We also identified an additional five national organizations that are not in TRAIN that represent the issues related to diseases that affect millions of Americans.

TRAIN is a group of unique nonprofit foundations that fund medical research across a spectrum of diseases, from breast cancer to Parkinson’s disease.[87] In many cases TRAIN’s member foundations have been created by patients and their families who are frustrated by the slow pace of change in the traditional medical research system. They represent the kind of organizations that are fast becoming the engine behind innovation in disease research – collaborative, mission-driven, strategic in their allocation of resources, and results-oriented. They are organizations that have a singular focus on, and a significant stake in, getting promising therapies from the laboratory bench to the patient’s bedside as rapidly as possible.

Figure 1 – FasterCures’ TRAIN Program

TRAIN has come together under the auspices of FasterCures – a nonprofit “action tank” whose mission is to save lives by saving time in the research, discovery and development of new medical solutions for deadly and debilitating diseases. The TRAIN network helps it members to more easily and effectively support each other’s efforts to produce better and faster results, and to bring their sense of the urgency about conducting more and better bench-to-bedside translational research to the medical research community as well as to the public at large.

FasterCures surveyed groups using email and telephone-based methods and attempted to reach representatives from a variety of diseases ranging from preventable to incurable. Specifically, representatives from the following groups were interviewed:

**Table I. *FasterCures’* Personalized Healthcare Qualitative Survey Respondents**
Organization	Organization Overview	Contact, Title/Role	Outreach Mechanism
Accelerated Cure Project for Multiple Sclerosis	Organizes the research process for multiple sclerosis and encourages collaboration between research organizations and clinicians.	Art Mellor, President & CEO, Co-Founder, Director	E-mail Correspondence
Alpha-1 Foundation	Identifies those affected by Alpha-1 Antitrypsin Deficiency (Alpha-1) and improves the quality of their lives through support, education, advocacy, and to encourage participation in research. The Association has over 70 volunteer-led support groups around the U.S	John Walsh, President	Phone Interview
Alzheimer’s Association	Mission is to eliminate Alzheimer’s disease through the advancement of research; to provide and enhance care and support for all affected; and to reduce the risk of dementia through the promotion of brain health. The organization’s achievements and progress in the field have given thousands of people a better quality of life and brought hope for millions more.	Jennifer Zeitzer, Associate Director, Federal Policy	Phone Interview
American Heart Association	Nation’s oldest and largest voluntary health organization dedicated to building healthier lives, free of heart disease and stroke. In fiscal year 2006–07 the association invested more than $554 million in research, professional and public education, advocacy and community service programs to help all Americans live longer, healthier lives.	Derek Scholes, Government Relations Manager	Phone Interview
Autism Speaks	Focuses on increasing awareness of autism spectrum disorders, to funding research into the causes, prevention, treatments and cure for autism, and to advocating for the needs of affected families.	Nancy Jones, Program Director	Phone Interview
COPD Foundation	Mission is to develop and support programs which improve the quality of life through research, education, early diagnosis, and enhanced therapy for persons whose lives are impacted by Chronic Obstructive Pulmonary Disease.	John Walsh, President	Phone Interview
Epilepsy Therapy Development Project	Mission is to advance new therapies for people living with epilepsy; supports the commercialization of new therapies through direct grants and investments in promising academic and commercial projects.	Joyce Cramer, President	Phone Interview
Friends of Cancer Research	Raises awareness and provides public education on cancer research in order to accelerate the nation's progress toward better tools for the prevention, detection, and treatment of all cancers .	Jeff Allen, Executive Director	Phone Interview
Hydrocephalus Association	Provides support, education and advocacy for people whose lives have been touched by hydrocephalus and the professionals who work with them; advocates for increased research and funding to advance understanding, improve diagnosis and treatment, and find a cure.	Dory Kranz, Executive Director	E-mail Correspondence
Lance Armstrong Foundation	Focuses on cancer prevention, access to screening and care, research and quality of life for cancer survivors. LAF has raised more than $260 million for the fight against cancer.	Adam Michael Clark, Director of Health Policy	Phone Interview
Michael J. Fox Foundation for Parkinson’s Research	Mission is to ensure the development of a cure for Parkinson’s disease within the decade through an aggressively funded research agenda. The Foundation has funded over $126 million in research to date.	Debi Brooks, Co-Founder	Phone Interview
National Health Council	Represents 119 national health-related organizations working to bring quality health care to all people. Its core membership includes some 50 of the nation's leading voluntary health agencies representing about 100 million people with chronic diseases and/or disabilities. Other Council members include professional and membership associations, nonprofit organizations with an interest in health, and major pharmaceutical and biotechnology companies.	Myrl Weinberg, President	Phone Interview
Parkinson’s Action Network	Serves as the voice of Parkinson’s on numerous public policy issues affecting the Parkinson’s community.	Mary McGuire Richards, Deputy Chief Executive Officer	Phone Interview
Prostate Cancer Foundation	Provides funding for more than 1,400 research projects at nearly 150 institutions worldwide; advocates for greater awareness of prostate cancer and more government resources, resulting in a twenty-fold increase in government funding for prostate cancer.	Jonathan W. Simons, President & CEO	Phone Interview
Susan G. Komen for the Cure	Largest grassroots network of breast cancer survivors and activists fighting to save lives, empower people, ensure quality care for all and energize science to find the cures. Invested more than $1 billion in the fight against breast cancer in the world.	Elizabeth Thompson, Managing Director, Public and Medical Affairs	Phone Interview

Additionally, FasterCures posted a description of the goals of this white paper along with several questions on PatientsLikeMe[88] to solicit candid feedback from patients. We received responses from 32 patients. The responses we garnered from this process are woven throughout this white paper. More patients are turning turn to online tools like PatientsLikeMe where they interact to help improve their outcomes. The data they provide helps researchers learn how these diseases act in the real world.

Overall Perspectives About Personalized Healthcare

Respondents identified a wide spectrum of current applications of personalized medicine for specific diseases. Our survey found that patient awareness and understanding of personalized medicine has begun, but it will be an ongoing process and that educational process will vary based on the disease. Everyone interviewed had some understanding of what personalized healthcare was, and the potential benefits it will offer as we transition from a trial-and-error, one-size-fits all approach to treatment to one that is tailored to individuals. Respondents on PatientsLikeMe were aware of it in a general sense, but didn’t necessarily know that it was called personalized healthcare.

There were some differences in whether people thought personalized healthcare was simply a way to understand the genetic and individual basis of disease or rather another way to segment patient populations and offer tailored therapies.

Even among groups who characterize themselves as less engaged on this issue, there was still widespread acknowledgment that this is the direction in which 21^st century medicine is heading. There was however, a sense that the leadership of the patient community lacked a clear sense of what was, and was not personalized medicine, identifying the need for additional work on definitions and illustrative examples. A wide spectrum of current applications of personalized medicine to specific diseases was represented by respondents including warfarin testing and BRAC1 for breast cancer.

Citing the Need for Patient-Centered Care

Some of the issues raised by the interviews were not always specific to personalized healthcare but instead represented challenges that patients have faced for years. Specifically, respondents expressed widespread frustration with the inability of the healthcare system to address each patient’s needs, and to efficiently and effectively coordinate care across providers and conditions. Personalized healthcare will not be immune to these challenges, and as innovative treatments and diagnostics grow more complex, it is a reasonable concern that the insufficiencies within coordination of care will become exacerbated.

