Skip Navigation

Pre-decisional Draft v 2.2

 

Background Information pertaining to Clinical Decision Support Proto-Recommendations

The topic of clinical decision support (CDS) cuts across the missions of many of the AHIC Workgroups, and each AHIC Workgroup has unique expertise and perspectives to address different aspects of CDS. For this reason, an Ad Hoc CDS Planning Group was formed in May 2007 to ensure that a common framework to address CDS would be used across Workgroups. The envisioned result was a vet of coherent and complimentary contributions that would result in recommendations to accelerate the implementation of robust and workflow-sensitive CDS interventions that will drive measurable improvement in key health care outcomes. Membership in the CDS planning Group includes representatives of the following five Workgroups that expressed significant interest in this topic:

  • Consumer Empowerment

  • Electronic Health Records

  • Personalized Healthcare

  • Population Health and Clinical Care Connections

  • Quality

Working Definition of Clinical Decision Support

  • CDS provides clinicians, staff, patients and other individuals with knowledge and person-specific information, intelligently filtered at appropriate times, to enhance health and health care.

  • CDS encompasses, but is not limited to, computerized alerts and reminders, clinical guidelines, order sets, patient data reports and dashboards, documentation templates, diagnostic support, and clinical workflow tools.

The CDS Ad Hoc Planning Group’s Approach

The Ad Hoc CDS Planning Group has held four meetings to date. In an effort to facilitate CDS efforts across Workgroups, a strategy was laid out to develop a comprehensive set of “proto-recommendations”, which are being directed to the attention of the five Workgroups for deliberation at Workgroup meetings during February and March 2008 for subsequent development into formal recommendations.

CDS-related efforts internal to the government are to be coordinated through a multi-stakeholder CDS Collaboratory coordinated by ONC, with formal co-sponsorship by the Agency for Healthcare Research and Quality (AHRQ) and the Personalized Healthcare Initiative. This group will build on a scan of government agency CDS-related activities conducted in 2007.

Areas of Emphasis

The following seven emphasis areas emerged to shape the generation of recommendations, and can be viewed as a proposed map of the domains that should be covered. The CDS Ad Hoc Group has identified several actionable recommendations in each area of emphasis:

  • Repository of Computable Rules/Clinical Practice Guidelines

  • CDS Oversight/Accreditation/Evidence/Data Quality/Transparency

  • Integration with Electronic Health Records (EHR) Systems and Incentives for Adoption of CDS Systems

  • Workflow

  • Ambulatory Care

  • Consumer Preferences

  • Driving Measurable Progress toward Priority Performance Goals

Other Relevant Organizations and Projects

In 2005-06, the Office of the National Coordinator for Health IT (ONC) in partnership with the Agency for Healthcare Research and Quality (AHRQ) supported the development of a CDS Roadmap. The American Medical Informatics Association (AMIA) convened experts in informatics, software, and evidence development from industry, academia, and government to develop a national plan of action for CDS. The CDS Roadmap identified three essential elements for achieving the promise of CDS: 1) access to the best knowledge available; 2) widespread adoption and effective use; and 3) continuous improvement of knowledge and CDS methods. The CDS Roadmap was presented to AHIC in the summer of 2006 and provides a useful framework from which to plan and organize efforts to address the barriers and enablers of CDS and related issues. The Roadmap was a major background resource to the development of recommendations offered in this document.

Overarching Themes and Recommendations

Factors that will accelerate the adoption and effective use of CDS include:

  • Adoption of national priorities for care improvement and explicit linkage of CDS to driving measurable improvements in these targets

  • Harmonization of quality measures and reporting on those measures

  • Continued implementation of provider payment mechanisms to reward care that is safe, high quality, efficient and coordinated

  • Presence of organizations, tools and resources, standards and information that can assist the small physician practice, hospital and other providers in implementing CDS effectively

  • Continued implementation of electronic health records

  • Implementation of high degrees of EHR interoperability

The CDS Ad Hoc Group has identified several actionable overarching recommendations.

Recommendation OA1.1: A public-private task force should be established to develop a plan and provide guidance for implementing the CDS recommendations.

Recommendation OA1.1.1: The task force should be composed of members of the Ad Hoc CDS Planning Group, AHIC-sponsored entities (CCHIT, HITSP) and organizations that represent the provider, payer, guidelines developers, medical informatics experts, consumer and life sciences, and clinical information system and CDS supplier communities and others.

Recommendation OA1.1.2: The task force should identify actions that its constituents can take together to advance the enablers (e.g., education, incentives and business cases) that will further the adoption and effective use of CDS, in addition to implementing the specific recommendations outlined below.

