Levin Thomas, Patricia Ojalvo-Guiberteau, Juan-Antonio Villatoro-García, Carmen Mata-Martín, María Estévez-Paredes, Francisco Buitrago, Félix Suárez-González, Jesús Cobaleda, Eva M Peñas-Lledó, Adrián LLerena
One in four primary care patients was nonadherent to the prescribed medications. At the patient level, polypharmacy was the strongest independent determinant of lower medication adherence across both PGx Level 1A and non-PGx Level 1A strata. Female sex was independently associated with lower odds of medication adherence among patients prescribed PGx Level 1A medications. At the medication level, the CYP2D6 genotype-predicted poor metabolizer phenotype was independently associated with lower odds of medication adherence only for PGx Level 1A medications.
PURPOSE: Medication nonadherence can offset the benefits of pharmacogenomic (PGx)-guided pharmacotherapy. Integrated assessment of demographic, prescribing, and PGx determinants of medication adherence in real-world, PGx-integrated primary care populations remains insufficiently characterized.
METHODS: This cross-sectional study assessed medication adherence among 1066 adult patients within the "MedeA" PGx implementation initiative across three primary care centers of the Public Health Service of Extremadura (SES) in Spain. Prescriptions were identified from electronic health records and verified for medication-by-medication adherence through structured patient interviews. Multivariable logistic regression models were applied at the patient level, and generalized estimating equation models were applied at the medication level, stratified by exposure to PGx Level 1A versus non-PGx Level 1A medications, to identify independent demographic, prescribing, and PGx determinants of medication adherence.
FINDINGS: One in four primary care patients was nonadherent to the prescribed medications. At the patient level, polypharmacy was the strongest independent determinant of lower medication adherence across both PGx Level 1A and non-PGx Level 1A strata. Female sex was independently associated with lower odds of medication adherence among patients prescribed PGx Level 1A medications. At the medication level, the CYP2D6 genotype-predicted poor metabolizer phenotype was independently associated with lower odds of medication adherence only for PGx Level 1A medications.
IMPLICATIONS: Implementing strategies to optimize medication burden, alongside PGx-informed, medication-level adherence assessment, may help enhance the potential clinical impact of precision prescribing in primary care; however, all PGx-related findings from this study should be considered exploratory and hypothesis-generating, requiring further confirmation via prospective studies.