The need for patient-focused care is increasingly more important as scientific discoveries bring us closer to personalized health care. “We need to address the medical and social goals of the whole person with multiple co-morbidities in the context of their individual life circumstances. We must try to get away from a purely medical model that offers only a disease-by-disease approach without consideration of personal desires such as living independently, remaining in the workforce or managing chronic pain,” offered Myrl Weinberg, President of the National Health Council, which represents over 120 member organizations including patient advocacy organizations.

"People with chronic conditions will interact with the health sector for the rest of their lives. If patients are an afterthought and not engaged at the front end of the research process, our collective opportunity to address the complicated medical and social needs of the whole person may be lost, and the scientific advances of personalized medicine and the expected benefits will be diminished,” said Weinberg.

Even more strongly, a patient said, “What I’ve experienced so far in most hospital environments is all but personalized… I felt more like cattle than a human being in general.”

If patients are an afterthought and are not engaged at the front end of the research process, the scientific advances of personalized medicine and the expected benefits to patients will be hindered. If patients are to be involved in clinical research leading to advancements in personalized healthcare, they need better information and a deeper understanding of it based on clear, concise, and accessible information.

A theme emerging from our analysis was that perspectives on personalized healthcare are directly shaped by the state of the science in a given disease area. All groups expressed knowledge of personalized healthcare and a majority had participated at some level in meetings and discussions on this topic. However, for diseases with a strong understanding of the mechanism causing the illness and associated targeted therapeutics, respondents offered an even more robust understanding and appreciation of personalized healthcare.

Many recognized the potential advances on the horizon for their disease area, but remarked that it still feels far enough away that it is difficult to reach and therefore difficult to plan for. “We are here and we are far away from personalized healthcare all at once,” mentioned one respondent. With some chronic diseases like heart disease it is difficult to project where the science will go, since its prevention and its treatment utilize both medical and public health approaches.

Co-morbidities are an increasing issue for many patient groups. For example, 65 percent of patients with chronic obstructive pulmonary disease (COPD) report six to ten co-morbidities, including conditions such as arthritis, diabetes, and cardiovascular disease.[89] For example, in the case of Alzheimer’s disease, 96 percent of patients have other conditions and data shows that Medicare spends up to three times more for an Alzheimer’s patient with diabetes.[90]

“Personalized healthcare of the future clearly needs to address co-morbidities,” asserted John Walsh, of the COPD Foundation. It will be important to recognize the interaction among different diseases and that personalized healthcare for one individual might require coordinating multiple treatments. Moreover, pharmacogenomics will play a crucial role in understanding efficacy and toxicity of drugs given to patients with co-morbid diseases.

Benefits of Personalized Healthcare to Patients

All respondents clearly understood the benefits of personalized healthcare described by the Personalized Medicine Coalition as the “right treatment for the right person at the right time.”[91] We found a consensus that it would be a significant advancement if the tools of personalized healthcare allow for earlier diagnosis and improved treatment success, including targeting drugs for use in people who will derive a benefit.

We found a dearth of understanding among respondents in the role personalized healthcare can play in avoiding drugs that will lead to adverse events. The removal of Vioxx from the market and the black box warning placed on other drugs attract big headlines in the media and patients are aware of these events. However, they do not always recognize that the identification of a drug causing severe side effects in a population subset is an advance in personalized healthcare. Some saw that future relabeling or warnings for medications could serve as teaching opportunities for the patient community about what personalized healthcare can offer.

Personalized healthcare has been defined as offering the promise of better care delivered more efficiently. In areas such as oncology, patients want better assurances that treatments will work for them. Particularly in cancer treatment, patients do not always have confidence that their treatment will be effective, thus they fear the side effects of a treatment that may not yield benefit. In the breast cancer community, survivors are focusing more on survivorship care plans that help them track the impact and potential for side effects of the treatments on their health down the road.[92]

Impact of Personalized Healthcare on Costs

Many respondents felt it is difficult to completely predict how personalized healthcare will unfold in the next 10-15 years and its impact on escalating healthcare costs. If personalized healthcare can help reduce costs, everyone regarded this as a positive and important benefit. Most respondents mentioned that they saw costs going up before going down as a result of personalized healthcare.

Patient advocates believe that personalized healthcare will ultimately lower costs by:

reducing the need for repeat visits,
reducing the number of adverse events and some hospitalizations, and ultimately resulting in better health outcomes,
saving patients and providers time, money, and wasted effort since most drugs are not working in some subset of certain patient populations, and
providing tools that give providers information about which subpopulations are likely to respond to therapy.

Respondents thought the cost to develop targeted, personalized therapies could be higher than the costs of developing existing treatments and might be labeled for use in smaller market sizes which could increase drug pricing. Thus there is concern that as therapies become more tailored, they may also become more expensive, and that investment in drugs for lower incidence populations won’t get pursued. Uncertainty about how payers will integrate targeted therapeutics into coverage and reimbursement decisions exists.

Concerns about Personalized Healthcare

Drug Development. Respondents acknowledged that the drug development models that currently exist will have to evolve to prepare for the personalized healthcare advances. There needs to be a process in place that considers the implications of the creation and characterization of subgroups of patients within a disease by both pharmaceutical and biotechnology companies and by the U.S. Food and Drug Administration (FDA). There are opportunities within FDA to make sure all the required policies are in place to promote the advancement of personalized healthcare practices. A robust post-marketing system needs to be in place to identify safety risks as these drugs are used by a more heterogeneous population. Also, the research building blocks with FDA drug safety efforts need to be aligned to learn more about how drugs are experienced in a large population.

From the scientific perspective, data continues to come in on most diseases about the variability within the particular disease class. Scientists and advocates are increasingly discussing the possibility of different subtypes of their particular disease areas. For example, Parkinson’s disease (PD) patients present to their doctors with their own personal mix of symptoms that roughly categorize them as PD patients. When treated, these patients often experience highly varied responses to medications. This known heterogeneity is still generally overlooked if not ignored as treatment protocols consider all these patients in a single category of disease. In fact, recent “failures” in clinical trials in PD might more appropriately be viewed as “inconclusive” findings with pockets of treatment success but insufficient (underpowered) evidence to propel the trial to its next stage of investment and/or investigation. The Michael J. Fox Foundation for Parkinson’s Research (MJFF) is focused on attempting to better understand and characterize the “subtypes” of disease with the particular goal of improved patient selection for clinical trials in mind.[93]

Also, some respondents raised the question of what needs to be done to facilitate the process of subgroup analysis and how to study different populations that respond differently to treatments. It was also acknowledged that even in areas where there are some targeted therapies identified, more research is needed. The work is not over when the initial finding is made. For example, new analysis of the data shows that women taking Tamoxifen can metabolize the drug differently.[94]

“It is clear we need to find ways to do clinical trials that are faster and cheaper,” asserted Debi Brooks, Co-Founder of MJFF. “One of our strategies is to fund creation of tools that can contribute to improved trial design in the first place.” In addition to the continued work to identify subtypes of disease, MJFF has a collaborative project underway where the Parkinson’s Institute and the company 23andMe are working to validate web-based surveys that could provide a proof-of-concept for tools to enable more robust data collection in the clinical trials process. In smaller disease populations that have potential subpopulations of disease, improved and innovative clinical trial design to increase the power of smaller sample sizes will help researchers complete studies faster.

