American Health Information Community

Personalized Healthcare Workgroup

Summary of the First Web Conference of this Workgroup

Thursday, January 4, 2007

PURPOSE OF MEETING

Following a preliminary meeting to orient the members on December 19, 2006, the purpose of this first official meeting was to receive and review informational presentations on issues and projects related to the charge of the Workgroup.

KEY TOPICS

1. Summary of the December Meeting

A summary of the orientation meeting in December was distributed by e-mail prior to January 4.

2. Introductory Remarks

Judy Sparrow, AHIC Director, summarized the rules for the Workgroup’s operations. In compliance with the rules of the Federal Advisory Committee Act, all work will be conducted in public, with meeting transcripts and summaries a part of the public record. Documents will be posted on the AHIC Web page, where future meetings will be announced. Members of the public are invited to listen to the Web broadcasts and a time at the end of each meeting is allocated for public input. The Personalized Healthcare (PHC) Workgroup is advisory to the AHIC, which makes recommendations to the Department of Health and Human Services (HHS) Secretary. HHS is responsible for implementing those recommendations that are accepted. The Workgroup meetings are to be conducted in such a way as to build consensus. Recommendations are expected to reflect consensus.

Co-chairs John Glaser and Douglas Henley restated the specific and broad charges of the Workgroup.

Specific Charge within the 12- to 18-month time frame

Make recommendations to the AHIC on the establishment of standards for reporting and incorporating genomic tests data into electronic health records (EHRs) and suggest incentives to encourage adoption of these standards. Genomic testing is becoming a routine part of the health care process, and standards are needed for use by the Federal Government and the private sector. The Workgroup also will consider privacy issues as well as how to reach consensus across diverse groups and interests on standards.

Broad Charge within the 2- to 3-year time frame

Make recommendations to the AHIC for the incorporation of genetic laboratory test data into EHRs and for the development and incorporation of analytical tools to support clinical decision-making at the point of care for the health care provider and patient. This builds upon the overall AHIC goal of interoperability.

Workgroup members had no comments or questions regarding their charge.

Prior to the beginning of the presentations, meeting participants introduced themselves, citing their current and past affiliations and briefly describing their interests and expertise related to the charge of the Workgroup.

2. Background Presentations

PowerPoint presentations are available at the AHIC Website.

Personalized Health Care Overview by Gregory Downing

Dr. Downing’s PowerPoint presentation described Secretary Mike Leavitt’s priorities for personalized health care. Health information technology (HIT) is a cross-cutting issue. His goal is to provide consumers and health care providers the means to make the delivery of health care more informed, individualized, and effective and to increase the choices presented to patients. His two-year objective is to establish concepts and priorities that support health care system transformation.

The PHC Workgroup has an important role in that transformation making recommendations for the adoption of standards for the inclusion of genetic test information in EHRs. The integration of HIT and genomic information will be transformative in its practice, as shown in the personalized health care pyramid.

HHS staff uses the term “genomics” in a broader context to mean not only DNA sequence but other features, including proteins, as being able to determine more mechanistically the basis in which health and disease have their origins. Complementing that are the advances in HIT, which make possible the aggregation and use of information.

Dr. Downing went on to review the history of the HHS planning process, which spanned a nine-month period. Various retreats, meetings, and discussions were held to promote the exchange of information across agencies on HIT activities, culminating in the decision at the September AHIC meeting to form a Workgroup dedicated to personalized healthcare issues.

In November, the Secretary released a Request for Information (RFI) soliciting opinions and ideas about the health care delivery system’s engagement in personalized healthcare planning. HHS staff also has engaged in community outreach activities with a number of patient advocate groups, and has been collecting information with the goal of developing a report for the Secretary relatively soon after the completion of the RFI on February 5. These information gathering activities are expected to assist with the identification of emerging opportunities, risks, and benefits for the incorporation of genetic test information into EHRs. Information is being obtained on the respective concerns of patients, physicians, and other providers as well as how data are being exchanged in ways that are meaningful and contribute to better health care.

Practical applications of medical genomic tests are already emerging for risk assessment, confirmatory diagnostic testing, and the selection of appropriate therapy. Many health care providers report that they fear the day in which genomic test information becomes a mainstay diagnostic test; they say that they do not have sufficient information or knowledge to use the information.

One of the advantages of genomic testing is that the structure of the genome provides a fairly simple lexicon of terminologies that can be used to describe unique sequences in our DNA. Furthermore, digital quality matches fairly well the exchange of digital information. A harmonized nomenclature system widely used for genes and diseases is continuing to evolve.

