Diabetes Prevalence by Leisure-, Transportation-, and Occupation-Based Physical Activity Among Racially/Ethnically Diverse U.S. Adults

Abstract

Objective: Leisure-time physical activity (LTPA) has been shown to prevent or delay the development of diabetes. However, little research exists examining how other domains of PA (e.g., occupation based [OPA] and transportation based [TPA]) are associated with diabetes prevalence across diverse racial/ethnic groups. We examined associations between OPA, TPA, and LTPA and diabetes prevalence and whether associations differed by race/ethnicity.

Research design and methods: Participants in the 2011-2016 National Health and Nutrition Examination Survey (NHANES) self-reported domain-specific PA. Diabetes status was determined by self-reported doctor/health professional-diagnosis of diabetes or a glycosylated hemoglobin (HbA_1c) measurement of ≥6.5% (48 mmol/mol). Multivariable log binomial models examined differences in diabetes prevalence by PA level in each domain and total PA among Latinos (n = 3,931), non-Latino whites (n = 6,079), and non-Latino blacks (n = 3,659).

Results: Whites reported the highest prevalence of achieving PA guidelines (64.9%), followed by Latinos (61.6%) and non-Latino blacks (60.9%; P < 0.0009). Participants achieving PA guidelines were 19-32% less likely to have diabetes depending on PA domain in adjusted models. Diabetes prevalence was consistently higher among non-Latino blacks (17.1%) and Latinos (14.1%) compared with non-Latino whites (10.7%; P < 0.0001), but interaction results showed the protective effect of PA was similar across PA domain and race/ethnicity-except within TPA, where the protective effect was 4% greater among non-Latino whites compared with Latinos (adjusted difference in risk differences 0.04, P = 0.01).

Conclusions: PA policies and programs, beyond LTPA, can be leveraged to reduce diabetes prevalence among all population groups. Future studies are needed to confirm potentially differential effects of transportation-based active living on diabetes prevalence across race/ethnicity.

Figure 1

Prevalence of diabetes by domain, level of PA, and race/ethnicity, adjusted for age, sex, marital status, income, education, and health insurance. Asterisks (*) indicate the P value for the contrasted predicted marginal comparing diabetes prevalence among those achieving and not achieving guidelines within each domain. *P < 0.05; **P < 0.01; ***P < 0.001.