Pharmacogenomic evidence is often summarized by asking whether a gene–drug pair is actionable. That formulation is useful for cataloguing knowledge but too coarse for many prescribing decisions. The same gene can support different recommendations across drugs, a pharmacokinetic association can remain clinically nonactionable, a genotype-predicted phenotype can be altered by assay limitations or phenoconversion, and the value of genetic information can change according to whether the immediate task is drug selection, initial dosing, titration, monitoring, or avoidance. This Perspective argues that actionability should therefore be indexed to three coordinates: the drug actually being considered, the clinically current phenotype that can be defended from the available genotype and contextual information, and the decision point at which information could alter management. Evidence from antidepressants, beta-blockers, methadone, antiplatelet therapy, statins, aminoglycosides, thiopurines, and fluoropyrimidines illustrates why biological relevance, prescribing relevance, and outcome relevance should not be collapsed. Genotype-to-phenotype translation and phenoconversion further show that phenotype is an inferential and sometimes dynamic intermediate rather than a fixed synonym for genotype. The article consequently develops a proposed Drug–Phenotype–Decision view of pharmacogenomic actionability. Its purpose is not to replace existing prescribing guidelines, but to make explicit the boundaries that determine when a genomic result can reasonably change a specific pharmaceutical decision.
IJPRAS