Additionally, until we have diagnostics that can identify who should receive which drug, patients want an improved adverse events reporting system that can contribute to research and development of such tests. One way to better understand extrinsic factors like drug-to-drug interactions, medical practice, diet, alcohol use and intrinsic factors like gender, genetics, and race is to establish systems that improve adverse event tracking. Currently the FDA is actively embarking on this task. In May 2008, FDA launched its Sentinel Initiative with the goal of creating and implementing a national, integrated, electronic system for monitoring product safety. This effort will strengthen FDA’s ability to monitor the performance of a product throughout its life cycle and enable real-time reporting of potential safety signals for medical products currently on the market.

Some respondents are concerned about genetic testing companies and want assurance these tests are accurate and that support systems and providers are ready and waiting after patients take the tests. The regulatory framework for these testing companies is still being created; the FDA does not evaluate these tests for accuracy, though a federal panel recently recommended stepped-up oversight. Different states have different regulations about the ordering of tests and the involvement of medical professionals; several states have ordered direct-to-consumer testing companies to stop selling their tests to residents of their states until they prove they have met that state's quality standards (which several companies subsequently did and received licenses to operate). Two major associations for genetics professionals disagree about whether any genetic tests are appropriate for sale directly to consumers without a medical intermediary. While regulators and medical professionals deliberate, the popularity of genetic testing in undeniably increasing, helped along by "genetic social networking" Web sites and program launches at venues such as the Mayo Clinic, Canyon Ranch Institute, and the Cleveland Clinic, opening whole new frontiers in the consumer information revolution.

Gatekeepers. Many respondents identified their patients’ need for a “medical home” to provide coordinated and targeted care. One patient said, “So, while providing more detailed tracking is helpful, one also needs a doctor who is receptive to that same tracking.” Some saw how this approach may create a situation where the provider serving as a gatekeeper may instead block or slow access to care. As patients have more and more access to information, and as they have mobilized, they want access to providers that will discuss options and a gatekeeper may stand in the way of that. Similarly, as therapies start to become available for subgroups of patients, there is concern about how the payer community will react. One respondent said, “What if treatment is only available if it works for everyone with our disease?”

There has been a lot of discussion in the past couple of years about comparative effectiveness. This is the approach that many healthcare stakeholders are turning to as a possible solution to curb healthcare spending. Comparative effectiveness research seeks to provide a cost-effective and efficient approach to identifying the best in drugs, devices, biologics, and medical procedures. However, as the drumbeat for comparative effectiveness intensifies, it is important to ensure that the law of averages doe not steer decision-makers away from treatment that demonstrates true patient benefit. Comparative effectiveness needs to allow for new research findings, as well as allow for diseases that may ultimately encompass hundreds of genetic variations and subtypes.

Privacy. There is lingering concern about whether individual test results and large datasets with personal information will be used against people for employment or insurance purposes. One respondent said that the passage of the Genetic Information Nondiscrimination Act (GINA) hasn’t assuaged those fears. (For more information about GINA, see page 18). However, a majority of the patient organization leaders we spoke with felt that privacy needed to be dealt with and closely monitored, but that it should not interfere with scientific and healthcare delivery advances. One respondent said, “We don’t want the politics of fear of privacy breaches to get in the way of the needed advances.”

Advances in 21^st century healthcare will heavily depend on advances in genetic research and other medical solutions that fuel the search for new treatments and cures. The passage of the GINA allows patients to more confidently participate in studies that search for linkages between genes and disease, to enroll in clinical trials for new targeted drugs, or to provide samples for DNA analysis to optimize their own disease prevention and treatment.

Due to the lack of EHRs in many care systems, respondents noted that often patients’ records were private, even to them. Some felt that the general consumer population was more concerned about privacy than patients, many of whom understand the value that pooled data can provide to the understanding of their disease. However, some still have concerns about posting their data onto some of the online personal health records systems. One patient said, “One of the risks that is going to emerge very quickly is the privacy status of medical records held by companies which function as control repositories.”

The impact of the Health Insurance Portability and Accountability Act (HIPAA) and privacy were raised in the context of conducting research studies. In many disease areas, sample collection is becoming standard practice, and yet there is still confusion of what is and is not allowable under HIPAA. There was concern about the impact of restrictions on the speed at which research can be conducted, and the fact that patients continue to lose ground in battling their conditions with these delays.

Educating Patients

In order to be truly effective with optimal impact, patient-centric and proactive healthcare practices must be supported by comprehensive education and communications efforts. The general public needs to understand genetic medicine - what it can and cannot do - and not be afraid of the power of this area of science. Healthcare providers need to be able to sift through the most recent advances in medicine and translate these into real-world scenarios, carefully putting the most promising developments into context for each patient. The doctor-patient relationship needs to be defined by clear and transparent lines of communication. It is vital that new developments brought about by personalized medicine approaches be managed and translated responsibly and effectively into tangible treatment protocols when appropriate.

Most felt that it would not be difficult to educate patients about the advances that will come from personalized healthcare. Patients are hungry for information, and many survey respondents mentioned how self-motivated their constituencies are. Many respondents cited the high motivation their constituencies have to accelerate the research process in order to have better treatments available.

One respondent felt that trusted messengers (e.g., medical associations, advocacy groups, the U.S. Surgeon General) could lead national efforts to educate consumers. It was pointed out that a major risk relates to unrealistic expectations by the patient. This patient said, “Sometimes, even with the right diagnosis and treatment, I won’t get better.”

It was felt that all stakeholders involved need to carry the messages to patients about the potential benefits personalized healthcare offers. Providers ranging from primary care physicians to specialists and all other providers that intersect with the patient communities need to be given tools to help them communicate these messages.

There is still a lot to learn about how patients will respond to detailed genetic profiling as that becomes a reality. One person said, “The jury is still out about how this will really be rolled out over time and how patients will manage this new information.”

Some groups talked about needing more documentation of successes in the area. “We need to have the demand for the science defined publicly so it is constituent driven.” Another respondent spoke of the flat funding for NIH and the concerns that it raises for the future pace of scientific advances. These comments speak to the need to engage fully with patients to be research advocates and suggests that the more motivated a patient is to get involved in a patient-oriented organization, the more likely they will be engaged in personalized healthcare.

IV. Potential Impact of Personalized Healthcare in Healthcare Delivery

There are several areas of healthcare that will be significantly affected by the adoption of a personalized medicine approach. Most notably, personalized healthcare alters the traditional model of healthcare delivery, shifting some responsibility toward the consumer while simultaneously requiring healthcare providers to process even more information. It also raises questions about:

when evidence is sufficient for use in health and disease management;
how best to gather and assess evidence about effectiveness and efficacy; and
how to appropriately regulate drugs used in personalized medicine.

Use of Genomics and Biomarkers to Predict Disease

An individual’s genetic and molecular profile, if accurately assessed, has the potential to predict predisposition to certain chronic diseases – for example, prostate cancer, glaucoma, Alzheimer’s disease, or heart disease – as well as guide disease prevention strategies and more effective use of therapies. Currently, many of these tests are predictive, rather than diagnostic, which means results are provided to otherwise healthy consumers as probabilities, or relative risks for an individual versus the general population. Most tests rely on SNP analysis or whole genome scans but others are based on non-DNA biomarkers associated with a particular pathological or physiological state.

