Deprescribing in frail adults is commonly framed as the removal of medicines whose harms or burdens exceed their likely benefits. That formulation is clinically useful but incomplete because discontinuation itself changes risk. Medication withdrawal can relieve adverse effects, administration burden, monitoring demands, drug interactions, and treatment workload, yet the same action may provoke physiological withdrawal, rebound phenomena, recurrence of the treated disorder, or loss of future preventive benefit. These consequences operate on different time scales. Withdrawal may emerge within days, recurrence may become apparent over weeks or months, preventive benefits may require months or years to accrue, and competing illness or mortality may prevent some delayed benefits from ever being realized. Frailty further alters the decision by changing susceptibility to drug harm, functional reserve, treatment priorities, monitoring capacity, and the consequences of unsuccessful withdrawal. This Current Opinion therefore treats deprescribing as a time-dependent competing-risk intervention rather than a unidirectional reduction in medication exposure. The central clinical task is to distinguish the type of risk created by stopping, determine when that risk is likely to become observable, compare it with the timing of continued-treatment benefit and harm, and preserve reversibility where uncertainty is substantial. Evidence from deprescribing trials remains heterogeneous: medication counts and potentially inappropriate prescribing can often be reduced, but improvements in mortality, function, quality of life, hospitalization, or other patient-important outcomes are less consistent. A clinically defensible deprescribing decision consequently requires an explicit decision horizon, class-specific understanding of discontinuation effects, and post-deprescribing surveillance capable of distinguishing transient withdrawal from recurrent disease or loss of necessary treatment.
