Algorithmic systems increasingly generate medication alerts, risk estimates, prioritizations, and treatment suggestions that can influence pharmacists, prescribers, patients, and organizations. Their safety cannot be secured solely through predictive performance, the presence of a professional reviewer, or the provision of a technical explanation. When advice is difficult to question, interrupt, reassess, or correct, it may acquire practical authority beyond its evidential strength. This article develops an original contestability framework for algorithmic medication advice in routine pharmacy care. Contestability is defined as the practical capacity of an affected or responsible person to obtain decision-relevant information, raise a challenge, activate a proportionate protective response, secure accountable review, and achieve correction or justified confirmation. The proposed framework connects six functions: registration of the advice object; access to reasons, evidence, alternatives, uncertainty, and limitations; human deliberation; challenge and escalation; correction or supersession; and organizational learning. Accountability, accessibility, equity, traceability, response time, and auditability operate as cross-cutting conditions. The framework distinguishes model output from medication decisions, explanation from justification, nominal oversight from meaningful review, and patient-level correction from system-level learning. Evaluation should assess whether contestability can be located, understood, initiated, answered, and closed under realistic workload and consequence conditions. Validation requires prospective pharmacy-workflow studies, patient and professional usability research, safety evaluation, equity analysis, external assessment, and longitudinal monitoring. Contestability is presented as a necessary governance capability rather than proof of algorithmic validity, clinical benefit, regulatory acceptability, or deployment readiness.