3
The Learning Health System
Key Points Highlighted by Individual Speakers1
- A learning health system (LHS) aligns science, informatics, incentives, and culture for continuous improvement, innovation, and equity. (McGinnis)
- A quality health system is patient engaged, safe, effective, equitable, efficient, accessible, transparent, measurable, secure, and adaptive. (McGinnis)
- Networks formed within the National Academy of Medicine Leadership Consortium are developing strategies to scale LHS principles and activities related to culture, informatics, incentives, and science. (McGinnis)
The second session of the workshop featured an overview of the learning health system (LHS) concept and efforts to shift health systems toward such a model. Odette Harris, professor of neurosurgery at Stanford University and deputy chief of staff for rehabilitation at the Veterans Affairs Palo Alto Health Care System, moderated the session.
___________________
1 This list reflects the rapporteurs’ summary of points made by the identified speakers, and the statements have not been endorsed or verified by the National Academies of Sciences, Engineering, and Medicine. They are not intended to reflect a consensus among workshop participants.
LEARNING HEALTH CARE SYSTEMS
J. Michael McGinnis, Leonard D. Schaefer Executive Officer of the National Academy of Medicine (NAM), discussed the development of the LHS definition and its core principles, as well as current efforts to expand and scale this concept. He remarked on the tremendous opportunity to accelerate progress in applying the LHS approach to traumatic brain injuries (TBI). Given the $4 trillion spent annually on health care in the United States, the resources and the technology required to develop systems that learn and advance from every patient experience are available, he said, provided the will to dedicate resources to this endeavor is mustered.
Strategy for Creating Learning Health Care Systems
McGinnis explained that the strategy for building LHSs was born from two reports from the Institute of Medicine (IOM) (2000, 2001). These reports—which emphasized a focus on quality and safety—established the core principles that health care should be patient centered, safe, effective, equitable, efficient, and timely. As a lack of robust effectiveness data for many medical interventions became apparent, the issue of effectiveness emerged as a particular concern. Responding to a charge from the insurance and manufacturing sectors, IOM established the Roundtable on Evidence-Based Medicine in 2005. This group soon identified an absence of evidence for many common medical practices. Given the impracticality of conducting 5-year, $100 million randomized controlled trials (RCTs) to fill each data gap, the roundtable looked to new research methodologies and technologies to accelerate the process for developing a continuous LHS. A 2006 workshop and subsequent activities advanced an understanding of the LHS concept, with McGinnis described as
a health system in which science, informatics, incentives, and culture are aligned for continuous improvement, innovation, and equity—with best practices and discoveries seamlessly embedded in the delivery process, individuals and families as active participants in all elements, and new knowledge generated as an integral byproduct of the delivery experience.
Furthermore, he noted that acceleration toward such a movement would require the involvement of manufacturers, insurers, health professionals, digital infrastructure stewards, and patients and families.
Working toward that end, the roundtable formed a vision of developing the defined LHS, McGinnis said. The group created a series of action collaboratives based on the four elements featured in the LHS definition: evidence mobilization (science), digital health (informatics), incentives and
systems, and inclusion and equity (culture). These action collaboratives developed agendas identifying key pressure points to use in their respective arenas to achieve an LHS. The transformation targets across collaboratives include digital infrastructure, health and social data, effectiveness research, technical innovation, financial incentives, person and family engagement, community activation, and the decision culture. In exploring these various elements, the collaboratives produced over 30 publications in the Learning Health System series.2
Each action collaborative established goals within four strategic action domains—digital, evidence, economics, and sociocultural:
- Digital strategic action goals focus on developing a virtual health data trust in which data are interoperable, accessible, and protected.
- Evidence goals center on increasing real world learning capacity, including RCTs in some—but not all—cases. McGinnis emphasized the importance of ensuring that expensive, time-consuming RCTs remain relevant when they reach completion, not having become outdated during the years the process entails.
- Economic action goals target alignment of resources with health outcomes and a shift from fee-for-service as the major driver to payment for outcomes that matter most to individuals and to communities.