As the technology for such testing – in particular genomic analysis – has advanced, the costs have decreased, which has spawned the growth of a new industry focused on personalized genomic services, frequently marketed directly to the consumer. Because in most cases the consumer can purchase the test and receive results without the direct involvement of a personal healthcare professional, several concerns have arisen.

Is the scientific evidence supporting the genomic-disease associated information sufficient for clinical use?[95]
Are consumers able to appropriately and effectively use such information in their own healthcare management?[96] and
Are healthcare providers sufficiently proficient in the application of probabilistic genomic information to respond to patient queries and develop a healthcare management plan appropriate for individual patients?[97]

Those advocating for more consumer involvement in test decisions believe that the slow pace of provider uptake and professional education, combined with more focus on consumer education and autonomy warrants such an approach.[98]

Pharmacogenomics

A specific field in personal medicine is pharmacogenomics, sometimes called molecular medicine. Pharmacogenomics is based on identifying genetic factors that directly influence a person’s response to a drug. It has the potential to enhance understanding of disease etiology and diagnosis as well as the determinants of drug effects so better prescribing decisions can be made. What makes pharmacogenomics both unique and a challenge is that it melds the worlds of diagnosis and treatment in new and different ways. It is an application of genetics and pharmacology that brings genetic testing into the purview of primary care, well beyond the more traditional bounds of rare diseases, where genetic testing has its historical roots.[99]

It is likely that in the future, drugs incorporating pharmacogenomic data will involve both a therapeutic agent and diagnostic test, wherein the diagnostic test will precede the prescription, which suggests a new model for healthcare delivery. Because pharmacogenomics can help physicians determine whether a proposed drug therapy is relevant to a given patient, this approach to clinical care has the potential to enhance preventative medicine and reduce the level of trial-and-error in patient management. As with the use of personalized genomics testing services, pharmacogenomics will increase the volume of information that will have to be processed and used by patients and their healthcare providers.

V. New Approaches and Opportunities to Transform the Drug Development Process

“…in the next 15 years the pharmacopoeia that we use for treating lots of disease will be very heavily influenced by the things we’re discovering right now about the molecular basis of disease. But that has the longest lead time, and so it won’t happen overnight for many conditions.”

-- Francis Collins, former Director of the National Human Genome Research Institute at the NIH[100]

New Approach to Clinical Trials

One of the challenges of personalized healthcare lies in assessing outcomes. First, because some of these interventions are being offered directly to the consumer it will be difficult to follow consumers to assess effectiveness and other outcomes. Thus, it will be critical that there be some publicly funded studies in these areas.

Second, because the very nature of clinical evidence will become more focused on individuals and subpopulations, personalized healthcare challenges the notion of randomized clinical trials as the gold standard for testing the safety and efficacy of new diagnostics and drugs. Simple reliance on biomarkers may be a poor method of predicting outcomes.

At least for some time it will be critical to evaluate large numbers of people before understanding the relative role of any given variant and its significance in personalized healthcare. Weak predictability combined with our lack of understanding of the causal relationship between genes and drug responses makes it difficult and costly to conduct appropriate validation studies. These studies are probably going to have to be large-scale, prospective studies that measure genetics and other biomarkers over time and follow up with patients for long-term outcomes.[101]

As such, analyzing evidence emerging from personalized medicine will require a different set of skills than those used in traditional clinical trials, combining diagnostic evidence with safety and efficacy evidence. Research will be needed to develop the best methods for collecting and analyzing evidence and large numbers of subjects will be needed for clinical trials.

Seizing Proven Opportunities

While nearly 10 percent of the drugs approved by the FDA include pharmacogenomic information in their labeling, only four have a sufficient body of evidence to support a requirement for genetic testing before treating a patient.[102] Many other drug labels reference validated biomarkers and associated diagnostic assays, but these are only ‘recommended’ to provide additional information—not because evidence has shown their impact on outcomes to be variable or unreliable, but because there is no evidence regarding outcomes at all. This highlights a fundamental imbalance in the progress of pharmacogenomic research: more and more studies are linking genotype to the mechanisms of drug metabolism and/or efficacy, but few are taking the critical next step of tying modified dosing or selective use of drugs based on genotype to improved patient outcomes. Stakeholders have identified the lack of clinical evidence base as a critical barrier to integration of personalized medicine into routine practice.[103] Making this connection to outcomes is necessary to realize personalized medicine’s promise.

The stakes are even higher since many of the drugs for which pharmacogenetic factors have been identified are often dangerous to patients and adverse reactions can be lethal. The FDA’s list of drugs with genetic biomarkers includes chemotherapy agents, anticoagulants, and neurologic agents—drugs whose side effects would exclude them from use were it not for the lack of suitable therapeutic options for patients with grave conditions. With more than 770,000 injuries and deaths due to adverse drug reactions and medication errors each year[104], elucidating whether genetic information can improve outcomes and reduce some of these events is critical to ensuring the safety of patients who take these drugs.

A growing body of research reveals the great promise of using an individual’s genetic information to guide his or her care; the next step for us is to seize that demonstrated opportunity by confirming whether this information can effect real change in short- and long-term patient outcomes. We can save patients’ time by building the evidence base as soon as possible so that caregivers can act on the promise of personalized medicine. We can save patients’ lives by defining how genetic tools can ensure a patient’s treatment is not only timely and beneficial, but safe.

VI. Making Personalized Medicine a Reality: The Need to Address Privacy

“These are catch-all diseases (e.g., cardiovascular disease, Alzheimer’s disease,
rheumatoid arthritis, and cancer) that look the same, but when you scratch below
the surface, you begin to understand that the underlying physiology of similar phenotypes can be fundamentally different.”

-- John Sninsky, Vice President of Discovery Research at Celera[105]

Precious patient resources are lost to medical research if individuals fear that genetic information, test results, or electronically stored health records might be used against them by insurers or employers. Public opinion has long reflected widespread anxiety about misuse of personal health information.

In a 2004 survey of 470 people with a family history of colorectal cancer, for example, about half said their concern about genetic discrimination was high, and that they would be significantly more likely to pay for genetic testing out of pocket, use an alias, or ask for test results to be excluded from their medical record.[106] Dr. Francis Collins, former Director of the National Human Genome Research Institute, has said that “at the NIH, fear of genetic discrimination is the most commonly cited reason that people decline to participate in research on potentially life-saving genetic testing for colon cancer and breast cancer. One-third of eligible participants have declined on this basis.”[107] People have been reluctant to know and act on genetic health risks, to their own detriment and society’s as a whole.

A patchwork of legislation at the state and national levels has tried to regulate the use and disclosure of personal health information, most prominently the 1996 HIPAA, which regulated the use and disclosure of such information by certain “covered entities.” Successfully navigating HIPAA and human research protections will be critical to advancing the science of personalized medicine.[108] And in 2008, after 13 years of effort, Congress passed and the President signed the GINA, which advocates have called the critical civil rights bill for the genome era.

