Article Text

Download PDFPDF

OP81 Applying hurdle models to estimate socioeconomic inequalities in moderate-to-vigorous physical activity: analysis using the health surveys for england 2008 and 2012
  1. S Scholes,
  2. JS Mindell
  1. Epidemiology and Public Health, University College London, London, UK


Background Guidelines recommend adults engage in at least 150 min of moderate-to-vigorous physical activity (MVPA) per week. It is unknown whether describing socioeconomic inequalities using the average amount of time persons spend in MVPA masks: (1) disparities in the proportion of persons that are active, or (2) disparities in the amount of time that persons who are active spend engaging in physical activity.

Methods Hurdle models are a new way of accommodating continuous physical activity data with: (1) an excessive amount of zeros (non-participation), and (2) a continuous positively-skewed part (the amount of time active persons spend being active). Using the Health Surveys for England (n=16,012; HSE 2008; 2012), we applied hurdle models to estimate inequalities in these two separate parts of MVPA data, and assess changes over time. Analyses were sex-specific and adjusted for body mass index and smoking. Separate analyses were performed for overall MVPA and for five activity domains, including walking and sports/exercise.

Results are presented as Marginal Effects (ME) with 95% Confidence Intervals (95% CIs). The MEs represent absolute differences between the highest- and lowest-income groups in: (1) the percentage of participants who were active, and (2) the average hours-per-week (hpw) spent in MVPA conditional on participants being active (i.e. hpw being greater than zero).

Results The proportion of participants who performed any activity was highest in the highest-income group. The ME for overall MPVA was 12.5 percentage points (pp) [95% CI 10.3 to 14.7 pp] in men; 11.6 pp [9.5–13.7 pp] in women. Similar patterns were found for walking [men: 19.8 pp: 16.7–22.8 pp; women: 15.0 pp: 12.4–17.6 pp] and for sports/exercise [men: 20.0 pp: 16.9–23.2 pp; women: 23.1 pp: 20.4–25.9 pp].

Differences in the amount of time spent in overall MVPA (amongst those doing any) also favoured high-income participants [men: 3.5 hpw: 2.4–4.7 hpw; women: 3.3 hpw: 2.5–4.2 hpw]. High-income participants spent on average 1 hpw more doing sports/exercise [men: 0.9 hpw: 0.0–1.8 hpw; women: 1.2 hpw: 0.7–1.7 hpw]. However, time spent walking (amongst those doing any) showed the opposite pattern in men being 1.9 hpw lower for those in the highest-income group [−2.8 to −0.9 hpw]. Patterns were similar in 2008 and 2012. Findings were robust to different model specifications (e.g. using two-part models).

Conclusion Inequalities in overall MVPA and in sports/exercise were pronounced for the hurdle of participation and for the amount of time spent being active. For walking among men, inequalities were sharpest for the hurdle of participation, highlighting the importance of interventions designed to increase walking among inactive low-income individuals. Our results will be updated when HSE 2016 data are available (spring 2018).

  • inequalities behaviours methods

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.