Background Evidence is unclear on whether inequalities in average levels of moderate-vigorous physical activity (MVPA) reflect differences in participation, differences in the amount of time spent active, or both. Using self-reported data from n=4874 adolescents aged 11–15 in England (Health Survey for England: 2008, 2012, 2015) and n=3065 adolescents aged 12–15 in the US (NHANES: 2007–16), we examined inequalities in these separate aspects for overall- and domain-specific MVPA.

Methods Socioeconomic position (SEP) was indexed by tertiles of equivalised household income (England) and the family income-to-poverty ratio (FIPR, US). Hurdle modelling is a novel way of analysing MVPA data with: (1) excessive zeros (non-participation), and (2) a continuous positively-skewed part (the amount of time active participants spend being active). We applied gender- and country-specific models to estimate inequalities in three aspects: (1) the probability of doing any MVPA, (2) the average hours-per-week (hpw) spent engaged in MVPA, and (3) the average hpw MVPA conditional on participation (MVPA-active). Using complete-case analyses adjusted for the complex survey design, absolute differences in MVPA (e.g. hpw) between adolescents in the highest versus lowest SEP were summarised using average marginal effects (AMEs) with 95% Confidence Intervals (95% CIs) after confounder adjustment (body mass index).

Results Inequalities in overall MVPA were observed in the US, but not in England. For example, the AMEs for girls in the US in the highest versus lowest SEP were 3.2 hpw (95% CI: 1.9 to 4.6 hpw) and 3.0 hpw (1.7 to 4.4 hpw) respectively for MVPA and MVPA-active. Inequalities in sports participation were evident for girls in both countries (AMEs for sports MVPA-active: England: 0.7 hpw (0.1 to 1.4 hpw); US: 2.5 hpw (1.4 to 3.6 hpw), and for boys in the US (AME: 2.0 hpw; 0.6 to 3.5 hpw). In contrast, boys in the highest versus lowest SEP spent less time on average in active travel (AME for MVPA: England: -0.3 hpw (-0.6 to 0.1 hpw); US: -0.6 hpw (-1.2 to 0.1 hpw)); this finding mainly reflected the difference between SEP groups in the probability of doing any active travel. Girls in the US in the highest versus lowest SEP showed a similar propensity to do any active travel, but on average spent less time engaged in active travel. Findings were robust to different model specifications (e.g. using two-part models).

Conclusion Monitoring inequalities in PA requires assessing different aspects of the distribution within each domain.