To summarize, GINA:

Prohibits use of an individual’s predictive genetic information in setting eligibility or premium or contribution amounts by group and individual health insurers;
Prohibits health insurers from requesting or requiring an individual to take a genetic test;
Prohibits use of an individual’s predictive genetic information by employers in employment decisions such as hiring, firing, job assignments, and promotions;
Prohibits employers from requesting, requiring, or purchasing genetic information about an individual employee or family member.[109]

The health insurance provisions of the bill will take effect in May 2009 and the employment provisions will take effect in November 2009. GINA does not apply to members of the U.S. military, or to other forms of insurance such as life, disability, or long-term care.

It is expected that passage of GINA will boost demand for genetic tests, leading to improvements in care and more participation in research that involves the collection of genetic information. But the passage of legislation is not enough. There has to be effective education of the public and providers about the protections that GINA confers. That includes compelling demonstration of the benefits genetic testing and personalized medicine will bring to them as individuals, as well assurance that new tests and personalized treatments will be paid for.

In addition, the application of GINA’s protections must be clear and consistent. Lessons must be learned from the experience with HIPAA, whose provisions regarding privacy have been misinterpreted and over interpreted in ways that have been detrimental to the conduct of medical research. In a 2007 survey published in the Journal of the American Medical Association, more than two-thirds of epidemiologists reported that the HIPAA Privacy Rule has made research more difficult, adding a great deal of cost and time to study completion without a countervailing positive influence on subjects’ privacy.[110]

And we need to continue to look beyond GINA at additional ways in which privacy concerns must be addressed in order to promote and facilitate the development of personalized healthcare. For instance, not addressed by GINA are all the security and privacy implications of the large databases of medical records tied to biological samples that will be required for the promise of personalized medicine to be realized.

VII. Genetic Literacy and the American Public

Patient-centered care requires that patients be informed, proactive partners with their physicians when facing health decisions. But a major hurdle for patient-centeredness in personalized medicine is a lack of ‘genetic literacy’ or a fundamental understanding of genetics and health in the general public. Informed patients are critical to patient-centered care, but as personalized medicine techniques become more sophisticated and information more complex, caregivers will face steeper challenges in communicating effectively with patients of all education levels and backgrounds. Improving the genetic literacy of the general public will be an important step in empowering patients to seek and understand personalized medicine.[111] As early as 1994, the National Research Council (NRC) was making calls for a "genetically literate public that understands basic biological research, understands elements of the personal and health implications of genetics, and participates effectively in public policy issues involving genetic information."[112]

Unfortunately, the past 14 years have not seen the NRC’s vision realized. A 2006 study on public attitudes about evolution showed that on an index of genetic literacy, American adults scored a median of 4 on a 0-10 scale, indicating that many adults are not well-informed of genetics principles.[113]Some studies have shown that minority populations of diverse cultures, in particular, have limited genetic knowledge despite a desire to know more about genetics and health.[114]

There are a number of programs aimed at addressing these deficits in genetic knowledge in the public: for example, March of Dimes has launched its Consumer Genetics Education Network (CGEN) Project, a five-year program to address genetic literacy in underserved populations and to increase access to culturally and linguistically appropriate genetics education programs and services.[115] The Health Resources and Services Administration funds the activities of the ‘Consumer Initiatives for Genetic Resources and Services’, a discretionary grant program through the Maternal and Child Health Bureau. Programs receiving grants provide education about genetics and genetic testing to patients, usually in the context of specific screening tests or conditions.[116] Genetic Alliance is one of the recipients of MCHB grants to improve genetic literacy, and is also working with funding from Centers for Disease Control and Prevention to develop the Access to Credible Genetics (ATCG) Resources Network, a genetics information resource for patients with rare genetic diseases and their families and physicians.[117] The National Human Genome Research Institute (NHGRI) at the National Institutes of Health also has active grants awarded to projects addressing genetic literacy among underserved groups.[118]

VIII. Personalized Healthcare: A Patient-Centered Action Plan

Personalized healthcare promises to be curative, predictive, and preventive. Our qualitative survey of patient organizations and patients themselves found a shared anticipation of the cutting-edge possibilities of personalized healthcare advances, especially as seeds of innovation yield tangible tools that move this approach forward. However, patient involvement is central to generating a sea-change in the traditional model of healthcare delivery.

Realigning the promise of personalized healthcare requires effectively and efficiently shifting some responsibility to the consumer while simultaneously requiring healthcare providers to process even more information.

We offer a framework for multiple stakeholders in the healthcare delivery system to act on to make personalized healthcare a reality:

Involve Patients in Medical Research. An individual’s genetic and molecular profile has the potential to predict predisposition to certain diseases, guide prevention strategies, and develop customized therapies. It is crucial for patients to understand their value to medical research and to actively participate by donating their biological specimens, being a part of clinical trials, and advocating for the use of EHRs. Accelerating and rewarding patient involvement in medical research will allow us to seize personalized healthcare’s promise to affect real change in short- and long-term patient outcomes.
Transform the Drug Development Process. Personalized healthcare challenges the long-held belief that randomized clinical trials are the gold standard for testing the safety and efficacy of new diagnostics and drugs. Understanding evidence emerging from personalized medicine will require a different set of skills than those used in traditional clinical trials, combining diagnostic evidence with safety and efficacy evidence. One of the challenges of personalized healthcare lies in assessing outcomes because some of these interventions are being offered directly to the consumer and because the very nature of clinical evidence will become more focused on individuals and subpopulations.
Protect Patient Privacy. Key to the widespread adoption of personalized healthcare is addressing public anxiety about misuse of personal health information. The privacy protections realized through the passage of GINA will lead to improvements in care and more participation in research that involves the collection of genetic information. We need to ensure that the application of GINA’s protections is clear and consistent.
Focus on and Deliver Patient-Centered Care. Personalized healthcare elevates the role of the patient to that of data source, proactive partner, and decision-maker. The role of the healthcare provider will evolve as well. The provider becomes the information filter, translating medical breakthroughs into real-world scenarios applicable at a personal level. However, our ability to deliver patient-centered care, and therefore personalized healthcare could be held back by the existing insufficiencies within our healthcare system.

[1] Scylla and Charybdis are two monsters from Greek Mythology viewed as virtually impossible for ships to pass between, as getting too close to either risked destruction of the crew and ship.

[2] Califf RM. Defining the balance of risk and benefit in the era of genomics and proteomics. Health Affairs. 2004;23(1) 77-87.

[3] Faulkner E. Addressing the realities of health care in the 21st century: a time for collaborative solutions. J Manag Care Med 2005;(8)2:11-2.

[4] Gelijns AC, Brown L, Magnell C et al. Evidence, politics, and technological change. Health Affairs. 2005:24(1):29-40.

[5] Ibid.

[6] Employer Health Benefits 2007 Annual Survey. Kaiser Family Foundation 2007.

[7] Health care spending in the United States and OCED countries. Kaiser Family Foundation 2007. http://www.kff.org/insurance/snapshot/chcm010307oth.cfm.

[8] National health expenditure projections 2007-2017. National Health Expenditures Survey. Centers for Medicare and Medicaid Services 2007. http://www.cms.hhs.gov/NationalHealthExpendData/Downloads/proj2007.pdf.

