How Does a Health System Learn?

How do health systems learn, and what does it mean to be a ‘learning health system’? Are we seeing a shift in the kinds of knowledge, practices, and professional cultures that characterise the NHS with the advent of continuous improvement enabled by biomedical informatics? What does data-driven innovation mean for patients?

On Thursday 16th January 2020, the Centre for Biomedicine, Self & Society hosted a half-day event at the Queen’s Medical Research Institute in Edinburgh to provoke discussion and debate of these questions. Generously funded by the Wellcome Institutional Strategic Support Fund, the meeting brought together an interdisciplinary group of participants, including clinicians, sociologists, clinical trialists, data scientists, health services researchers, and information systems managers, amongst others. The discussion was captured in a large-scale visual record by Clare Mills from Listen Think Draw.

Image

The impetus for the event was the growing traction of “learning health systems”, defined by the Institute of Medicine as those 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 by-product of the delivery experience”[1].  Since the term was first coined, it has grown in popularity across the globe as a way to embed and constantly update evidence-based medicine within health care systems. Described as an “ultra-large scale cyber-social system”[2], it brings with it a raft of attractive promises such as better care at lower cost, a democratization of health, seamless and continuous improvement, personalized medicine, and discovery as a natural outgrowth of every clinical encounter.  If achieved, it is nothing short of a paradigm shift in the organisation and delivery of healthcare.

Speaking at the event, Professor Aziz Sheikh challenged the room to consider the relationship between healthcare and health and to expand the focus from a hospital-centric idea of healthcare to a broader understanding of health systems, which could include, for example, environmental health. Using the example of an Ultra Low Emission Zone in central London to improve children's respiratory health, he suggested the need to think about the big, and often unevaluated, issues in regards to health policy.  Rather than creating many discrete, small-scale examples of learning health systems, can we create the first learning health system in the NHS at scale? Alluring as this may be, there are few worked examples of learning health systems, and the implications for society - organisational, ethical, regulatory, economic – are as yet unknown.

At the Centre for Biomedicine, Self & Society, we are exploring what it means to envision a ‘self-learning health system’ and the kinds of data practices and other work involved to make such a vision reality. During the meeting, we encouraged participants to move beyond the narrow definition of a ‘Learning Health System’ to consider more broadly the relationship between learning and care:

• How are different professional cultures and forms of expertise coalescing in healthcare to change the practice of medicine?
• How are machine learning and social learning combining in new ways to change the care that patients receive?
• How are new data practices reconfiguring the roles and relationships between doctors, patients and scientists?
• What does the learning health system mean in terms of access to and equality in health care?
• What are the ethical and regulatory implications of these changes in practice and the blurring boundaries between research and treatment?

The meeting provided a dynamic discussion to begin to break open these questions.  Key areas of concern centred on trust and trustworthiness, particularly in relation to data sharing, privacy and the use of algorithm-based predictive modelling in the clinic; the importance of public involvement and ensuring that users’ voices are heard during system-wide technological change; and information governance barriers to achieving learning health systems in practice.  It is clear that collaborative and interdisciplinary research between the social and biomedical sciences is needed to provide theoretically-informed and empirically-rich responses to these concerns. If you would like to know more or would be interested to work with us to understand how health systems learn, please contact Catherine.montgomery@ed.ac.uk.

“How Does a Health System Learn?” was organised by Professor Steve Sturdy, Dr Nayha Sethi, Dr Lukas Engelmann and Dr Catherine Montgomery of the cross-cutting CBSS theme, Beyond Data, and Stephanie Sinclair, CBSS Public Engagement and Knowledge Exchange Co-ordinator.