OBJECTIVES : Frequent exercise is associated with pain management, yet the characteristics associated with exercise among individuals with pain are not well understood. This study identified predictors of frequent exercise among older United States adults with pain.

METHODS : This study involved a retrospective, cross-sectional database design using 2017 Medical Expenditure Panel Survey data. Eligible participants were alive for the full year, aged ≥50 years, and had pain in the past four weeks. The outcome variable was moderate-vigorous intensity exercise for at least 30 minutes five times a week (yes, no). Hierarchical logistic regression models were used to identify significant predictors of frequent exercise. National estimates were obtained by adjusting for the complex survey design. An a priori alpha level of 0.05 was used.

RESULTS : This study included 5,051 subjects for a weighted population of 56,751,256 non-institutionalized adults alive aged ≥50 with pain in 2017. Of these, 23,758,914 (41.9%, 95% confidence interval [CI]=40.0%-43.7%) reported doing moderate-vigorous intensity exercise for at least 30 minutes five times a week. Predictors significantly associated with doing exercise included: male versus female gender (adjusted odds ratio [AOR]=1.507, 95% CI=1.318-1.724); Hispanic vs non-Hispanic ethnicity (AOR=0.780, 95% CI=0.622-0.979); employed versus unemployed (AOR=1.274, 95% CI=1.040-1.560); 0 versus 5 chronic conditions (AOR=1.547, 95% CI=1.094-2.268); 2 versus 5 chronic conditions (AOR= 1.547, 95% CI=1.226-1.952); any limitations versus no limitations (AOR=0.827, 95% CI=0.694-0.985); little/moderate versus quite a bit/extreme pain (AOR=1.358, 95% CI=1.137-1.621); and excellent/very good (AOR=2.408, 95% CI=1.875-3.093) or good (AOR=1.337, 95% CI=1.087-1.646) versus fair/poor physical health.

CONCLUSIONS : This study found approximately 23 million of the 56 million older United States adults with pain performed moderate-vigorous exercise for at least 30 minutes five days a week in 2017. Several characteristics of these individuals can be used to predict exercise, which may form part of an appropriate personalized pain management protocol for pain patients.