What is a Learning Healthcare System (LHS), and how does it apply to hospital medicine? An LHS is defined by the Institute of Medicine as a “system in which science, informatics, incentives, and culture are aligned for continuous improvement and innovation, with best practices seamlessly embedded in the delivery process and new knowledge captured as an integral byproduct of the delivery experience” (1).
LHSs comprise several elements. The first is an infrastructure for capturing data from every patient encounter, most commonly from electronic medical records (EMRs). A second element is the use of sophisticated analytic tools to process clinical data, identify important signals, and iteratively deliver better evidence at the point of care. Examples include systems in which clinical guidelines are flagged when individual patient characteristics are recognized and that contain EMR-embedded decision support tools. Incentives and culture are interrelated concepts that recognize that an LHS does not just collect and deliver data: LHS clinical care occurs in a particular context and culture that incentivizes and rewards progressive improvement (1). LHSs are intentionally designed to have all of these elements, so data collection, analysis, implementation, evaluation, and learning occur systematically and in a continuous cycle.
Some LHS elements may seem familiar to hospitalists who use clinical data for quality improvement, conduct operationally relevant research, or who instruct learners in core competencies, such as practice-based learning. However, LHSs differ from these pursuits in several ways. First, the LHS data infrastructure, analytics, culture, and policy environment are embedded within an organization, health system, program, or department. Second, the emphasis is on large, population level data sets, high-performance computing, and advanced mathematical modeling, so knowledge can be automatically pushed to clinicians in real-time. Third, LHSs essentially never stop the plan-do-study-act cycle: They are designed for continuous learning so each patient encounter is better than the one before (2).
Researchers and clinician leaders working to reduce mortality in sepsis have leveraged the LHS framework to identify new opportunities to improve early, appropriate, and targeted treatment (3). For example, in the Kaiser Permanente Northern California integrated health care delivery system, early work sought to develop a robust infrastructure to support the implementation and “hard-wiring” of early sepsis care through education, workflow development, reporting scorecards, and EMR tools. After an initial phase focused on care for septic shock patients, the tools were used to improve outcomes among sepsis patients who had not yet developed shock. This work involved concurrent deployment of EMR risk scores, including an early warning score system for general inpatients developed using machine-learning algorithms, to identify high-risk inpatients with and without sepsis. A regional team reviews the alerts from all 21 hospitals to minimize workflow disruptions to front-line clinicians and mitigate alert fatigue, problems that are already substantial in the inpatient setting. In an effort to continue to identify novel opportunities to improve sepsis care, real-world data are being used to evaluate sepsis and presepsis risk prediction, to provide targeted fluid management and optimal triage of intermediate-risk patients, and to mitigate adverse long-term outcomes following sepsis. Over time, this program has improved overall outcomes across 21 Kaiser Permanente hospitals while identifying new opportunities to combat sepsis.
Many health care systems are beginning their LHS journey, providing an opportunity for hospitalists, nurse scientists, informaticists, and health systems researchers to shape the future of this innovative model. However, hospitalists bring a unique perspective to this work as they provide high-stakes care for some of the highest-risk and highest-cost patients in the health care system in a “data dense” environment. Further, hospitalists are “systematists,” for whom effective practice means integrating information from several sources, leading teams to promote high-quality care at the bedside, and promoting a culture of continuous improvement amid tremendous complexity. In many ways, hospitalists are the ideal human partners to guide LHS development through collection, interpretation, and implementation of data, and leading cultural change along the way. This is not to say that hospitalists have all of the skills required to function at the center of LHSs. Given the relative youth of the field, many hospitalists are early in their careers. We encourage hospitalists to look to implementation science, behavioral economics, informatics, and organizational psychology for additional skills to support the creation of LHSs.
Through massive data infrastructure investments, development of machine learning technology that can be integrated into EMRs, and attention to incentives and culture, LHSs are moving health care to new ground. Organizations across the nation are developing LHS infrastructures, doing the work, and publishing case studies that demonstrate the potential of the LHS vision (4). With their deep understanding of hospital systems, hospitalists are expected to play an important role in advancing LHSs in U.S. health care.
Acknowledgments
Financial Support: Drs. Gilmartin and Burke are supported by VHA Career Development Awards from the U.S. Department of Veterans Affairs.
Disclosures: Dr. Liu reports grants from the National Institutes of Health during the conduct of the study. Authors not named here have disclosed no conflicts of interest. Forms can be viewed at www.acponline.org/authors/icmje/ConflictOfInterestForms.do?msNum=M19-3873.
Footnotes
Publisher's Disclaimer: Disclaimer: The contents of this manuscript do not represent the views of the Department of Veterans Affairs or the U.S. government.
References
- 1.Mullins CD, Wingate LT, Edwards HA, et al. Transitioning from learning healthcare systems to learning health care communities. J Comp Eff Res. 2018;7:603–14. doi: 10.2217/cer-2017-0105 [DOI] [PubMed] [Google Scholar]
- 2.Greene SM, Reid RJ, Larson EB. Implementing the learning health system: from concept to action. Ann Intern Med. 2012;157:207–10. doi: 10.7326/0003-4819-157-3-201208070-00012 [DOI] [PubMed] [Google Scholar]
- 3.Liu VX, Morehouse JW, Baker JM, et al. Data that drive: Closing the loop in the learning hospital system. J Hosp Med. 2016;11 Suppl 1:S11–7. doi: 10.1002/jhm.2651 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Olsen L, Aisner D, McGinnis JM. Institute of Medicine (US). Round-table on Evidence-Based Medicine The learning healthcare system: workshop summary. Washington: National Academies Pr; 2007. [PubMed] [Google Scholar]