Counterfactual Medication Planning before Treatment Is Changed

Medication decisions are commonly informed by predicted risks, expected responses, previous treatment experience, and population-level evidence. Yet the practical question faced before treatment is changed is not merely whether an adverse event or therapeutic failure is likely. It is how the patient’s future might differ under continuation, dose modification, switching, augmentation, tapering, discontinuation, delayed action, or additional information collection. Current predictive, causal, medication-safety, and clinical decision-support approaches address parts of this problem but do not provide an integrated structure for comparing alternative medication trajectories while preserving uncertainty, professional responsibility, and patient priorities. This article proposes a Counterfactual Medication-Planning Framework for organizing treatment-change deliberation. The framework begins with an explicitly defined decision time and patient–medication state, identifies feasible treatment strategies, specifies outcomes and time horizons, and constructs parallel multidimensional trajectories. Each trajectory is qualified through an assumption and confounding ledger and an uncertainty envelope. An identifiability and comparability gate determines whether alternatives may proceed to human deliberation, require additional information or restriction, or should not be compared. The framework separates trajectory generation from medication authorization and places pharmacists, prescribers, patients, and accountable organizations within a bounded decision and monitoring process. Evaluation would require causal, technical, medication-safety, human-factors, equity, transportability, implementation, and lifecycle-governance evidence. Important boundaries include unmeasured confounding, incomplete longitudinal records, poor treatment overlap, uncertain adherence, unstable patient preferences, rare or delayed harms, and the impossibility of observing both treatment futures in the same patient. The contribution is an original non-empirical synthesis intended to make the assumptions, uncertainties, responsibilities, and abstention conditions of pre-change medication planning explicit. It requires empirical validation before clinical use.