Habitual Physical Activity Is Associated With Prescription Medication Use In Older Adults

Abstract

Increased prescription medication (Rx) use has been associated with adverse drug events, drug-drug interactions, and rising health-care costs. Few studies have examined the potential relationship between physical activity (PA) level and Rx use. Thus, the purpose of this study was to determine if habitual PA is associated with Rx use in older adults. METHODS: In 171 older adults (116 women, 55 men, 72 ± 8y, BMI: 27 ± 5 kg/m2) physical activity (PA) was measured via 1) self-report (PA-SR) using the Community Healthy Activities Model Program for Seniors (CHAMPS) and 2) objectively (PA-O) using 7-day accelerometry (Actical, Phillips Respironics, Bend, Oregon). PA-SR data were calculated as kcal/wk of moderate-to-vigorous PA (MVPA). Accelerometry data were processed as min/day of MVPA and sedentary behavior, as well as the ratio of Sedentary/MVPA behavior (Sed/Mod). Prescription medication use (Rx) was collected during health history screening. Bivariate and partial correlations (controlling for age and BMI) were utilized to determine if measures of PA were significantly correlated with Rx. Hierarchical linear regression was utilized to determine the predictive power of age, BMI, and PA for estimating Rx use. RESULTS: Average Rx use was 2.5 ± 0.16 meds. Average physical activity was 1,632 ± 1,280 kcal/wk of moderate-to-vigorous PA for PA-SR and 55 ± 35 min/day of MVPA for PA-O. Age (r = 0.21), BMI (r = 0.47), PA-SR (r = -0.31), PA-O (r = -0.45), and Sed/Mod (r = 0.47) were significantly (p < 0.05) correlated with Rx use. Partial correlations (controlling for age and BMI) showed that PA-O and Sed/Mod, but not PA-SR, remained significantly correlated with Rx use. Utilizing hierarchical linear regression, a model containing BMI, Sed/Mod, and age was the most predictive of prescription medication use—explaining 42% of the variance. Of these independent variables, BMI and Sed/Mod ratio explained 21.7% and 21.6% of the variance in Rx use when included in a regression model individually, while age explained 3.8% of the variance in Rx use. CONCLUSION: These data extend previous reports that PA is inversely associated with Rx use. They further suggest that BMI and objective measures of PA are more predictive of Rx use than age in older adults.