Efforts are underway by various groups to develop standards for technology platforms, whether for registration and market entry or for clinical diagnostic frameworks with laboratory applications. These efforts will facilitate the incorporation of the laboratory test data into genetic records and EHRs.

Dr. Downing summarized the Secretary’s goals:

1. Link clinical and genomic information to support personalized health care by establishing an interoperable public/private data network of networks to deliver information on individual medical outcomes and linking findings to genetic laboratory tests and establishing a common pathway for data integration through personal EHRs.

2. Support the appropriate use of genetic laboratory tests in medical practice by encouraging policies and practices that provide sufficient protections to consumers that genetic test information is used for their medical benefit and providing oversight of genetic testing to ensure analytical and clinical validity.

He concluded by mentioning the several advisory groups whose charge is related to the Workgroup and then described ongoing activities of the National Institutes of Health (NIH), the National Library of Medicine (NLM), the Food and Drug Administration (FDA), and other agencies.

Nationwide Health Information Network (NHIN) Overview by John Loonsk, Office of Interoperability and Standards Director

Dr. Loonsk presented an overview of the responsibilities and activities of the HHS Office of the National Coordinator for Health Information Technology (ONC). The national HIT agenda is about not only the establishment of standards but also the adoption and implementation of standards. This process constitutes a cycle or series of continuing cycles. The AHIC has identified opportunities for adoption or breakthroughs, which in turn have led to the establishment of a workgroup. The workgroups have developed three use cases to articulate the problems so that various other aspects of this national apparatus can work on them.

The three use cases focus on (1) consumer empowerment, a use case about registration information and medication history; (2) provider use of electronic lab results reporting; and, (3) in the area of population health, a use case around using clinical care data for public health biosurveillance. The EHR Workgroup recently requested another use case on emergency responders’ use of EHRs. ONC is anticipating that the AHIC will prioritize three additional use cases at the January meeting.

The NHIN is concerned with facilitating the movement of data between the different organizations in the network to advance the broad HIT agenda of having information follow the consumer and available for clinical decision-making.

This year the activities of the NHIN focused on these use cases. Four different consortia developed prototype architectures for how the different organizations involved in health care could connect to, transmit, and share data. Software development was contracted and the new software systems will be demonstrated at the January AHIC meeting. In 2007, the NHIN will engage directly with State and regional health information exchanges in the context of trial implementations; a new request for proposals soon will be announced to implement these trials.

Another part of this process is the Health Information Technology Standards Panel (HITSP), which is addressing several classic standards issues. HITSP is taking the AHIC use cases and harmonizing the standards associated with them. Another issue is identifying standards gaps, which is related to the charge of the PHC Workgroup. It is not necessarily HITSP’s role to fill standards gaps, but rather to identify them and work with standards bodies to fill those gaps. A third classic problem with standards is that their specificity has not been always adequate to accomplish easy interoperability or easy exchange.

Another issue is certification criteria. In 2006, certification criteria for ambulatory care EHRs were developed and tested. The certification commission has subsequently certified some 35 ambulatory care EHRs. This year the focus will be on certification for inpatient EHRs as well as specialty ambulatory care EHRs. Next year certification work will focus on updating ambulatory care and inpatient EHRs, as well as on network certification criteria.

Following the presentation, Dr. Loonsk answered several questions. In response to a query about whether the use cases dealt with adult or pediatric cases, he said that the use cases to date have not been sufficiently specific to differentiate between adult and pediatric medicine. That level of specificity is expected to occur at the standards harmonization stage; if there is a standard relevant in the pediatric domain to fulfilling the use case but not in the adult domain, or vice versa, that will be articulated.

In response to a question about how standards are adopted, Dr. Loonsk acknowledged the tension between encouraging the development of standards and using HITSP to make sure that there is a valid process for anointing the standards to be used. HITSP includes some 260 participating organizations. Not all the standards activities are about harmonization; working with standards development organizations to make sure the standards are available is important as well. He also clarified that the work on standards does not include quality issues or quality standards for laboratory testing. LOINC is widely accepted for laboratory standards.

A process question was raised: “Given that the membership of the Workgroup overlaps with other AHIC groups and various other advisory groups, how should these relationships be exploited?” The Co-chairs suggested that this question be addressed during the February meeting when the Workgroup members will begin to delineate tasks and a roadmap for completing them. Once the tasks have been identified, members can determine how best to leverage the variety of professional relationships represented in the group.

Finally, Dr. Loonsk clarified that HITSP has bylaws and an established process for making decisions. The Certification Commission for Health Information Technology is a public-private process. The NHIN has sponsored a series of activities to obtain public input.

