Digital Pharmacy, Artificial Intelligence and Medication Safety: A Theoretical Framework for Clinical Pharmacy Practice, Dispensing Automation and Pharmaceutical Care in Resource-Constrained Health Systems

Abstract

Artificial intelligence (AI) is entering pharmacy practice through interaction and dose checking, prescription verification, automated dispensing, telepharmacy and patient-facing medication assistants. It arrives in health systems where pharmacists are scarce and medication errors are common. This theoretical paper develops a framework for the role of AI in clinical pharmacy and pharmaceutical care under resource constraint. It draws on pharmacy-practice and medication-safety theory, on the clinical-decision-support and human-factors literatures, and on recent research on AI in pharmacy, including the performance of large language models on pharmaceutical tasks. The analysis separates three functions that pharmacists perform: verification, the checking that a prescription is appropriate and safe; supply, the dispensing of the right product correctly; and care, the counselling, monitoring and optimisation of therapy for the individual patient. It argues that AI strengthens verification, automates supply, and cannot substitute for care, which is where the pharmacist's professional value increasingly lies, and that resource-constrained systems face a particular temptation to use AI to replace rather than extend scarce pharmacists. On this basis the paper advances a set of propositions about the conditions under which AI improves medication safety without degrading pharmaceutical care, and about the risk of using tools trained on drug data, formularies and prescribing patterns from other health systems, with Nigerian pharmacy practice anchoring the analysis. It concludes that, where pharmacists are scarce, the appropriate strategy is to automate the verification and supply functions so that scarce professional capacity is freed for the care that only pharmacists can provide, and that tools whose data derive from other health systems should be validated locally before they inform practice. A further emerging role is identified, in which pharmacists mediate between patients and the unreliable medication information they increasingly obtain from AI directly.