Causal Inference in Studies with Functional Unmasking: Psychedelics and Beyond
medRxiv Preprint Server December 5, 2025 Gabriel Loewinger, Mats J. Stensrud, Sandeep M. Nayak et al. preprint
Functional unmasking (unblinding) in clinical trials for mental health treatments, especially with psychedelics, can bias results because participants often know they received the active drug due to its unmistakable acute effects. This undermines confidence that outcomes reflect true therapeutic properties rather than placebo-like effects. A counterfactual conceptualization formalizes the shortcomings of existing solutions like dose-response and active controls, and shows how modern causal inference approaches can isolate effects free of this contamination. Feedback mechanisms between perceived benefits and expectancies can make traditional methods obscure or exaggerate therapeutic benefits. The proposal motivates trial designs and statistical methods to mitigate the impacts of functional unmasking.