- Sociocultural goals focus on full and equitable health engagement.
Applying Core Principles to Practice
After much exploration of these complex issues, several action collaborative initiatives are poised to move forward, McGinnis said. The spread and scale of core principles is most imminent. As shown in Figure 3-1, the core principles of a quality health system originally identified in the IOM reports have been expanded. Initial patient centered terminology shifted to engaged to reflect active patient participation in a system fundamentally oriented to a partnership between patients, families, and the clinical enterprise. Timely has been replaced by accessible to encapsulate both timing and access in the ready availability of services. And four additional principles have been added: transparent, accountable, secure, and adaptive. A transparent system provides clear information related to the nature, use, costs, and results of services. Accountable refers to the reliable assessment of consequential activities and outcomes. McGinnis also emphasized that a system focused on outcomes does not attempt to measure every step of a
___________________
2 See the Learning Health System series at https://nam.edu/programs/value-science-driven-health-care/learning-health-system-series/ (accessed February 5, 2024).
process, but rather measures what matters most for the individual, the care delivery system, and the community. A secure system offers validated access and uses safeguards for digitally mediated activities. An adaptive system has an organizational culture with continuous learning and improvement at its core.
The roundtable is currently partnering with the American Hospital Association, the American Medical Association, state medical boards, and a broad range of participants to ensure that institutions and providers nationwide are stewards in creating learning health organizations, said McGinnis. This effort will also engage the National Institutes of Health, the Centers for Disease Control and Prevention, the Health Resources and Services Administration, and major organizations in the NAM Leadership Consortium. Given the fragmented nature of the U.S. health care system, the establishment of common commitments provides reference points for organizations working to improve the ways in which they engage with patients, he stated. The action collaboratives are building a series of networks focused on culture, informatics, incentives, and science to advance strategies to scale LHS principles and activities. As this effort progresses, these networks will also provide a mechanism for gathering information. He emphasized that the professionals working to develop and spread new LHS initiatives are pioneers and are valuable in helping to identify opportunities.
The collaboratives have developed starter application templates describing how the core principles apply within the arenas of culture, informatics,
incentives, and science, McGinnis explained. He provided examples of how each of the 10 core principles applies to the areas of digital health and evidence mobilization (see Box 3-1). McGinnis emphasized that organizations should adapt and expand these starter applications to their circumstances. In working toward the core principles across the arenas of evidence, informatics, incentives, and culture, organizations can establish LHSs. Furthermore, organizations can consider how these applications apply across sites including the clinic, home, businesses, and communities.
BOX 3-1
Applying Learning Health System Principles in Two Areas
Digital Health
- Engaged: Digital health records reflect engagement when discretion on control and use of personal data resides with the individual or their designee.
- Safe: Safety involves data stewardship protocols that safeguard against use resulting in harm.
- Effective: Digital health records collect and maintain data according to validated stewardship protocols.
- Equitable: Data systems are designed to identify and counter bias or disparities.
- Efficient: Data systems acquire only those service licenses that enhance health system interoperability.
- Accessible: Records feature data that are available at the times, locations, and on devices most proximate to decisions.
- Measurable: Systems continuously monitor digital health performance for accuracy and interoperability.
- Transparent: Transparency is achieved by making the sources and uses of personal data clearly evident.
- Secure: Systems establish data sharing protocols that are transparent and are considered secure by users.
- Adaptive: Adaptability is achieved when data strategies are regularly calibrated to ensure continuity, currency, utility, and security.
Evidence Mobilization
- Engaged: Mobilization occurs when individuals, circumstances, and personal goals shape health and health care.
- Safe: Safety is fostered by health services and research that contain safeguards against unintended harm.
- Effective: Health services both reflect and enhance the evidence base.
- Equitable: Equity involves developing and applying evidence with care and standards to eliminate bias.
- Efficient: Evidence mobilization develops and applies evidence using resource-optimization strategies.
- Accessible: Accessible evidence refers to the availability of best evidence at the point of choice to guide health services delivery.
- Measurable: A measurable system digitally records and assesses health services for continuous learning.
