Objective: To identify the strongest predictors of opioid use among older United States' (US) adults (≥50 years) with pain.
Design: Cross-sectional retrospective database design.
Setting: 2017 Medical Expenditure Panel Survey data.
Participants: Civilian, noninstitutionalized sample of US adults aged ≥50 years alive for the calendar year with pain in the past 4 weeks.
Interventions: Hierarchical logistic regression models assessed significant predictors of opioid use, which included: predisposing, enabling, need, personal health practices, and external environmental factors.
Main outcome measures: Opioid use status (opioid user vs. nonopioid user).
Results: Among 51,372,861 civilian, noninstitutionalized US adults alive aged ≥50 years with pain in 2017, the opioid use prevalence was 27.4 percent (95 percent confidence interval = 25.8-29.0). Predictors of opioid use included: white versus other race (adjusted odds ratio, AOR = 1.430), Hispanic versus non-Hispanic ethnicity (AOR = 0.648), up to high school versus higher than high school education (AOR = 1.259), functional limitation versus no limitation (AOR = 1.580), lit-tle/moderate versus quite a bit/extreme pain (AOR = 0.422), good versus fair/poor perceived mental health status (AOR = 1.429), smokers versus nonsmokers (AOR = 1.523), and residing in the northeast versus west US (AOR = 0.646).
Conclusions: This study of 51 million older US adults with pain indicated that several factors including race, ethnicity, education, functional limitations, pain severity, mental health status, smoking status, and region of the country were pre-dictors of opioid use. Future research is needed in additional clinical populations and to investigate where these findings diverge from previous studies.