AHRQ Research Summit on Learning Health Systems: Executive Summary

September 15, 2017
5600 Fishers Lane, Rockville, Maryland

On September 15, 2017, the Agency for Research and Quality (AHRQ) convened, in Rockville, MD, a meeting of experts to identify and discuss aspects of a learning health system (LHS). An LHS features methods and strategies, especially using data and evidence, for continually making improvements in a health care system. Health care delivery organizations striving to become LHSs also develop a culture and leadership committed to learning and growth.  The meeting participants shared experiences relating to attempts to improve systems, discussed attributes of a successful LHS, and received descriptions of health care delivery organizations attempting to become an LHS. They engaged in activities designed to generate ideas regarding research, tools, training, and data. They focused on how AHRQ could support development of those ideas to advance learning health care organizations.

Welcome and Goals of the Meeting

Gopal Khanna, M.B.A., Director of AHRQ, expressed a strong belief in the power of data upon which researchers can base their work. Clinical data can support decisions, suggest options, and, along with insight, inform clinical practice leading to better outcomes. Health care organizations must provide leadership and culture that support utilizing data and evidence to improve. They must adapt, adopt, and apply evidence, and this should become systemic.

The three main goals for the meeting were:

• To identify aspects of learning health care delivery organizations and practices that are necessary to achieve improved performance objectives relating to quality, safety, cost, and the workforce.
• To recognize roles of the meeting participants in helping to create an LHS.
• To explore how AHRQ can act as a catalyst for the advancement of learning health care delivery organizations and practices.

Discovering the Ingredients for Improvement in Health Care Delivery Organizations

Daniel Wolfson, M.H.S.A., the American Board of Internal Medicine Foundation, served as facilitator of the meeting. He asked the meeting participants to consider the following: “Think of a time when your organization, team, or individual performance experienced improvement. What conditions or assistance helped that improvement?” The meeting participants produced many responses, including the following:  There must be clear lines of authority and leadership; there must be a champion who can obtain needed resources; there must be a clear purpose; and good, measureable data must be obtained. Regarding data, the participants stressed issues of patient-centeredness and patient need. The program should be able to access evidence for what does or does not work. Standards and standardization of data are needed. 

Conversations With Health Care Leaders: Vision of an LHS

Mary Naylor, Ph.D., R.N., University of Pennsylvania School of Nursing, and Michael McGinnis, M.D., National Academy of Medicine, discussed their experiences and perspectives relating to the LHS, engaging in a conversation before the meeting participants. Dr. McGinnis referred to a National Academy of Medicine committee and report that envisioned an LHS as including science, informatics, incentives, and culture in alignment, leading to improvement and innovation. He noted recent progress in those four areas. Dr. Naylor cited accelerating efforts to determine how to help local systems use evidence. She stressed a need for humility among researchers, recognizing that research teams do not have all the answers. Partnerships need to evolve.

The meeting participants raised many issues in response to the conversation. The LHS implements evidence, improves care, and performs effectiveness research. There is a need for a patient focus and for a longitudinal perspective. Following patients through transitions can produce learning. We need to build infrastructure that delivers best care, helps in reporting, and generates new knowledge. Delivery organizations need to make available information that supports strategic efforts, as in evaluating new medical products. We must transform data sets to make them more useful, and we need more data from the community sector (social and behavioral needs).

Interview: Learning Health Care Delivery Organizations in Practice

Karen Feinstein, Ph.D., Jewish Healthcare Foundation, interviewed two experts with deep experience in these issues, Lucy Savitz, Ph.D., Kaiser Permanente Northwest, and Peter Pronovost, M.D., Johns Hopkins Medicine. The respondents focused on current practices in health care delivery organizations and made many points.

Dr. Pronovost cited structure and culture as keys to success. An LHS should not be about individual projects but rather about a larger structure. Quality improvement is about processes, whereas the LHS is about structure and culture, as in supporting clinicians to engage in teams for learning. The successful LHS features a culture of respect and a hunger to learn. Clinicians today lack many skills needed for team care. An LHS must present new narratives and remove dysfunctional narratives. Technology is an important enabler, yet it will be insufficient if we do not consider a sense of purpose. Dr. Pronovost proposed creating a management structure for quality based on the concept of the fractal, in which simple rules are repeated at different levels. He proposed that administrators focus on costs and short-term value while considering smaller actions and investments. As for Federal funding issues, Dr. Pronovost cited a need to embed funding concerns in a leadership’s idea of the delivery system, leading to investment in infrastructure to support the LHS. The system needs to drive down costs and grow productivity. As an example of successful AHRQ work, Dr. Pronovost cited support for efforts to reduce central-line blood stream infections.

