Racial, Ethnic, and Nativity Disparities in Physical Activity and Sedentary Time among Cancer Prevention Study-3 Participants

Erika Rees-Punia; Elizabeth G. Kirkland; Melissa H. Rittase; Carlos X. Torres; Sicha Chantaprasopsuk; Matthew Masters; Alpa V. Patel


Med Sci Sports Exerc. 2022;54(7):1139-1146. 

In This Article

Abstract and Introduction


Purpose: Understanding racial/ethnic and nativity disparities in physical activity (PA) is important, as certain subgroups bear a disproportionate burden of physical inactivity–related diseases. This descriptive study compared mean leisure-time moderate-to-vigorous intensity physical activity (LTMVPA) by race/ethnicity and nativity.

Methods: The Cancer Prevention Study-3 (78.1% women; age, 47.9 ± 9.7 yr) includes 4722 (1.9%) Asian/Pacific Islander; 1232 (0.5%) Black/Indigenous (non-White) Latino; 16,041 (6.5%) White Latino; 9295 (3.8%) non-Latino Black; 2623 (1.1%) Indigenous American; and 210,504 (85.7%) non-Latino White participants across the United States and Puerto Rico. Participants completed validated LTMVPA and 24-h time use surveys at enrollment (2006–2013). Differences in LTMVPA across race/ethnicity and nativity were examined by ANCOVA with paired Tukey tests adjusting for age and sex. The proportion of time spent sitting, sleeping, and on PA by race/ethnicity was also compared.

Results: There were significant differences in LTMVPA by race/ethnicity (race main effect, P < 0.001; nativity, P = 0.072; interaction, P < 0.001). Pairwise comparisons showed that White participants born abroad were the most active (23.8 MET-h·wk−1) and non-White Latino participants born abroad were the least active (17.9 MET-h·wk−1). Among Latinos, participants born in Puerto Rico were 6.6–7.3 MET-h·wk−1 less active than participants born in Mexico, the United States/Canada, or other countries. There were variations in time use by race/ethnicity, with the largest difference in time spent sitting while watching TV. Black participants spent 14.8% of the day (~3.5 h) sitting watching TV, which was 78 min longer than Asian/Pacific Islander participants.

Conclusions: This study suggests that there are differences in LTMVPA accumulation by race, ethnicity, and nativity. Results can be used to identify demographic groups that may benefit from culturally tailored PA interventions.


Many adults across the United States do not meet the minimum amount of recommended moderate-to-vigorous intensity physical activity (MVPA).[1] According to national data from the Behavioral Risk Factor Surveillance System, the prevalence of physical inactivity differs vastly by race and ethnicity, with 31.7% of Latinos, 30.3% of Black Americans, and 23.4% of White Americans reporting no MVPA.[2] Similar patterns of racial/ethnic disparities in physical activity (PA) behaviors have also been demonstrated in small (less than 8000 participants) epidemiologic cohort studies utilizing accelerometry.[3,4] Understanding the racial/ethnic disparity in PA accumulation is important, as Black, Latino, Asian/Pacific Islander (API), and Indigenous Americans bear a disproportionate burden of physical inactivity–related chronic diseases, such as colon cancer, breast cancer, type 2 diabetes, and hypertension, relative to White Americans.[5–10]

Prior studies have been instrumental in identifying the presence of a broader racial/ethnic disparity in PA, but there are limitations to existing data. First, surveillance surveys tend to use single-item assessments of PA. Responses are often categorized into very broad cut points of PA time, such as meeting or not meeting PA guidelines, limiting our understanding of the magnitude of the disparity.[2,11,12] Similarly, many surveillance surveys do not include an assessment of sedentary time (any waking behavior characterized by an energy expenditure ≤1.5 metabolic equivalents, while in a sitting, reclining, or lying posture),[13] leaving gaps in our understanding of potential racial/ethnic disparities in this distinct behavior. Finally, surveillance and smaller cohort studies often categorize participant race/ethnicity very broadly—most often "Black," "White," "Hispanic," and "other." Importantly, these categories are socially constructed, and although they may be relevant for the US population for use within the census, health registries, and other legal documentation, there is growing recognition of the vast diversity within these broad categories.[14] Thus, patterns related to race/ethnicity may be better understood when studied in conjunction with other demographic information, such as nativity, current geographic region, and sex. In fact, some studies suggest that nativity (i.e., country of birth), in particular, is more strongly associated with PA behaviors than race/ethnicity,[12,15–17] but may be differentially related to PA depending on race/ethnicity.[11]

The Cancer Prevention Study-3 (CPS-3) is a large, modern cohort of over 303,000 adult participants across the United States and Puerto Rico. Capitalizing on the racial, ethnic, and geographic diversity in this cohort, the aims of this study were to 1) compare sex- and age-adjusted mean leisure-time moderate-to-vigorous intensity aerobic physical activity (LTMVPA) across racial/ethnic and nativity groups, including an in-depth exploration of LTMPVA within Latino and API subpopulations; 2) compare sex- and age-adjusted LTMVPA across said groups stratified by sex, age, geographic region, and occupational PA; and 3) compare the proportion of time spent per day sitting, sleeping, on light PA (LPA), and on MVPA by racial/ethnic group.