Physiology’s role in unlocking the potential of AI in health

In 2023, the Physiological Society noted its concern that artificial intelligence (AI) healthcare tools are being developed, approved, and adopted without sufficient physiological input. Research Consulting, alongside independent consultant Georgina Humphreys, were commissioned to consult with experts in physiology to run an evidence-gathering research project on the emerging technology of artificial intelligence in the field of diagnostics. The key objectives of this work were to tease out both potential opportunities and risks, with a focus on maximising equitable health outcomes.

This work involved a consultation with over 30 experts, to (i) explore the relationship between physiology, AI and diagnostics; (ii) understand the opportunities and challenges for improved diagnostics with respect to inequalities in health outcomes; and (iii) to identify how Government can respond to the opportunities and risks identified. The consultation was delivered through a series of one-to-one interviews and interactive focus groups, generating qualitative data that informed a final report.

The final report was prepared as a concise policy document, including five actionable recommendations targeting Government departments and political party manifesto commitments. The report was formally launched at an event in the House of Lords, including discussion around our key finding: that the limited inclusion of physiological evidence in the development of healthcare AI tools can lead to reduced trust, challenges with applicability and to the identification of spurious correlations without sufficient physiological plausibility.