Heart disease is one of the leading causes of death in the U.S. Contributing to the toll exacted by heart disease is medical misdiagnosis. Efforts are underway to address this problem: recommendations from the Institute of Medicine, for example, led to a national system to hold facilities accountable and assist them in developing protocols to reduce future errors. On a local scale, cities like Chicago – in which heart disease is the primary cause of death – developed data compendiums to examine coronary diseases in the context of leading health problems. As recent national data on the prevalence of heart disease make clear, much nonetheless remains to be done. A good next step would be to adopt what I call nudge 2.0, an innovative and cost-effective measure to improve coronary disease diagnoses.

The roots of nudge 2.0 lie firmly in Chicago. About fifteen years ago, University of Chicago-affiliated academics, the economist Richard Thaler and the then-law professor Cass Sunstein, introduced the concept of a policy “nudge.” A nudge is a method of subtly improving human decision making, or reducing human fallibility. Misdiagnosis in heart disease (and elsewhere) often occurs from physician reliance on gut feelings or coarse rule-of-thumb guidelines in complex cases. Appropriate nudges would be well-suited to overcome common medical decision shortcomings in the context of heart disease.

Advances in artificial intelligence (AI) in recent years demonstrate that computers can quickly and thoroughly draw upon huge amounts of information and make appropriate recommendations based on that “knowledge.” In the world of medical diagnosis, a nudge 2.0 would involve AI generating a default or baseline diagnosis, individually personalized for patients at risk of heart disease. The sensible choice of defaults is a staple of nudge-based policy, and one that shares the usual nudge criterion of not precluding other courses of actions. Nudge 2.0 would “only” suggest a diagnosis, complementing rather than replacing the doctor’s judgment.

The AI-enhanced nudge 2.0 would extend the traditional default nudge through personalization. A personalized nudge would prove fruitful in the context of diagnosis as it would take into consideration complex and multidimensional individual medical information: a personalized nudge facilitates a more precise diagnosis, compared with applying uniform treatments to all patients exhibiting similar symptoms. Nudge 2.0 could thus prove to be a game-changer in diagnosis-making by communicating with cardiologists and offering guidance via an AI-enhanced, personalized recommendation.

Potential issues in introducing nudge 2.0 to the field of heart disease are data breaches and malicious usage of information. Storing, receiving, and transmitting the large amounts of sensitive medical data required for implementing an approach that combines AI and nudging could prove problematic. To ensure data privacy, it is necessary to implement strong security measures, ensure transparent disclosure of the algorithm, and establish responsible data collection and usage policies. The route forward involves implementing proven security measures, represented by encryption and multi-factor authentication, alongside conducting regular risk analyses. Health providers must additionally have clear policies in place regarding the usage and collection of patient data.

Nudge 2.0, which relies on AI for diagnosis suggestions, additionally raises concerns regarding the erosion of physician skills. A high-quality nudge 2.0 may obviate the need for doctors to master diagnostic skills, as they increasingly rely on technology for medical diagnoses. But perhaps the erosion will come too quickly, when the nudges themselves are not sufficiently accurate to merit high confidence. Any such concerns can be reduced with continued attention to ensuring that nudges complement doctors’ judgments, rather than replace them. In the long run, however, physician training should shift to ensure that they become especially proficient in areas where AI does not excel without their assistance – but this is a problem that has accompanied all medical technology developments over the centuries.

Nudge 2.0 presents high hopes for improving health outcomes. If applied to the problem of heart disease, nudge 2.0 could reduce medical errors in cardiology and contribute to addressing one of the main causes of death in the U.S. The Chicago School of Nudge, in conjunction with recent advancements in AI, has the potential to convert “subtle” nudges into transformative impacts on lives.