Title : Nudges, algorithms and the heart: Behavioural economics and AI governance in cardiovascular care
Abstract:
Cardiovascular disease remains the leading cause of death worldwide, and much of its burden depends on routine decisions about diet, medication adherence, physical activity and screening. This presentation examines how behavioural economics and artificial intelligence can work together to support those decisions, and what governance is needed when they do. Behavioural economics offers a well-studied set of tools: default options, framing effects, social norms, commitment devices and well-timed prompts, often grouped under the term nudges. Machine learning models, including predictive systems trained on electronic health records, wearable sensor streams and large language model interfaces, can now tailor when and how these nudges reach a patient. Personalised nudging may improve adherence at scale, but it also raises questions of autonomy, algorithmic bias, explainability and accountability that clinical governance frameworks have not yet fully addressed. The presentation sets out the Governance Gap, the distance between what deployed clinical AI systems do and what oversight structures can verify. It then introduces the AI-G Index, a proposed scoring framework that assesses systems across data provenance, subgroup performance and fairness testing, human oversight, and post-deployment monitoring. Using illustrative cardiology use cases, including hypertension management and cardiac rehabilitation, the talk considers where a behavioural nudge becomes a form of algorithmic decision-making and how responsible deployment can be designed in from the outset. The session closes with practical recommendations for clinicians, health system leaders and AI developers who want to deploy behavioural interventions that are effective, transparent and equitable.

