Validation of self-reported adherence in chronic patients visiting pharmacies and factors associated with the overestimation and underestimation of good adherence.

BACKGROUND Studies validating indirect methods to identify nonadherence in chronic patients who visit pharmacies are lacking. The aim of this study was to validate self-reported adherence and assess the variables associated with both overestimation and underestimation of good adherence when using this method. MATERIALS AND METHODS An observational, cross-sectional study was undertaken to validate self-reported adherence in 132 community pharmacies throughout Spain in 6237 chronic patients. The Morisky-Green test was used as the validation method and through a 2 × 2 table, the validity indicators, predictive values, and likelihood ratios were calculated. To assess the variables associated with both overestimation and underestimation of good adherence, multivariate logistic regression analysis and calculation of the area under the ROC curve were used to evaluate discriminatory capacity. RESULTS Sensitivity was 27.8% (95% CI: 26.2-29.4) and specificity was 93.9% (95% CI: 93.1-94.7). Discrepancy analysis obtained a significant overestimation of good adherence (p < 0.001). The factors associated with overestimating good adherence were performing a mnemonic trick (p < 0.001), not self-medicating (p < 0.001), a high level of physical activity (p < 0.001), and an older age (p = 0.014). Factors associated with underestimation were self-medication (p < 0.001), desiring more information (p < 0.001), smoking (p = 0.014), not engaging in physical activity in the low (p = 0.006) or high (p < 0.001) categories, having a younger mean age (p = 0.007), and taking two to three (p = 0.029) or four or more (p < 0.001) chronic treatments. CONCLUSION Self-reported adherence has good specificity but poor sensitivity. The associated profiles of the discrepancies were obtained to identify both good and poor adherence.