Medication optimization in multimorbidity is frequently described as a search for inappropriate prescribing, unnecessary medicines, or opportunities for deprescribing. These approaches are valuable but incompletely represent a more difficult clinical problem: changing one component of a regimen can improve one therapeutic objective while worsening another. The competing consequences may involve disease control, adverse effects, drug interactions, prescribing cascades, functional reserve, treatment workload, patient priorities, or benefits that occur on different time horizons. Frailty further alters the consequences of therapeutic perturbation because modest physiological changes may have greater functional significance and because prognosis can change whether delayed preventive benefit remains clinically relevant. Evidence from medication-review trials also shows that improvement in prescribing indicators does not consistently translate into improvement in patient-level outcomes. Regimen optimization should therefore make competing objectives explicit, distinguish medication-level signals from patient-level consequences, and preserve uncertainty instead of collapsing it into a universal score. Patient priorities determine which compromises are acceptable, while reversibility and decision horizon determine how uncertain choices can be tested. The clinically defensible target is not maximal treatment of every condition independently, but a regimen whose expected benefits, harms, workload, feasibility, and temporal consequences remain acceptable for the individual patient and are reassessed as circumstances change.
