During the final session of the workshop series, Eric Larson, the planning committee chair, and Michael McGinnis from the IOM reviewed several points that were mentioned repeatedly over the course of the series.
The business imperative for continuous improvement
As a sector, health care has an unprecedented opportunity to use its current and developing data and methodological resources to drive continuous improvement, McGinnis noted. A key target of this opportunity is to move toward the triple aim through the delivery of high-value care in a manner informed by the best, most timely evidence—evidence that can be useful to decision makers, clinicians, and patients. This includes, he highlighted, learning from other sectors and working together to drive innovation in the area of learning from routinely collected data.
Culture that produces value at speed
Larson acknowledged that health care systems face tremendous financial pressures today, placing a premium on research that can be deployed rapidly to increase system efficiency, improve patient outcomes and satisfaction, and reduce costs. The nation’s health care system needs performance change and time horizons that are more rapid than those created by the standard research structure, he said. Demonstrating that integrated knowledge-generation efforts such as PCORnet can enable fast, focused studies will be a key to winning support from CEOs for creating a sustainable learning system.
Failing fast and forward
Health system leaders are most interested in studies that not only will generate usable information quickly but have the potential to provide valuable knowledge even when they fail, Larson reiterated from earlier workshop discussions and presentations. Although maintaining a willingness to fail can be difficult in health care, it is an essential attribute of a learning system. Key components of such a system, he summarized, are transparency to stakeholders, failing fast, and learning from experience and sharing those lessons.
The economic case for infrastructure for continuous learning
Accountability and continuous quality improvement are at the foundation of a responsible business model, McGinnis stated, noting that most health care organizations have recently made substantial investments in informatics and analytics capabilities. This expensive basic infrastructure could create important added value if the informatics and analytics capabilities are applied to knowledge-generation efforts that have an impact on moving toward higher-value care and that contribute to the improved efficiency of integrated research. The inclusion of cost information, such as activity-based costing, to this infrastructure was suggested by several workshop participants.
Alignment of research with organizational goals and priorities
Larson highlighted a point made by several workshop speakers that every organization has a finite bandwidth, not only in terms of financial resources but also with regard to institutional energy. Therefore, he noted, better alignment of research initiatives with performance improvement initiatives and consideration of implementation potential and provider burden in research design can maximize the impact of that research.
Sustainability in patient partnerships
Partnerships with patients at all stages of priority setting, knowledge generation, and implementation can be drivers of sustainability of continuous learning and improvement in health care. Larson highlighted that, throughout the workshop, speakers mentioned that patients want practical research that addresses their concerns and questions, they are eager to participate when these conditions are met, and they have taken an increasingly active role in driving research priorities themselves. He noted that new platforms and methodologies provide opportunities to better capture the patient experience and perspective both in routine care and from outside the health care system and to use this to drive learning.
Implementation at scale
Larson noted that several workshop speakers suggested that research without implementation does little to advance the development of a continuously learning health care system, and implementation that does not go beyond individual hospitals or health plans falls short of the transformational potential needed to effect major change. In that regard, he suggested that there is a need to advance the science of implementation and to make better use of the knowledge that already exists within the confines of individual health systems.
Good governance to enable big impacts
Effective, skillful governance promotes sustainability of research, summarized McGinnis, both in terms of the ability to develop shared research assets to conduct studies and in terms of developing a community of researchers and stakeholders who reuse and develop those assets. Governance structures, he noted, need to be flexible enough to adapt to the legitimate needs of those projects while also being inclusive enough to gather knowledge from each of the hospitals, health care systems, and networks of systems that are part of PCORnet and collate them into some organized whole in order to develop common models of analyzing data. One key governance focus, he highlighted, has to be on shortening the cycle time from when research generates evidence to when health system management triggers organizational change.
The ethical imperative for improvement
That ethics-relevant policies must be in place in a learning health care system and that those policies need to be transparent about ongoing learning and engage patients in a manner that informs decisions about which studies need consent and further protections, Larson noted as a theme mentioned by workshop speakers. As these learning activities proceed, ongoing evaluation of what types of learning activities need review and consent should become a standard part of the health care system, he suggested. IRBs play an important role in providing oversight of clinical research, but too often IRBs do not have the expertise to understand how comparative effectiveness research differs from investigator-initiated, hypothesis-driven clinical trials.
Partnerships to drive learning
A hallmark of successful continuously learning health systems is the partnership that develops among clinicians, patients, and health system leaders, noted McGinnis. Strong partnerships, particularly those that in-
clude patients, he highlighted, will be just as important for realizing the full potential of PCORnet as the network’s ability to generate the data needed to inform a learning health system.
A continuous improvement workforce
McGinnis suggested that realization of a continuously learning health system will depend on a workforce that is trained and experienced in the skills needed to integrate research and the practice of health care delivery. This, he noted, has implications for the education and training of future clinicians, researchers, and health care administrators as well as in the design of career incentives for young professionals.