Pharmacogenomics is increasingly used to guide medication selection and dose individualization, yet routine implementation produces inconsistent clinical and patient outcomes. This realist review examines how pharmacogenomic programmes operate, for whom, and under which conditions they generate therapeutic benefit, limited uptake, or unintended consequences. The initial programme theory proposed that benefit becomes more likely when a clinically consequential drug–gene relationship is translated into a timely, understandable, and feasible recommendation; incorporated into prescribing workflows; supported by laboratory, informatics, professional, and reimbursement infrastructure; and accepted by clinicians and patients. Evidence was identified through iterative, purposive searching and appraised for explanatory relevance, richness, and rigour. Contexts, programme resources, actor responses, proposed mechanisms, outcomes, negative cases, and rival explanations were compared across therapeutic and implementation settings. The synthesis indicates that testing does not act independently. Pharmacogenomic resources must reduce uncertainty, create perceived actionability, strengthen prescriber confidence, support patient trust, and coordinate responsibility before they can alter treatment. Even then, prescribing change produces patient benefit only when it materially affects a credible pathway to toxicity prevention or improved response. Delayed results, ambiguous recommendations, alert fatigue, phenoconversion, uncertain ancestry transferability, weak reimbursement, and unequal access can interrupt this pathway. The refined programme theory therefore treats pharmacogenomic utility as an emergent property of evidence validity, therapeutic consequence, timing, interpretation, behavioural response, organizational capability, population relevance, and access. Pharmacogenomics may improve routine therapeutics when these conditions align, but implementation activity alone cannot establish clinical utility.
IJPRAS
