Fractured by Design: How Post-Hoc Subgroup Analyses Are Distorting Individualized Treatment Decisions
Clinicians across the United States are increasingly anchoring individualized treatment decisions to subgroup findings that were never prespecified before trial data collection began. The statistical fragility of these post-hoc analyses—compounded by inadequate guideline warnings—creates a quiet but consequential distortion in evidence-based practice. This article examines the mechanisms behind the problem and the structural reforms needed to address it.