[9] Technological change and the growth of health care spending. U.S. Congressional Budget Office. January 2008

[10] Knowing what works in health care: a roadmap for the nation. Institute of Medicine. National Academy of Sciences 2008.

[11] The value of diagnostics: innovation, adoption and diffusion into health care. Advanced Medical Technology Association. Jul 2005.

[12] Cutler DM, Rosen AB, Vijan S. The value of medical spending in the United States, 1960-2000. NEJM 2006: 355(9).

[13] Realizing the Potential of Pharmacogenomics: Opportunities and Challenges. Secretary’s Advisory Committee on Genetics Health and Society 2007. http://www4.od.nih.gov/oba/SACGHS/reports/SACGHS_PGx_Report.pdf

[14] Decision support systems references a broad variety of software-based systems that aim to improve end user (in this case physicians and other clinical decision makers) decision making by systhesizing, simplifying and/or involving algorithms that facilitate manipulation of complex or potentially confusing information.

[15] Overcoming barriers to electronic health record adoption. Health Care Financial Management Association 2006. http://www.hhs.gov/healthit/ahic/materials/meeting03/ehr/HFMA_OvercomingBarriers.pdf

[16] Based on objectives defined in the Medicare Prescription Drug, Improvement and Modernization Act (MMA) of 2003 and other initiatives focused on health care quality and performance improvement.

[17]Personalized Health Care. Opportunities, Pathways, Resources. United States Department of Health and Human Services. 2007. Accessed September 28, 2007: http://www.hhs.gov/myhealthcare/news/phc-report.pdf.

[18] Faulkner E. The road to value-based healthcare: destination apparent, journey uncertain. J Manag Care Med 2006;(9)3:27-9.

[19] The US Department of Health and Human Services (HHS), Office of the Assistant Secretary for Planning and Evaluation (ASPE) has commissioned this white paper as one in a series intended to evaluate and conceptualize business and management processes necessary for integration of personalized medical practices into health care. These efforts also address issues central to the forward-looking aspects of Secretary Michael Leavitt’s Personalized Health Care Initiative that emphasize planning for the integration of personalized health principles into the delivery of health care. By identifying barriers to personalized health care and best practices to overcome them, HHS will be better prepared to communicate and plan for health systems change in a manner that appropriately leverages new technology and medical innovation, supports viable financial models, and engenders highly efficient, quality-focused, and personalized health care delivery practices.

[20] Elstein AS. On the origins and development of evidence-based medicine and medical decision making. Inflamm. Res. 53 Suppl (2) 2004: S184–9.

[21] Clancy, C. The evolution of evidence-based medicine and personalized healthcare. Presented at the 21^st Century Medicine: Personalized and Evidence-Based conference. Washington, DC 2007.

[22] Khoury MJ, Gwinn M, Burke W, Bowen S, Zimmern R. Will genomics widen or help heal the schism between medicine and public health? Am J Prev Med. 2007:33(4):310-7.

[23] Developing Biomarker-based Tools for Cancer Screening, Diagnosis and Treatment. IOM Workshop Summary 2006.

[24] Gwinn M, Khoury MJ. Genomics and public health in the United States: signposts on the translation highway. Community Genet 2006;9(1):21-6.

[25] Garrison LP and Austin MJF. Linking pharmacogenetics-based diagnostics and drugs for personalized medicine. Health Affairs 2006;25(5):1281-1290.

[26] Ibid.

[27] Glickman SW, Ou FS, DeLong ER, et al. Pay for performance, quality of care, and outcomes in acute myocardial infarction. JAMA. 2007;297(21):2373-80.

[28] Tanne JH. Performance related pay doesn't improve quality of primary care, US study finds. BMJ. 2008;337:a1160

[29] Linares A, Gauthier P, Isaacs C, Barnett P, et al. The benefits of pay-for-performance programs are worth the cost and hassles to provider organizations to participate. Health Data Manag 2007;15(12):10.

[30] McGlynn EA, Asch SM, Adams J, Keesey J, et al.The quality of health care delivered to adults in the United States. N Engl J Med 2003;348(26):2635-45.

[31] Stulberg J. The physician quality reporting initiative--a gateway to pay for performance: what every health care professional should know. Qual Manag Health Care 2008;17(1):2-8.

[32] Ferrara J. Personalized medicine: challenging pharmaceutical and diagnostic company business models. MJM 2007;10(1): 59-61.

[33] Watkins JB, Choudhury SR, Wong E, Sullivan SD. Managing biotechnology in a network-model health Plan: a US private payer perspective. Health Affairs 2007;25(5):1347-52.

[34] Mullins CD, Lavallee DC, Pradel FG, et al. Health plans' strategies for managing outpatient specialty pharmaceuticals. Health Aff 2006;25(5):1332-9.

[35] Ibid.

[36] Stern D, Reissman D. Specialty pharmacy cost management strategies of private health care payers.J Manag Care Pharm. 2006;12(9):736-44.

[37] Fish L. The case for cost sharing for biologic therapies. Amer J Manag Care 2006;12(6):159-161.

[38] OncoType Dx. Wikipedia 2008. Accessed on August 15, 2008 at: http://en.wikipedia.org/wiki/Oncotype_DX.

[39] When new technology revolutionizes patient care:a blood test that provides new information about transplant rejection. Executive Healthcare Management 2008. Accessed on August 15, 2008 at: http://www.executivehm.com/pastissue/article.asp?art=270586&issue=210.

[40] About EGAPP. National Office of Public Health Genomics. Centers for Disease Control and Prevention. http://www.cdc.gov/genomics/gtesting/EGAPP/about.htm.

[41] Diagnostics have historically represented only approximately 2-3% of overall US health expenditures and have not until recently garnered significant concern from payers and policy makers. http://www.socalbio.org/pdfs/thevalueofdiagnostics.pdf

[42] Bell G. Managing office administered drugs: an economist’s perspective. JMCM 10(2) 2007.

[43] Reporting hospital quality data for annual payment update. Centers for Medicare and Medicaid Services. www.cms.gov.

[44] CMS increasing required data reporting. Brigham and Women’s Hospital. Accessed August 15, 2008 at: http://www.brighamandwomens.org/publicaffairs/publications/DisplayMSN.aspx?articleid=1614&issueDate=11/1/2007%2012:00:00%20AM.

[45] US Department of Health and Human Services Medicare hospital value-based purchasing plan development. Centers for Medicare and Medicaid Services 2007. http://www.cms.hhs.gov/AcuteInpatientPPS/downloads/hospital_VBP_plan_issues_paper.pdf.

[46] Ibid.

[47] Bussey HI, Wittkowsky AK, Hylek EM, Walker MB. Genetic testing for warfarin dosing? not yet ready for prime time. Pharmacotherapy 2008;28(2):141-3.

[48] The value of diagnostics: innovation, adoption and diffusion into health care. Advanced Medical Technology Association. Jul 2005.

[49] Information technology for genetic and genomic based personalized medicine. The Harvard Medical School – Partners HealthCare Center for Genetics and Genomics 2008. http://www.parliament.uk/documents/upload/stGMHarvardMedicalSchool.pdf.

[50] Kawamoto K and Lobach DF. Clinical decision support for genomics and personalized medicine. Institute for Genome Sciences and Policy Medicine Forum. Duke University 2007.

[51] Ibid.

