Factors associated with potentially inappropriate medications use by the elderly according to Beers criteria 2003 and 2012

Background Potentially inappropriate medications (PIMs) should be avoided by the elderly because they possess a significant high risk for this population when a safer alternative is available. Therefore, the identification of prevalence and factors associated with PIMs should be considered as they provide valuable information that can be used to develop strategies to ensure patients’ safety. Objective To identify the prevalence and the clinical and socioeconomic–demographic factors that may be associated with PIMs use in the elderly, according to Beers criteria 2003 and its updated version 2012. And, as a secondary objective, a comparison between both criteria was performed. Setting Pharmacy of the Basic Health District Unit of the western district of Ribeirão Preto. Methods This cross-sectional observational study was conducted with the elderly, assisted by the Brazilian public health system. Data from patients were collected through a structured interview form. Beers criteria 2003 and 2012 were used to classify PIMs. The association between PIMs used and independent variables were analyzed by odds ratios. The differences between PIMs use according to Beers criteria 2003 and 2012 were analyzed by McNemar’s test and the agreement by kappa coefficient. Main outcome measure Prevalence and factors associated with PIMs use in Brazilian elderly outpatients. Results One thousand elderly patients were interviewed. High prevalence of PIMs use was observed, 48.0 and 59.2 % according to Beers criteria 2003 and 2012, respectively. The factors associated with PIMs use, common for both criteria, are female gender, self-medication, use of over the counter drugs, complaints related to adverse drug event, psychotropic medication, polypharmacy and some categories of drugs. PIMs use is different between Beers criteria 2003 and 2012 (McNemar’s test, p < 0.01), although a substantial agreement between these classifications was observed (kappa coefficient 0.635, 95% confidence intervals (0.588, 0.681). Conclusions Our study showed a high prevalence of PIMs use, which is associated with various clinical and social–demographic factors. When comparing both criteria through McNemar’s test, PIMs use was considered different. The differences may have occurred because medications with high prevalence of use in Brazil were included in Beers criteria 2012 .