Behavioral Valuation of Benefits and Harms in Cardiovascular Prevention: A Trial-Based Cumulative Prospect Theory Sensitivity Analysis

Document Type : Original Article

Authors

Abstract
Aims: To examine how probability weighting, reference points, relative outcome burdens, and assumptions about co-occurring events affect the valuation of cardiovascular prevention benefits and harms.

Methods: A secondary decision-modeling study used published aggregate counts from six randomized comparisons covering aspirin, rosuvastatin, blood-pressure lowering, and a polypill plus aspirin. Each comparison retained its own population, endpoints, and follow-up. Expected normalized endpoint burden was compared with cumulative prospect theory using complete outcome distributions. Outcomes were not assigned empirical disability-adjusted life-year weights. Severity thresholds, alternative reference points, and feasible dependence scenarios were examined.

Findings: In the aspirin comparisons, crude vascular-event reductions were 10.98, 2.40, and 1.87 per 1000 participants in ASCEND, ASPREE, and ARRIVE, respectively; corresponding bleeding increases were 8.91, 10.26, and 5.11. Bleeding definitions differed. Under independence and additive burdens, changing probability weighting from linear to a Prelec parameter of 0.7 shifted the harm-severity thresholds from 1.232 to 1.029, 0.234 to 0.222, and 0.366 to 0.249. The direction of threshold change was not universal across the other comparisons. Dependence assumptions could reverse selected behavioral valuations. With a healthy-state reference, a common loss-aversion coefficient changed scale but not sign.

Conclusion: The model identifies assumption-sensitive benefit–harm trade-offs within selected randomized comparisons. The illustrative scenarios do not estimate patient preferences, clinical net benefit, or public-health priorities.

Keywords

Subjects


Articles in Press, Accepted Manuscript
Available Online from 16 November 2026