[52] Infrastructure to monitor utilization and outcomes of gene-based applications. Agency for Healthcare Research and Quality 2008.

[53] Faulkner E. The road to personalized health care: translating promise into practice. J Manag Care Med 2007;(10)6:25.

[54] Personalized Health Care. Opportunities, Pathways, Resources. United States Department of Health and Human Services. 2007. Accessed September 28, 2007: http://www.hhs.gov/myhealthcare/news/phc-report.pdf.

[55] Frueh F, Amur S, Mummaneni P, et al. Pharmacogenomic biomarker information in drug labels approved by the United States Food and Drug Administration: prevalence of related drug use. Pharmacotherapy 2008;28(8):992-998.

[56] Genomic updates to drug labeling could result from Medco/FDA partnership.

[57] Ghosh AK. Dealing with medical uncertainty: a physician's perspective. Minn Med. 2004;87(10):48-51.

[58] Ghosh AK and Ghosh K. Translating evidence-based information into effective risk communication: current challenges and opportunities.J Lab Clin Med. 2005;145(4):171-80.

[59] Based upon data from a 2007-08 comprehensive web-based survey of the membership of the National Association of Managed Care Physicians. Of the 150 total responses, 62 were from managed care organization (MCO) decision makers (predominately medical and pharmacy directors), 31 were from health system and hospital administrators and provider decision makers, 6 were from large US employers or purchaser organizations, and 10 represented commercial life sciences manufacturers. (publications in progress).

[60] Payer principles. Personalized Medicine Coalition. Accessed August 28, 2008: http://www.personalizedmedicinecoalition.org/sciencepolicy/payer_principles.php.

[61] Faulkner E. Oncology biomarkers: reimbursement implications for diagnostics and therapeutics. Presented at the GTC Bio Oncology Biomarkers: From Discovery to Validation Conference, San Francisco, CA 2008.

[62] Appleby C. Making the case to managed care. Biotechnology healthcare Dec. 2004:16-25.

[63] Diffusion and Use of Genomic Innovations in Health and Medicine. Institute of Medicine 2008.

[64] Realizing the Potential of Pharmacogenomics: Opportunities and Challenges. Secretary’s Advisory Committee on Genetics Health and Society 2007.

[65] The value of diagnostics: innovation, adoption and diffusion into health care. Advanced Medical Technology Association. Jul 2005.

[66] A better Medicare for healthier seniors: recommendations to modernize Medicare's prevention policies. Washington, DC: Partnership for Prevention 2003.

[67] Renner P, Renner P, Didebahn R, Bullock AN Challenges in developing quality performance measures for geriatric populations. Abstr Acad Health Serv Res Health Policy Meet 2002;19:33.

[68] Quality matters: patient-centered care. The Commonwealth Fund 2007;Volume 23

[69] Schoenbaum SC. Can care be patient-centered and clinically efficient? Bulletin of the Royal College of Pathologists July 2007;139:27–30.

[70] Re-engineering the clinical research enterprise: translational research. NIH Roadmap for Medical Research. National Institutes of Health 2008. www.nih.gov.

[71] Translating research into practice II: fact sheet. Agency for Healthcare Research and Quality 2001. http://www.ahrq.gov/research/trip2fac.htm.

[72] Promoting innovation and competitiveness: President Bush’s technology agenda. 2004. http://www.whitehouse.gov/infocus/technology/economic_policy200404/chap3.html.

[73] Bradley EH, Webster TR, Baker D, et al. Translating research into practice: speeding the adoption of innovative health care programs. Commonwealth Fund 2004;(724):1-12.

[74] Faulkner E. The road to personalized health care: translating promise into practice. J Manag Care Med 2007;(10)6:25.

[83] Diamond C, Shirky C, “Health Information Technology: A Few Years of Magical Thinking?” Health Affairs, September/October 2008; 27(5): w383-w390.

[84] See the FasterCures’ white paper, Clinical Trials Recruitment and Retention: Best Practices and Promising Approaches, September 2006, http://www.fastercures.org/objects/pdfs/meetings/FC_ClinicalTrials_report_art_spg.pdf.

[85] Laurie Ryan, Program Director, Alzheimer’s Disease Clinical Trials, National Institute on Aging, NIH, Presentation Comments, Institute of Medicine Forum on Drug Discovery, Development, and Translation workshop “Breakthrough Business Models: Drug Development for Rare and Neglected Diseases and Individualized Therapies,” June 23, 2008, Washington, DC.

[86] See the Department of Health and Human Services report, Personalized Health Care: Opportunities, Pathways, Resources, September 2007, www.dhhs.gov/myhealthcare/news/phc-report.pdf.

[87] See www.fastercures.org for more information on FasterCures’ TRAIN program.

[88] PatientsLikeMe is the leading treatment, symptom and outcome sharing community for patients with life-changing conditions, and creates new knowledge by charting the real-world course of disease through the shared experiences of patients with ALS, Multiple Sclerosis, Parkinson's, HIV, and Mood conditions (including depression, bipolar, anxiety, OCD and PTSD). The company endeavors to create the largest repository of real-world disease information to help accelerate the discovery of new, more effective treatments. See www.patientslikeme.com.

[89] Personal communication with John Walsh, President of COPD Foundation, September 4, 2008.

[90] Personal communication with Jennifer Zeitzer, Associate Director, Federal Policy, Alzheimer’s Association, September 10, 2008.

[91] See the Personalized Medicine Coalition report, The Case for Personalized Medicine 2006, http://www.personalizedmedicinecoalition.org/communications/pmc_pub_11_06.php.

[92] Personal communication with Elizabeth Thompson, Managing Director, Public and Medical Affairs, Susan G. Komen for the Cure, September 18, 2008.

[93] Personal communication with Debi Brooks, Co-Founder, Michael J. Fox Foundation for Parkinson’s Research, September 4, 2008.

[94] Personal communication with Elizabeth Thompson, Managing Director, Public and Medical Affairs, Susan G. Komen for the Cure, September 18, 2008.

[95] Cecile A et al., “A Critical Appraisal of the Scientific Basis and Personalize Health Interventions,” The American Journal of Human Genetics, 82:593-599, March 2008.

[96] Gosline A, “Genome Scans Go Deep into Your DNA,” Los Angeles Times, April 14, 2008, www.latimes.com/features/health/la-he-genome14apr14,0,2443364.story.

[97] Harvey EK et al., “Providers’ Knowledge of Genetics: A Survey of 5915 Individuals and Families with Genetic Conditions,” Genetic Medicine 9(5): 259-267, 2007; Scheuner, et al., “Delivery of Genomic Medicine for Common Chronic Adult Diseases,” Journal of the American Medical Association 299(11):1320-1334, 2008.

[98] Wolfberg AJ, “Genes on the Web—Direct-to-Consumer Marketing of Genetic Testing,” New England Journal of Medicine, 355(6):543-545, 2006.

[99] Hanna KE, “Pharmacogenomics and the Evolving Regulatory Paradigm,” Research Practitioner 8 (6):210-216, November-December 2007.

[100] Collins F, “Genomic Research and Personalized Medicine: An Expert Interview with Francis Collins, MD, PhD.” Medscape Genomic Medicine, May 8, 2008.

