HOW REPRESENTATIVE ARE PHYSICAL ACTIVITY TRIAL PARTICIPANTS? - RESULTS FROM THE PACE-UP PRIMARY CARE TRIAL

Background The generalisability of evidence from a randomised trial can be compromised if there are key differences between participants and the target population. Physical activity is important to health yet physical activity intervention studies often have low recruitment rates. If health promotion interventions preferentially reach those from more affluent backgrounds they could exacerbate inequities in access to care and increase health inequalities. Identifying inequalities in uptake of interventions is therefore very important. Methods PACE-UP is a large population-based walking intervention trial in inactive 45–74 year-olds recruited through seven general practices in south west London. Anonymised demographic data are available through general practice records on all those invited. Patients who initially indicated they would like to take part in the trial are classified as participants. Logistic regression was used to obtain odds ratios (ORs) for participation by demographic factors. In addition, 1125 non-participants filled out a questionnaire providing information including physical activity levels and reasons for non-participation. Results 8279 people were invited to participate in PACE-UP in five of the seven practices and 1364 (16.5%) agreed to participate. 49% of invitees were aged 45–54 years. Of these 13.1% agreed to participate, compared with 18.9% of 55–64 year-olds and 20.8% of 65–74 year-olds. ORs are 1.5, 95% CI 1.4–1.8, and 1.7, 95% CI 1.5–2.0, respectively. Participation was strongly associated with deprivation, as measured by quintile of Index of Multiple Deprivation score, with 23% of those in the least deprived quintile wanting to participate compared with 13% in the most deprived quintile: OR 2.0 (95% CI 1.6–2.6). Ethnicity was recorded in practice records for 6639 invitees. 69% were white, 12% Asian and 13% black. Practices with a higher proportion of patients of black ethnicity had lower participation rates; Spearman’s rho=0.90, p = 0.037. Responders to the questionnaire had similar age profile to the participants. Main reasons given for non-participation were: being very physically active (n = 660), and not having enough time (n = 460). Of those who stated they were already active, only 49% were classed as active by the General Practice Physical Activity Questionnaire suggesting that some of these non-participants were over-estimating their physical activity levels. Fuller analysis from all seven practices will be presented. Conclusion Individuals who agreed to participate in a large primary care trial to promote physical activity were older and from less deprived areas than non-participants. If interventions to increase physical activity fail to reach less affluent populations, this may increase existing health inequalities.