Standards and Best-practice Guidelines Development by Janet Warrington

Dr. Warrington began by stating that her presentation was intended to describe the activities related to genomic technology and the development of standard controls and best practices in which she herself was involved, many of which are upstream of the immediate goals of the Workgroup. Most of these activities focus on RNA and DNA. She urged caution in the use of the term “standards”; staff of the National Institute of Standards and Technology avoids the use of the term as a noun. They use it as an adjective, for example, as in “standards controls.” The term should be used only in reference to a very specific work product.

She went on to describe the composition, process, publications, and expected deliverables of the External RNA Controls Consortium, Clinical and Laboratory Standards Institutes MM16 and MM17, the Clinical and Laboratory Genomic and Genetic Standards Group, the Organization for Economic Cooperation and Development Biotechnology Industry Advisory Committee, the Working Party on Biotechnology, the MicroArray Quality Control Project, the E.U. Project EMERALD, and the Microarray Gene Expression Data Society.

Ontology and Standards Development at NLM by Betsy Humphreys

Ms. Humphreys referred Workgroup members to an NLM Web page, which contains information about the library’s participation in various standards development efforts. NLM provides financial support and makes certain clinical terminology standards freely available for use, in some cases within the United States and in other cases internationally, including SNOMED Clinical Terms (CT) and LOINC. The LOINC group is interested in expanding LOINC to accommodate the transmission of genetic tests data for clinical reporting.

SNOMED CT is undergoing a transition in ownership from the College of American Pathologists to an international health data standards organization. That new organization will work with LOINC to clarify areas of responsibility between the two standards groups.

Marc Williams noted that current coding systems, such as ICD9 or -10 and SNOMED CT, are inadequate from the perspective of genetic disorders. Ms. Humphreys said that the change in the ownership of SNOMED likely will improve that situation. The Workgroup is in a position to promote action on the issue. SNOMED has not always had the resources and expertise to deal with genetics testing. Andy Wiesenthal said that Kaiser Permanente has expanded the SNOMED terms for internal use.

Tim O’Leary, Department of Veterans Affairs, questioned whether SNOMED is capable of dealing with the characterization of a whole human genome. A mutation nomenclature working group, more or less associated with the Human Rights and Governance umbrella, is working on some of these problems. He suggested that the Workgroup inform itself about the activities of this group.

Health Level Seven (HL7) Activities by Grant Wood, Clinical Genetics Institute, Intermountain Health

Dr. Wood explained the Institute’s goal is to link genomic and family medical history data to the EHR to provide advanced genetics-based tools for the clinician. Basic questions are, “How do we store genomic data?”, “What data do we store?”, and, “What format would be the most useful for the clinician?”

HL7 is an international community of health care clinical and informatics experts collaborating to create standards for the exchange, management, and integration of electronic health care information. The HL7 and clinical genomics mission is to create standards for communication between clinical and genomic data. The focus of the clinical genomics work is the personalization of the genomic data and linking these data to clinical information. He explained how this might work.

The genotype model describes various types of genomic data relating to a chromosomal locus, and is proposed to be the basic unit of genomic information exchange in health care built with personalized medicine in mind. The entry port of the model consists of a genotype locus but also will store individual alleles; biosequencing data such as DNA, RNA, or proteins; sequence variations like mutations, single nucleotide polymorphisms, and haplotypes; genoexpression data; and related clinical phenotypes such as diseases and allergies.

An important design characteristic is that this is a shared model or reusable component which works regardless of use case. At the same time, it allows the representation of fine-grained rod genomic data, thus accommodating the more complex use cases such as high-resolution tissue typing, somatic mutations, personalized expression profiles, or genomic based clinical trials.

The clinical phenotype in the model is an actual observation valid for the patient. The genotype model also will work for pharmacogenomic-based clinical trials and submissions. Dr. Wood believes that any integration of genomic data in the EHR must include the raw sequence data. Because our knowledge of the interpretation of clinically significant variants or mutations is being updated constantly, we need to have the full sequence data with every test result.

The genotype model is designed from the ground up to provide dynamic reporting of data for clinical decision support. The EHR could inquire about genetic testing for CYP2C9 or VKORC1, which are the two genes that have the largest impact on warfarin dose. If the testing has been done, the program uses an evidence-based algorithm to determine a starting dose.

