Development and internal validation of GENADHECAR: a screening instrument for identifying probable statin non-adherence in patients with ischemic heart disease

Abstract

Background: No validated and practical tool exists to assess adherence to statins in secondary prevention of cardiovascular disease. Current reference methods for assessing adherence, including plasma drug measurements, are not feasible for routine clinical care or large clinical trials. Objective: To develop and internally validate a predictive screening instrument for identifying patients at risk of statin non-adherence. Methods: This multicenter cross-sectional study reports the results of the second phase of the GENADHECAR project in four Spanish hospitals. An initial set of 42 adherence-related items was refined and evaluated in a sample of patients with ischemic heart disease. Responses were compared against a pharmacokinetic reference method (plasma statin concentrations) to assess recent medication intake. A multivariate predictive model of non-adherence was then developed, selecting the best predictors and covariates to be included in the final screening instrument. Results: A total of 309 patients were included, of whom 49 (15.9%) were classified as non-adherent according to the pharmacokinetic reference method. The final GENADHECAR screening instrument comprised three questionnaire items (use of reminders, forgetfulness, and medication access) and two sociodemographic variables (sex and educational attainment). Internal bootstrap validation yielded an optimism-corrected AUC of 0.694 (95% CI 0.617–0.767), with adequate calibration (Hosmer–Lemeshow p = 0.994). At the selected probability threshold (0.15), sensitivity was 0.73, specificity 0.61, and positive and negative predictive values were 0.26 and 0.92, respectively. Overall, the model showed moderate discrimination and a high capacity to identify patients at low risk of non-adherence. Conclusion: To our knowledge, GENADHECAR is the first screening instrument with internal validation to identify statin non-adherence using a pharmacokinetic reference method. It offers a brief and practical approach for identifying patients at risk of non-adherence and may help overcome some limitations of conventional self-report methods in this population. Further external validation is needed before routine clinical implementation.