Withdrawal-Based Truth in Medicine: A Structural Criterion for Cure Falsifiability

Medicine evaluates treatments with increasing statistical and technical sophistication, yet it lacks a binary falsification rule for one of its most consequential assertions: the claim of cure. Improvement under treatment, durability over time, and long-term follow-up do not establish whether a medical state persists independently of continued intervention. As a result, cure claims often remain epistemically underdetermined. This paper introduces a withdrawal-based truth criterion for cure claims. A cure claim is truth-apt only if the claimed state remains stable after complete withdrawal of the intervention responsible for producing it. If withdrawal results in loss, degradation, or reversal of the state, the claim is false by structure. If withdrawal cannot be performed or meaningfully assessed, the claim is undecidable rather than validated. The framework is strictly epistemic and law-level. It provides a necessary falsification condition, not a sufficient definition of cure, and advances no treatment recommendations, clinical mandates, or policy proposals. Evidence obtained under continued intervention—statistical improvement, biomarker normalization, or long-term stability—remains essential for evaluating treatment effectiveness but is inadmissible for establishing cure truth. By treating causal independence, time, and irreversibility as structural invariants, the withdrawal criterion exposes failure modes that persist under maintenance, including masked compensation, survivorship artifacts, and treatment-dependent remission misclassified as cure. The framework clarifies the boundary between care and truth, preserves ethical constraints, and restores falsifiability to cure claims without diminishing therapeutic value. The central result is a minimal and non-negotiable principle:A cure without withdrawal is undefined.