If the genotype information is not available, the physician is prompted to order the test. The resulting sequence data from the test can be encapsulated in any format. Last fall the clinical genomics group became involved with the AHIC EHR Workgroup. Family history made the top 10 priority list of that Workgroup. Today, genomic information is not sufficient to make predictions for one’s future health. However, because family history can include environmental and cultural factors, Dr. Wood believes that it is the best tool to start with in personalized medicine. Genomic information can be added as it comes online.

An HL7 model for family pedigree also exists and a number of applications are in use. The Avon Comprehensive Breast Evaluation Center at Mass General has incorporated the family history model into a cancer risk application. The Harvard Partners Center for Genetics and Genomics stores genetic data in its electronic medical records (EMRs). Most health care providers will use external reference laboratories for genetic testing. In order for structured test results to transfer in these cases, interfaces will need to be established between laboratories and provider organizations. Intermountain Health Care and Harvard Partners currently are collaborating to establish an inter-institutional interface for passing structured genetic results.

The interface will enable further testing of these standards and may serve as a prototype for other organizations. The goal is to create interoperability across the country.

Confidentiality, Privacy and Security (CPS) Workgroup by Jodi Daniel, Office of Policy and Research

Dr. Daniel is the ONC Director of Policy and Research. Privacy and security of health information is one of the prominent issues at ONC. She summarized the basis for the establishment of the CPS Workgroup and described its relationship to the PHC Workgroup. The CPS Workgroup was formed in May 2006 at the recommendation of three of the initial AHIC Workgroups Chronic Care, Consumer Empowerment, and EHR. The three Workgroups found that they were discussing similar CPS issues but did not have the expertise to make specific recommendations. They recommended that the CPS Workgroup have representation from the National Committee on Vital Health Statistics, provider groups, State public health officials, Federal agencies, consumer groups, and technology.

The Broad Charge was to make recommendations to the Community regarding the protection of personal health information, in order to secure trust and support appropriate interoperable electronic health information exchange. The CPS Workgroup began meeting in August 2006. Its first topics were identity proofing and user authentication. Hearings were held and the Workgroup members agreed to tackle patient identity proofing first and left user authentication for a later time. Identity proofing is the process of providing sufficient information, history, credentials, and documents to establish and verify, correctly and accurately, the identity of an individual for use in electronic environments.

The first set of recommendations from the CPS Workgroup will be considered by AHIC at the January 23 meeting. Dr. Daniel suggested that as CPS issues are raised in the deliberations of the PHC Workgroup, the Workgroup may defer to the CPS Workgroup or consult with it. An ongoing dialogue is expected. Dr. Daniel indicated her intent to follow the proceedings of the PHC Workgroup and to facilitate coordination as necessary. In addition to coordination at the staff level, Deven McGraw is a member of both workgroups.

Greg Downing then commented at length, explaining that ONC did not intend to create specific rules to protect the CPS of laboratory tests; rather, the interest was in overall protection of the EHR. However, new technology and the feasibility of linking various information sources have advanced since the enactment of the Health Insurance Portability and Accountability Act (HIPAA). HHS staff is attempting to identify the gaps in the current levels of protection. This is a complex area of federal and state laws, regulations and statutes. The results of a gap analysis are critical to both workgroups.

De-identification of Data and Privacy Concerns by Amy McGuire

Dr. McGuire began by defining several terms and referencing The Stored Tissue Issue by Bob Weir and Robert Olick. Identified data is information attached to traditional identifiers such as name, address, Social Security number, and medical record number. When samples or genetic DNA information are imbedded in the EMR and connected to phenotypic information that would identify an individual easily, the data are identified.

There are two ways to de-identify data. Data can be linked to traditional identifiers through a numeric code, with the key to the code secured and protected, or data can be made anonymous, meaning that they cannot be linked to traditional identifiers. It has been assumed that de-identification is an appropriate means of protecting privacy, but two forms of protection for identifiable data come under the common rule for research involving human subjects (45 CFR 46) and HIPAA. The common rule applies to research involving human subjects, and a human subject is defined as a living individual about whom an investigator obtains either data through intervention or interaction with the individual, or identifiable private information.

A 2004 Office for Human Research Protections guidance document stated that genomic data are not identifiable under the common rule if the data are collected in a coded fashion, and the investigator is prohibited from deciphering the code. According to another 2004 HHS guidance document, blood, tissue, and the results of an analysis of blood or tissue are not in and of themselves considered individually identifiable health information.

More recently, the questions have been raised on whether DNA is itself a unique identifier and whether there needs to be some reconsideration of what is considered identifiable in the case of genetic data that are not linked to other types of identifying information.

