OBJECTIVES : To estimate the prevalence of opioid use and determine the characteristics associated with opioid use among older US adults (≥50 years) with comorbid hypertension and pain.

METHODS : This study involved a retrospective, cross-sectional database design using 2017 Medical Expenditure Panel Survey (MEPS) data. Eligible participants were alive for the calendar year, aged ≥50 years, had a diagnosis of hypertension, and reported having pain in the past four weeks. The outcome variable was opioid users versus non-users. Hierarchical logistic regression models were used to identify significant predictors of opioid use. National estimates were obtained by adjusting for the complex MEPS design. An alpha level of 0.05 was set a priori.

RESULTS : This study included 2,733 subjects; 803 opioid users and 1,930 non-opioid users. The total weighted number of non-institutionalized adults alive aged ≥50 with hypertension and pain in the US in 2017 was 26,997,191; of which 7,930,252 (29.4%, 95% confidence interval [CI]=27.3%, 31.4%) were opioid users. Predictors significantly associated with opioid use included: age 50-64 versus ≥65 years (adjusted odds ratio [AOR]=1.310, 95% CI=1.025-1.674); white race versus others (AOR=1.698, 95% CI=1.288-2.237); Hispanic versus non-Hispanic ethnicity (AOR=0.448, 95% CI=0.321-0.626); 1 versus ≥5 chronic conditions (AOR=0.515, 95% CI=0.312-0.849); functional limitation versus no functional limitation (AOR=1.457, 95% CI=1.121-1.893); little/moderate versus quite a bit/extreme pain (AOR=0.438, 95% CI=0.344-0.559); good versus fair/poor perceived mental health status (AOR=1.512, 95% CI=1.071-2.133); smoker versus non-smoker (AOR=1.435, 95% CI=1.076-1.914); Northeast versus West census region (AOR=0.509, 95% CI=0.361-0.718); and Midwest versus West census region (AOR=0.661, 95% CI=0.479-0.913).

CONCLUSIONS : This study found that approximately 8 million of the 27 million older US adults (≥50 years) with hypertension and pain used opioids in 2017, and that several characteristics of those individuals can be used to predict opioid use. These characteristics may be targets for interventions to help address the opioid epidemic.