Title : Beyond data: Behavioural economics and AI in cardiovascular care
Abstract:
Cardiovascular disease accounts for approximately 18 million deaths annually, yet up to 80% of premature cardiac events are preventable, and medication adherence in chronic cardiovascular conditions remains at just 50–60%. This gap is not a knowledge deficit; it is a behavioural one. Drawing on behavioural economics, this presentation examines how present bias leads patients to discount future cardiac risk, how optimism bias undermines engagement with preventive protocols, and how clinicians operating under time pressure rely on heuristics that produce systematic distortions, manifesting in the overuse of stents in stable angina and the underuse of statins in high-risk populations. Artificial intelligence is increasingly deployed across cardiovascular care, from predictive modelling for heart failure readmissions to AI-assisted imaging showing detection accuracy improvements of 10–20% in selected studies; however, AI models trained on historical clinical data inherit the inequities and cognitive errors embedded in past decisions, and through self-reinforcing feedback loops, these patterns become automated, scaled, and institutionalised. The real opportunity lies not in deploying AI alone, but in combining it with behavioural insight: embedding nudges at point-of-care interfaces to counter anchoring and availability bias, deploying personalised adherence interventions using patient-specific timing and framing, and designing real-time risk communication tools that meaningfully reduce optimism bias — approaches shown to shift clinical prescribing behaviour by 5–15%, consistently outperforming complex educational interventions. AI will not fix decision-making in cardiology. It will expose it. The central question is not whether we are building smarter tools, but whether we are building tools that produce better decisions, or merely faster ones, at scale, and with the veneer of objectivity.