Dr. Savitz cited a need to change the business model to get to success. Embedding procedures in the LHS requires determining how best to invest in research. We must define terms and determine who governs how data are used and who owns the data. Dr. Savitz called for new kinds of information (e.g., social determinants) and the science to support them. She noted the problem of fragmentation as we attempt to address an issue that is broad. She called for spending more time with care teams. A clinical program infrastructure should feature connections across the care continuum and a structure to support change. We must create strategic alignments and listen to voices from different parts of the organization. As for Federal funding issues, Dr. Savitz cited a need to align the timing of grant funding with research opportunities and actions. As an example of successful AHRQ work, she cited the Agency’s support for field-based delivery system research via Accelerating Change and Transformation in Organizations and Networks (ACTION) partnerships.

AHRQ Exemplars

Six representatives from AHRQ presented brief descriptions of current AHRQ programs and resources that are aligned with the LHS idea. These were (1) CDS Connect: Using Clinical Decision Support To Move Evidence into Practice, (2) Comparative Health System Performance Initiative: Identifying and Understanding High-Performing Systems, (3) Comprehensive Unit-Based Safety Program: Accelerating the Adoption of Evidence-Based Practices To Prevent Healthcare-Associated Infections, (4) EvidenceNOW: Advancing Heart Health in Primary Care Practices, (5) Learning Health System Competencies: Training the Next Generation of Researchers, and (6) Measurement-Powered Quality Improvement. Learn more about these presentations and projects.

Development of a Shared Purpose

With consideration given to AHRQ’s competencies and capacity to address needs, meeting participants were asked to develop a shared purpose that the Agency might promote and support as it collaborates with learning health care delivery systems.

Participants brainstormed many purposes and, through an iterative process, developed the following single purpose: 

• In collaboration with health care delivery organizations, AHRQ will develop the science and provide tools and training to support creation and development of learning health systems that can generate, use, and share evidence and best practices to systematically improve patient care and health outcomes.

Generating Ideas: Open Space

Through a series of co-creation exercises, attendees developed and prioritized ideas regarding research, tools, training, and data that AHRQ could produce to advance learning health care organizations. Participants brainstormed ideas and were asked to develop further the higher rated ideas. For each idea, participants worked in small groups to develop/describe future AHRQ initiatives that might address it or support it. The descriptions addressed the following aspects: (1) A problem statement/need, (2) AHRQ actions, (3) Audience/stakeholders, (4) Partners with AHRQ, and (5) Evaluation. Once these descriptions were developed, Summit attendees again voted to identify the most critical ideas.

The top five ideas as voted on by Summit participants included:

1. Develop a business case for investing in LHS and disseminate to delivery system chief executive officers.
2. Catalyze the development of a public commons in the form of a common data platform that accelerates sharing of data, tools, and resources.
3. Curate a guide to becoming a LHS with practical guidance and tools for health care delivery organizations.
4. Accelerate the spread of evidence-based best practices across LHSs.
5. Advance patient-reported outcomes and patient experience measures to support the ability of LHSs to pursue meaningful patient-centered improvement.

Additional Ideas

Toward the end of the meeting, the facilitator asked participants to share additional bold ideas for AHRQ that may not already have been discussed. The participants offered many suggestions, including the following:

• Go beyond AHRQ’s funding of people to do things to consider new ways to effect change. Consider other relationships with health care organizations to stimulate change, for example, matching grant programs, which feature greater collaboration and a vested interest in outcomes.
• Define most LHS activities as normal operations rather than research. These normal operations should feature “trials” that do not require oversight of the common rule or an IRB. The LHS should use the highest quality methods for developing evidence. That should not ordinarily be called “research.”
• Address all social determinants of health, respecting the health-and-wellness concept. We should foster the 360-degree, whole-person system.
• Change funding mechanisms to align better with innovative, rapidly changing health systems that demonstrate capacity and resources to become an LHS.
• Large technical companies including Amazon, Apple, and Facebook are doing work in areas such as electronic health records, health care, and interoperability. Perhaps AHRQ could form collaborations with such industry groups.