Recommendation OA1.1.3: The task force should develop strategies for increasing the health care community’s understanding of the importance of CDS.

1. Repository of Computable Rules/Clinical Practice Guidelines

A standard CDS Knowledge Repository or “knowledge service” of common computable rules, algorithms, and agreed upon clinical practice guidelines offers numerous potential benefits: 1) avoid duplication of effort in re-creating rules for specific activities used in decision making (e.g., dosing algorithms and routine management options), 2) allow vendors/software developers to use common resources and build on them to enhance services, 3) indirectly facilitate standards development for CDS incorporation/integration with electronic health records (i.e., if common pathways for algorithms are used for basic functions, adoption and usability may increase if standards are available), and 4) potentially, provide a structure to store and exchange national feedback and experience with various CDS interventions.

Recommendation 1.1: Establish a standard CDS Knowledge Repository model that is freely and widely available to the public for collecting, organizing, and distributing computable clinical knowledge and CDS interventions, targeted toward direct incorporation of new CDS capability into electronic health record systems. The repository will serve as an aggregation of human readable logic for guidelines, rules, and alerts, and computable forms of the same. Development of knowledge in specific clinical areas will be driven by national high priority needs.

Recommendation 1.1.1: Establish an expert multi-stakeholder panel to develop the necessary structure, standards, and design elements for a CDS repository.

Recommendation 1.1.2: Produce a model repository demonstrating feasibility of the concept. Such a model repository would focus on a starter set of clinical guidelines and would allow its subscribers to search by clinical topic, review and download specific pertinent CDS interventions for use in electronic health record systems.

Recommendation 1.1.3: Work with HITSP to establish any necessary data and formatting standards.

Recommendation 1.1.4: Consider requiring or encouraging Medicare quality demonstration project sites to make use of and contribute to the repository.

Recommendation 1.1.5: Work with CCHIT and other bodies to move toward adding the ability to download from such a repository to the criteria for certified EHRs, to encourage vendors to make their standard CDS tools compatible with the developed format.

Recommendation 1.1.6: Work with NQF and other bodies toward ensuring that clinical practice guidelines and quality measures take note of practical implementation considerations and publish relevant CDS interventions in approved repositories using this format.

In order to have a repository of clinical practice guidelines and associated useful interventions, it is necessary to have a standard practical format for representing CDS interventions of various types in the repository. Examples of interventions include computerized alerts and reminders, clinical guidelines, order sets, patient data reports and dashboards, documentation templates, diagnostic support, clinical workflow tools and other computed information that assists clinical decision making.

Recommendation 1.2: Identify a specific set of CDS intervention types, and create a standard format for representing each type in a standard CDS Knowledge repository.

Recommendation 1.2.1: Develop the necessary structure, standards, and operations for a standard practical CDS intervention format.

Recommendation 1.2.2: Work with CCHIT, EHR vendors, and other organizations to ensure that EHR vendors can and should make their standard CDS tools compatible with the developed format, so as to facilitate easy downloading of interventions from a CDS intervention repository.

Recommendation 1.3: Establish a regular mechanism for vetting and authorizing particular persons and groups to edit and enter interventions into the standard CDS Knowledge repository.

Recommendation 1.3.1: Establish standard practices for authorization of individuals and groups to add or modify information in an approved CDS repository.

Recommendation 1.3.2: Establish appropriate requirements for description of the origin of the intervention, evidence strength, and support by approved clinical and quality organizations in CDS interventions stored in approved CDS repositories.

Recommendation 1.4: Develop a national repository of public domain evidence-based best practices for CDS, including knowledge elements, implementation guides, and assessment techniques, to establish a common ‘floor’ of CDS (initial focus areas may include USPSTF guidelines, CMS CMP focus areas of DM, CHF, CAD, Immunizations, etc.)

Recommendation 1.4.1: Enhance the national research agenda to support discovery of best practices for knowledge management related to CDS.

Recommendation 1.4.2: Enhance the national research agenda to support discovery of best practices for personalized medicine in relation to CDS.

2. CDS Oversight, Accreditation, Evidence, Data Quality, and Transparency

Currently, there are no clear guidelines or oversight for the knowledge or algorithms that underpin CDS interventions. The source, accuracy, validity, and comprehensiveness of the knowledge and algorithms are often not readily apparent to the purchaser or end-user of the CDS interventions. In the areas of diagnostic/prognostic decision-making for some computer based algorithms, FDA regulatory authority has been considered. There is a need for CDS knowledge transparency or review that provides assurances that the information is used properly but that does not create new obstacles for software development.