[101] Garrison LP and Austin MJ, “Linking Pharmacogenetics-Based Diagnostics and Drugs for Personalized Medicine,” Health Affairs 25(5):1281-1290, 2006, p.1285.

[102] See FDA‘s Table of Valid Genomic Biomarkers in the Context of Approved Drug Labels, http://www.fda.gov/cder/genomics/genomic_biomarkers_table.htm.

[103] Deverka PA et al., “Integrating Molecular Medicine in the US Healthcare System: Opportunities, Barriers and Policy Challenges,” Clinical Pharmacology & Therapeutics, 2007; 82(4): 427-34.

[104] See Agency for Healthcare Research and Quality report, Reducing and Preventing Adverse Drug Events to Decrease Hospital Costs. Research in Action, Issue 1, March 2001, http://www.ahrq.gov/qual/aderia/aderia.htm.

[105] Mullin R, “Personalized Medicine,” Chemical & Engineering News, February 11, 2008.

[106] Apse KA, et al., “Perceptions of genetic discrimination among at-risk relatives of colorectal cancer patients,” Genetics in Medicine, 6:510-516, 2004.

[107] Kibak P, “After long wait, GINA becomes law,” Clinical Laboratory News, July 2008.

[108] Carhart S, “Coming Century to Witness Major Changes as Hospitals Adapt to Personalized Medicine,” BNA’s Health Law Reporter, 17(25): 1155, 2008.

[109] See Genetics and Public Policy Center, www.DNApolicy.org

[110] Ness R, “Influence of the HIPAA privacy rule on health research,” Journal of the American Medical Association, 298(18):2164-2170, 2007.

[111] Haga, SB, “Teaching resources for genetics,” Nature Reviews Genetics, 7, 223–229, 2006.

[112] National Research Council report, Assessing genetic risks: implications for health and social policy, 1994, Washington, DC: National Academies Press.

[113] Miller JD, et al., “Public Acceptance of Evolution,” Science, Vol. 313. no. 5788, pp. 765 – 766, August 11, 2006, Vol. 313. no. 5788, pp. 765 – 766.

[114] Catz DS, et al., “Attitudes about genetics in underserved, culturally diverse populations.” Community Genetics. 8(3):161-72, 2005.

[115] See March of Dimes CGEN Project Website http://www.marchofdimes.com/professionals/15829_29466.asp

[116] Health Resources and Services Administration, Maternal Health Bureau, Consumer Initiatives for Genetic Resources and Services Abstract search, https://perfdata.hrsa.gov/mchb/DGISReports/Abstract/AbstractSearch.aspx, Last Accessed September 10, 2008.

[117] See Genetic Alliance Access to Credible Genetics Resources Network http://geneticalliance.org/atcg.

[118] See National Human Genome Research Institute Active Grants Database, http://www.genome.gov/10001799 Last accessed August 25, 2008.

Return to Table of Contents

Commissioned

Personalized Health Management:A Geisinger View

INTRODUCTION

PHM OBJECTIVES

Objectives 1 and 2: Foundational

Objectives 3 and 4: Supporting the Consumer

Obtaining Data and Information

Consumer Choice

Objective 5: Supporting the Clinician

Risk Assessment and Stratification

Personalized Prediction Models

Visual Display of Clinical Information

Clinical Decision Support (CDS)

Objective 6: Shared Decision Making

Objective 7: Continuous Management

PHM IMPLEMENTATION AT GEISINGER

CONCLUSIONS

References

Role of Professional and MedicalSpecialty Societies in theEra of Personalized Health Practice

US Cost and Quality Challenges: The Context for Personalized Health Care

Balancing Standardized vs. Personalized Health Care

How Professional Societies Can Contribute to Elements Important to Personalized Health Care

1. Clinician education, training, and certification:

2. Clinician decision support and information management:

3. Patient education and decision support:

4. Health outcomes research and new health technology evaluation:

5. Health quality and pay-for-performance standards:

6. Educate and inform evolving health management practices and operational models:

7. Stakeholder collaboration, communication, and coordination:

8. Rational health policy development that supports viable business models and care delivery practices focused on personalized health care:

Assumptions Regarding Future Dynamics of Health Care Delivery

Factor 1: Gene-based and Other Molecular Tests are Routinely Used in Patient Management

Factor 2: US HTA and Reimbursement Infrastructure Sufficiently Enables Personalized Health Care

Factor 3: Prevention and Risk Assessment Approaches that Incorporate Genetic Testing are Standard Practice

Factor 4: Electronic Health Records and Decision Support Systems are a Mainstay in Hospital and Multi-physician Practices

Factor 5: Provider Education and Certification is Increasingly Tied to Health Care Quality and Best Practices Initiatives

Factor 6: Professional Societies Play a Key Role in Evidence Evaluation and Implementation of Knowledge into Clinical Practice

A Framework for Professional Societies to Play a Role in Enabling Delivery of Personalized Health Care

The Road Ahead: Enabling the Personalized Health Care Environment

The Role of the Academic Medical Centerin Advancing Personalized Health Care

AMCs as sources discovery in personalized medicine (T1)

AMCs role in biobanking and patient registries: important sources of discovery and validation for novel molecular tools for PHC

Resourcing the clinical validation of the next generation of personalized medicine (T2)

AMC-initiated T2 research studies

Conflict of Interest in T2 research at AMCs

AMCs as platforms to study implementation of PHC delivery (T3) and outcomes (T4)

Developing new economic models

Quality, Performance, and PHC

Health Professional and Public Education

Patient Centered Medical Home (PCMH)

Health information technology (HIT) and electronic health records (EHRs)

Payors as integrated partners in T3 and T4 research.

A Call for Specialized Centers as Catalysts Accelerating Genome Information to Medicine

Personalized Medicine and Disruptive Innovation:Implications for Technology Assessment

ABSTRACT

The Promise of Personalized Medicine

Technology and Innovation in Health Care

Disruptive Innovation in Health Care?

Administrative Barriers to Innovation

The Policy Maker’s Role

Exploring Potential Policy Options

Example: The Complex Development Path for a Potentially Disruptive Innovator

A Way Forward for Disruptive Innovations

Conclusions

Assessing Risk and Return:Personalized Medicine DevelopmentAnd New Innovation Paradigm

Introduction

Current Challenges in Productivity and Investment Returns

Early-Stage Funding Challenges in Stratified Medicine Development

Venture Philanthropy—Early-Stage Funding/Proof of Concept

Venture Philanthropy and Nonprofit Venture Affiliates

Venture Philanthropy and Nonprofit Venture Intermediaries

Risks and Impact on Return: De-risking Clinical Trials

Development Risk and Clinical Trial Design

Time/Cost Correlation

Venture Philanthropy—Clinical Trial De-risking Mechanisms

Venture Philanthropy and Patient Registry/Database

Venture Philanthropy and Clinical Trial Networks

Assessment and Recommendation

Investing in Personalized Health Care Innovation

Background and Purpose

Personalized Health Management:
A Geisinger View

Role of Professional and Medical
Specialty Societies in the
Era of Personalized Health Practice

The Role of the Academic Medical Center
in Advancing Personalized Health Care

Personalized Medicine and Disruptive Innovation:
Implications for Technology Assessment

Assessing Risk and Return:
Personalized Medicine Development
And New Innovation Paradigm

Venture Capital Investing in Healthcare^[1]