Trends in Prevalence and Determinants of Potentially Inappropriate Prescribing in the United States: 2007 to 2012.

OBJECTIVES To estimate the prevalence and determinants of the use of potentially inappropriate medications (PIMs) in older U.S. adults using the 2012 Beers criteria. DESIGN Retrospective cohort study in a random national sample of Medicare beneficiaries. SETTING Fee-for-service Medicare beneficiaries from 2007 to 2012. PARTICIPANTS U.S. population aged 65 and older with Parts A, B, and D enrollment in at least 1 month during a calendar year (N = 38,250 individuals; 1,308,116 observations). MEASUREMENTS The 2012 Beers criteria were used to estimate the prevalence of the use of PIMs in each calendar month and over a 12-month period using data on diagnoses or conditions present in the previous 12 months. Generalized estimating equations were used to account for the dependence of multiple monthly observations of a single person when estimating 95% confidence intervals (CIs), and logistic regression was used to identify independent determinants of PIM use. RESULTS The point prevalence of the use of PIMs decreased from 37.6% (95% CI = 37.0-38.1) in 2007 to 34.2% (95% CI = 33.6-34.7) in 2012, with a statistically significant 2% (95% CI = 1-3%) decline per year assuming a linear trend. The 1-year period prevalence declined from 64.9% in 2007 to 56.6% in 2012. The strongest predictor of PIM use was the number of drugs dispensed. Individuals aged 70 and older and those seen by a geriatrician were less likely to receive a PIM. CONCLUSION From 2007 to 2012, the prevalence of PIM use in older U.S. adults decreased according to the 2012 Beers criteria, although it remains high, still affecting one-third each month and more than half over 12 months. The number of dispensed prescriptions could be used to target future interventions.