While recognizing that there are often proprietary considerations for commercially-developed tools, important quality and safety issues apply to the field in general, such as: What are reference sources for the information used in the tools? What is the credibility of those sources? How often is information updated? What processes ensure currency of information provided?

Greater transparency may promote higher rates of adoption in health care practices as a result of improved user/purchaser confidence in CDS quality and utility. While the establishment of formal oversight or credentialing process is premature, near term steps include: 1)examine the options and assess their pros and cons; 2) consider what should be accredited (i.e., groups and processes, guidelines themselves); 3) consider how knowledge and algorithms can be reused or modified to meet local circumstances.

Recommendation 2.1: To enhance understanding about the evidence used in developing clinical decision support tools, AHIC should request information from CDS developers about the content development questions outlined above. The information should be collected in a way that addresses proprietary concerns.

Recommendation 2.2: AHIC or HHS, working with other interested parties, should work with CDS developers to build consensus on, and mechanisms for, informing CDS users and purchasers in a useful way about the answers to the important questions outlined above regarding CDS content.

3. Integration with EHR systems and incentives for adoption of CDS Systems

Health information technologies designed to improve clinical decision making are particularly attractive for their potential to address the growing information overload clinicians face, and to provide a platform for integrating evidence-based knowledge into care delivery. However, despite the theoretical and demonstrated benefits of such technologies, the existing literature shows low adoption of these technologies.

The following recommendations will ensure that optimal, usable, and effective clinical decision-support is widely available to health care providers, patients and individuals to enable high quality, cost-effective health care decisions.

It is noted that development of a standard CDS Knowledge Repository, as noted in Recommendation 1.1, is critical to advancing integration of CDS into EHR systems. Equally critical is workflow integration, discussed in detail in Section 4 to follow.

Recommendation 3.1: Incorporate certified CDS technologies into health care information technology, and clinical processes. (Note: Need to specify whether this is calling for an independent/separate certification.)

Recommendation 3.2: Provide financial and other incentives to those clinicians who achieve demonstrated quality, efficiency and safety outcomes through use of CDS technologies. Clinicians include physicians, nurse practitioners, pharmacists, dentists, and allied health practitioners. This may be accomplished through the creation of a variety of quality dashboards or feedback mechanisms about the utilization of CDS alerts.

Recommendation 3.3: Support research to develop new methods that leverage CDS in EHR and Personal Health Record (PHR) systems.

4. Workflow

Advice generated by CDS resources must fit into clinical workflow or it will not be heeded. It is necessary, but not sufficient for CDS systems to generate advice that is clinically valid. The advice must arrive “in the right place at the right time” and be delivered to the right person who understands the advice and is in a position to act on it.

The recommendations here are targeted toward ensuring that practice guidelines and clinically-important knowledge are developed with consideration of how the knowledge can be best delivered and applied in real-world workflow.

Recommendation 4.1: Develop a model which depicts standard workflows and types of CDS interventions that are applicable to each workflow. The generalized workflows would be created at a high level, rather than depicting workflows for a specific clinical condition or procedure. A use case could be employed to communicate the model. Example workflows include: Pre-Visit, Start of Visit/Intake, Clinician’s History & Physical (H&P) and plan, Documentation, Orders/RX, Consult requests, Therapies/Procedures, End of Visit/Checkout, RX Dispense, Results arrive, Post-Visit/Home Care.

Recommendation 4.2: Develop strategies to ensure that guidelines, when developed, include recommendations for which CDS standard intervention types and workflows are best suited for advice computed from the guidelines.

Recommendation 4.3: Build a resource of shared information about best-practice CDS implementation methods what works well, what does not, and lessons learned. Such a resource may have static elements, such as a reference website or written guide, and dynamic elements, such as a community for rapid exchange of information and ideas. Make use of existing programs and resources that have worked in this area, such as IHE, the Davies program, DoQ-IT, and others.

Recommendation 4.4: Direct Medicare demonstration programs related to quality improvement and HIT to document their successful and unsuccessful experiences with CDS, indicating objectives, the type of intervention used and the workflow to which applied, details of the implementation, and user acceptance and impact.

5. Ambulatory Care

Significant attention is needed to stimulate CDS adoption in ambulatory care settings. In large institutional settings, centralized efforts to deploy CDS can reach thousands of practitioners, and often the high costs of these deployments can be justified over the very large number of patients they will affect, and the high consequence of untoward clinical events that may be averted. In ambulatory care settings, and in particular the small office environment, there are additional challenges that can become barriers to CDS adoption. Since CDS is an enhancement to basic EHR systems, CDS adoption in ambulatory settings is limited by the lower adoption rate of EHRs in these settings. The costs and challenges of implementation are distributed over a smaller number of patients, and there may be a mistaken perception that the benefits of CDS are less profound.

