Fragility Analysis Should Complement Statistical Reporting in Thrombosis Trials: Lessons from Reduced-Dose Direct Oral Anticoagulants for Cancer-Associated Venous Thromboembolism.

Randomized trials and meta-analyses in thrombosis are traditionally interpreted using effect estimates, confidence intervals, and p-values. While indispensable, these metrics often provide limited insight into the robustness of trials reporting statistically significant findings. Fragility analysis offers a complementary approach by quantifying the minimum number of event-status changes that would reverse statistical significance. Despite increasing adoption across oncology, ophthalmology, orthopedics, and cardiovascular medicine, fragility analysis remains underutilized in thrombosis studies. Using a recent meta-analysis comparing reduced- versus full-dose direct oral anticoagulants (DOACs) for extended treatment of cancer-associated venous thromboembolism as an illustrative example, we demonstrate how fragility analysis provides additional context regarding the stability of trial findings. More importantly, we argue that fragility analysis should become a routine adjunct to conventional statistical reporting in thrombosis trials and systematic reviews to facilitate interpretation of evidence, improve transparency, and enhance shared decision making.