- Transparent: Evidence is open and accountable as to source strength and applicability.
- Secure: Security involves services and results that are tracked, reported, and stored with validated safeguards.
- Adaptive: An adaptive system features evidence, algorithms, and service protocols that reflect the evolving knowledge base.
SOURCE: Presented by Michael McGinnis, October 12, 2023.
McGinnis closed with the example of applying the learning health system approach to an assessment of the impact of the COVID-19 pandemic across the health system, which identified the need for a commission to explore the challenges that system fragmentation, perverse or misaligned incentives, and inequities posed to the pandemic responses (NAM, 2023). To address this need, NAM has established the National Commission on Investment Imperatives for a Healthy Nation to integrate lessons learned and advance alignment across sectors via five work streams: individual and community health goals, inclusive and equitable systems, digital and data architecture, funding and accountability, and private equity health investments.
DISCUSSION
Principle Prioritization
Kathy Lee, senior health policy analyst at the Office of the Deputy Assistant Secretary of Defense for Health Readiness Policy and Oversight, asked about core principles to prioritize during the initial stages of creating a learning health care system. McGinnis replied that the core principles can serve as a checklist for health organizations to use in reviewing their performance within each dimension. The results of such a review may lead different organizations to varied emphases. An initial priority for a TBI system should be engaging patients, families, and communities, he suggested, noting that a financial reimbursement system for medical care will not precede grassroots demand and political will for this care. Lee asked about the size and scope desired in a learning health care system. McGinnis clarified that the entire
system of factors that impinge on health status are involved, including public health, social services, organizations, and constituent groups.
Cross-Sector Collaboration
Christina Master, pediatrician and sports medicine specialist at Children’s Hospital of Philadelphia, noted the critical role of electronic health records and health information systems in caring for children with mild TBI. She asked about strategies for engaging partners from industry and academia, given that goals may differ and, in some cases, competitive interests may be at play. McGinnis emphasized the value of discussing common commitments with partners across sectors, specifically and transparently voicing the implications of moving forward. The development of generative artificial intelligence (AI) and large language models holds potential to transform the way business is conducted and how patients and families are engaged in the learning enterprise, he said.
Measuring and Replicating Excellence
Ramon Diaz-Arrastia, director of the Traumatic Brain Injury Clinical Research Center at the University of Pennsylvania, remarked that the fragmented U.S. health system presents challenges, but it also enables innovation and excellence to emerge organically. Given that a top-down system is unlikely to function well over the long term, he asked how best to identify centers of excellence and disseminate their models. McGinnis replied that while a top-down system is unlikely to take root, a more integrated system delivering evidence-based interventions is possible. Such a system should operate with a governance structure that generates transparency regarding the results of interventions and common approaches to health care delivery, he said, and an inherent challenge in this process is the need to narrow down measurement strategies to a relatively small number. For instance, IOM issued a report that examined 2,000 measures required by Medicare, Medicaid, and other insurance systems, then identified key clusters among those measures and described how AI and machine learning could be used to ease data collection and assessment (IOM, 2015). Regarding centers of excellence, McGinnis noted that roundtable action collaboratives have identified best practices related to informatics, evidence, culture, and incentives and will feature these in an upcoming annual report.
Health Outcomes and Spending
The session moderator Harris commented that health system incentives can be at odds with a patient-engagement model; she asked how to
address this pervasive barrier. McGinnis emphasized that an understanding of what patients and families need in order to meet their goals must drive the process, and incentives then must be structured accordingly. Incentives for outcomes will only be built into systems with full partnership and advocacy from the patient and family community. Currently, he said, the U.S. health care system spends $4 trillion annually, and a quarter of this expense is waste (CMS, 2023; Shrank et al., 2019). This waste is reflected in the country’s rank of 35th worldwide in performance (in terms of outcomes) despite outspending all other countries. Furthermore, the current incentive system has enabled intractable inequities. He emphasized that political will is needed to change the incentive system and, to this end, experts should partner with patients and families to identify the consequences of this system in a cohesive and compelling manner.