It is noted that development of a standard CDS Knowledge Repository as noted in Recommendation 1.1, and the development of appropriate incentives as noted in Recommendation 3.3, are critical to advancing the adoption of CDS in ambulatory care settings.

Recommendation 5.1: Develop methods to train clinicians in the effective use of CDS, particularly when used within an electronic health record.

Recommendation 5.1.1: Develop curricula in clinical informatics and continuous quality improvement with information technology for medical and health education, and continuing medical education programs.

Recommendation 5.1.2: Enhance the culture of improvement as a core component of practice through program initiatives in local and state medical societies, professional societies, and Federal initiatives (National HIT Resource Center, DOQ-IT U., and others).

Recommendation 5.1.3: Develop a mechanism by which the effective use of CDS is factored into clinician certification and licensure processes.

6. Consumer Preferences

Processes, tools sets, and advances in clinical analytics can significantly improve care and personal health.Information about patient preferences and attributes should be incorporated into the clinical decision support process. It is important to recognize that health-related decisions are made by clinicians caring for patients, by consumers managing their own health, and through shared decision making processes.

The following recommendations will maximize opportunities for more consumer centric health care that is individualized to consumer needs and preferences.

Recommendation 6.1: Develop interoperability standards for a minimum data set which can be incorporated into certified EHRs and characterizes the most important personal attributes that contribute to individualized care. (Examples include: age, gender, family history, prior illness, current illness, genetic markers, current medications, allergies, and end of life instruction.) Interoperability standards should span EHRs and PHRs.

Recommendation 6.2: Support research on approaches that different types of consumers are most likely to use with respect to accessing and applying decision support in managing their own care.

Recommendation 6.3: Incorporate consumer preferences for accessing and applying decision support into CDS systems used by clinicians.

Recommendation 6.4: Deploy and evaluate known methods of supporting communication between patients and their clinicians for the purposes of shared decision making regarding health and care. Personal attributes and attitudes that contribute to decision processes and outcomes should also be identified. Consumers should be included in the evaluation process.

7. Driving Measurable Progress toward Priority Performance Goals

The National Quality Forum (NQF) has initiated efforts to establish national priorities and performance goals for several common chronic conditions, and to identify existing measures that can be used to assess progress in meeting each performance goal. NQF also envisions creating a feedback loop for measure developers to help identify areas that need improved measurement for national priorities. To further close the gap between desired and actual performance, the impact of prioritizing conditions on a national level could be evaluated and reported using a national dashboard of metrics. If measure development, payment policy, CDS, and evaluation efforts across various stakeholders could be better aligned, system level changes to achieve a high performance health care system would be more likely to succeed.

The set of activities that could help address these components includes the following:

  • Identify national priority clinical conditions for which CDS should support better outcomes.

  • Identify and monitor the results of performance metrics for the high priority conditions.

  • Identify and assess the effects of CDS-related efforts intended to drive improvements in the performance metrics for the high priority conditions.

  • Develop a framework for optimizing how CDS interventions to improve (and data collection approaches to assess) targeted outcomes can be incorporated into clinical workflow and care processes. (For more information, please refer to Recommendation 2.1 in the AHIC Quality Workgroup’s recommendations letter dated March 13, 2007.)

  • Disseminate the results of the above activities to pertinent stakeholders, and foster information sharing and collaboration to accelerate progress on applying CDS to performance improvement nationwide on targeted metrics.

Recommendation 7.1: The National Quality Forum’s National Priority Partners Committee intends to prioritize conditions in the first half of 2008. By September 30, 2008, representatives of federal agencies (“the CDS Collaboratory”) should identify priorities for federally funded CDS initiatives based on the National Quality Forum’s set of national prioritized conditions and on existing government priority areas with relevant funded programs, particularly in areas such as pay for performance, public health, and personalized health care. The CDS Collaboratory should also identify methods of data collection and reporting of the metrics for the purposes of developing a feasible evaluation plan to monitor the impact of government funded CDS programs. The CDS Collaboratory should make public its list of top priorities for CDS initiatives.

Recommendation 7.2: Once the priorities and evaluation plan from Recommendation 7.1 have been completed, the CDS Collaboratory should facilitate alignment of the various initiatives, methods and metrics within the federal government. The CDS Collaboratory should establish a mechanism to periodically measure progress within the federal government against the metrics and provide performance measurement feedback to the CDS initiatives. The CDS Collaboratory should then publicly disseminate the results of the progress measurement as well as information about federal initiatives and programs.