Diversity of scale and context of implemented learning healthcare systems: A narrative literature review

Article type
Zurynski Y1, Smith K1, Vedovi A1, Ellis L2, Gul H2, Braithwaite J1
1Partnership Centre for Health System Sustainability, Australian Institute of Health Innovation, Macquarie University
2Centre for Healthcare Resilience and Implementation Science, Australian Institute of Health Innovation, Macquarie University
Background: The concept of a learning healthcare system (LHS) has been gaining traction for over a decade. In 2013, the US National Academy of Medicine identified four main areas necessary to establish an LHS: Science and Informatics, Patient-Client Relationship, Incentives, and Culture. As healthcare institutions increasingly look to develop learning cultures, improve clinical processes, and exploit the untapped potential of electronic health records in the age of artificial intelligence, a growing literature has emerged on the theory and—crucially—the emerging practical implementation of LHSs.

Objectives: We aimed to explore a sample of healthcare systems that self-identified as emerging or implemented LHSs to describe the current field.

Methods: We carried out a narrative/scoping review (January 2016 to May 2019), using search terms (“learning health system” and “learning health care system”) in PubMed and Scopus. We identified 276 publications; 79 (28.6%) were excluded as they mentioned LHSs only in passing (e.g., among the keywords, abstract, and/or conclusion). There were 197 remaining publications with an explicit discussion of LHSs, and 67 of these presented case studies or profiles of self-identifying emergent or established LHSs.

Results: The 67 articles described 51 LHS systems: 33 in the United States, 12 in other countries, and six multi-country LHS networks. Twenty-two described data networks or IT architecture platforms at local, national, or international levels to support collection, storage, sharing, and/or analysis of health data; 11 were practice-based research networks (PBRNs) or other ‘learning communities’, typically focused on particular conditions or disciplines. Five were combinations of a PBRN/learning community and a data network/platform. There were five LHSs operating at the level of single medical centre or clinical unit. We identified eight large-scale systems: four were private corporations and four were public systems at national community or specific population levels.

Conclusions: The LHS concept is a journey rather than a destination, and it has been implemented to some extent in several different contexts and at varying scales. Almost half of the systems centred on data usage and interoperability to provide fundamental infrastructure to underpin LHSs. PBRNs and PBRNs supported by data networks are also emerging; however, large-scale, systemwide LHSs are scarce. The use of the term “LHS” was applied to different systems regardless of their scope or scale. This is not surprising in this emerging and evolving field. Our review included LHSs described in the peer-reviewed literature; the grey literature may reveal additional examples.

Patient or healthcare consumer involvement: As this is a scoping review of the literature, consumers were not directly involved. The Australian Institute of Health Innovation and the Partnership Centre for Health System Sustainability collaborate with the Consumers Health Forum of Australia.