Dr. McGuire referred to an article by Owen and Altman published in Science in 2004. As the technology develops to secure data, it also develops to breach security measures. She suggested that there are three areas in which privacy protection can be increased. One is increased data security. Another is legislation. Two different types of legislation are relevant. The first is limiting access to genetic information. One example of this is through HIPAA. If we broaden our interpretations of what is identifiable, private, and protected health information under HIPAA, access to genetic information could be limited. There may be other State or Federal legislative mechanisms that could be used as well.

The second type of legislation is prohibiting discrimination on the basis of genetic information. Some States have prohibited both employment and insurance discrimination on the basis of genetic information.

The third avenue for increasing privacy protection is to do more around consent. Well-designed consent processes and forms and an overall transparency are essential in order to encourage people to participate in genetic testing and genetic research. We should be open about what the privacy risks are, to the extent that we know them, and should have people decide voluntarily whether to assume those risks. The risks are difficult to define, because we do not know what the future uses of genetic information may be.

Discussion indicated that the information presented by Dr. McGuire should be distributed to the CPS Workgroup. Dr. Daniel said that the presentation raised issues of concern to both Workgroups. The more technical issues related to genetics probably should remain under the purview of the PHC Workgroup, but data security is included in the charge of the CPS Workgroup. In any case, the workgroups will need to coordinate.

Ms. Humphreys pointed out that DNA cannot be protected in a local environment. NLM has been involved in activities to assist with DNA identification of victims (World Trade Center and Katrina), and in developing methods for quality control through the National Center for Biotechnology Information.

Several members agreed that increased legal protection was both possible and necessary, to protect against not only discrimination in employment and health insurance but also forensic use and the possible encroachments of Federal security initiatives.

Dr. McGuire pointed out that some institutions have established Internal Review Board practices that are more protective than the interpretation of the common rule.

3. Steps to begin Workgroup Deliberations

The next meeting is scheduled for February 21, 2007. The meeting is expected to be interactive as members begin to delineate the overall agenda and tasks for the Workgroup. Decisions about what to undertake will be affected by the available staff resources.

The Co-chairs offered several topics for consideration at the next meeting, beginning with the “standards agenda”: the range of areas (family history, transmission, and so on) in which standards should be developed and applied, the landscape of standards, implementation of standards, gaps in nomenclature; criteria for assessing standards, criteria for framing the Workgroup’s recommendations, standards to apply to use cases, and privacy issues unique to the charge.

Other ideas included surveying existing standards efforts to identify conflicting activities and conducting an environmental scan to identify near-term advances in diagnostic tests or other genetic breakthroughs; for example, the recent American College of Obstetricians and Gynecologists recommendation for screening pregnant women for Down syndrome. The Workgroup also needs to establish boundaries for its work in order to know when the work is completed.

Dr. Williams asked that his use cases on family history be distributed to the Workgroup members. Staff agreed to do so. Dr. Williams agreed to assembly information on the current status of representation of family history in the EHR.

Staff Action Item #1: The staff will work with the Co-chairs to prepare an agenda for the February meeting.

4. Public Comments

None.

5. Adjournment

Rob Kolodner, the ONC Director, was present at the conclusion of the meeting and thanked the participants for their willingness to serve on the Workgroup. After covering all agenda items, the meeting adjourned at 3:45 p.m.

SUMMARY OF ACTION ITEMS

Staff Action Item #1: The staff will work with the Co-chairs to prepare an agenda for the February meeting.

Meeting Materials

Agenda

Proposed Workgroup Charges

Background Documents:

• Lin Zhen, Owen Art B., Altman Russ B. Genomic research and human subject privacy. Science. 2004 (305), 183.

• McGuire Amy L., Gibbs Richard A. No longer de-identified. Science. 2006 (312), 370371.

• Shabo, Amnon. Clinical genomics data standards for pharmacogenetics and pharmacogenomics. Pharmacogenomics. 2006 (7), 247253.

PowerPoint Presentations:

• Personalized Healthcare Overview, Gregory Downing

• Personalized Healthcare - Operationalizing NHIN and Related Activities, John Loonsk

• Standard Controls and Best Practice Development for Genomic Technologies, Janet Warrington

• Related NLM Activities, Betsy Humphreys

• Introduction to HL7 Clinical Genomics, Grant Wood

• De-identification of Data and Privacy Concerns, Amy McGuire

Personalized Healthcare Workgroup

Members and Designees Participating in the Web Conference

Disclaimer: The views expressed in written conference materials or publications and by speakers and moderators at DHHS-sponsored conferences do not necessarily reflect the official policies of the DHHS; nor does mention of trade names, commercial practices, or organizations imply endorsement by the U.S. Government.