BackgroundPhysical inactivity is a major public health concern and more innovative approaches are urgently needed to address it. So far, physical activity interventions have largely ignored the extent to which an individual's social network can affect physical activity behaviours. The sparse evidence base has mainly relied on self-reported, cross-sectional data for social networks. However, novel technologies and analytical techniques afford new and promising means to obtain and analyse these data. MethodsThe purpose of this study was to investigate the characteristics and effect of social networks on physical activity behaviour in adults in a workplace setting. The specific objective was to investigate the structure of the social networks by individual factors (eg, sex and age) and healthy behaviours (eg, physical activity, smoking, and body-mass index). Data were obtained from 406 adults who participated in a 12-week physical activity intervention. Sensors were placed along footpaths, in the gym, and in the workplace. Participants scanned their loyalty card at the sensor when undertaking physical activity (eg, walking), which created a transaction (timestamp) logging their activity. With the near field communication and radio frequency identification technology, we generated the social network with time between card swipes and frequency of similar locations of card swipes. When a participant's social network data were unavailable, we excluded them from the analyses. Self-reported physical activity was measured with the Global Physical Activity Questionnaire (GPAQ) at baseline, week 12, and 6 months. Regression modelling investigated the effect of number of so-called physical activity buddies and frequency of doing activity with a buddy. The statistical models accounted for four confounding variables: age, sex, education level (highest qualification attained), and staff grade. FindingsWe obtained follow-up data for 353 (87%) participants. The mean age of included participants was 43·32 years (SD 9·37). 272 (67%) were women. At baseline, 215 (53%) participants were categorised as having low activity levels with GPAQ. Regression analysis indicated that a rise in the number of buddies during the 12-week intervention was associated with increased physical activity. Coefficients showed that each additional friend was associated with a rise of 1·9 min (95% CI −0·12 to 3·8) of physical activity. InterpretationIndividuals who had a physical activity social network used their loyalty card more often and for longer than did those without such a network. Increased numbers of friends and social bonds for physical activity were associated with increased activity levels. Strategies to foster friend support for physical activity could be important for increases in activity. However, the focus of the research was the effect of buddies on physical activity behaviours. As such, the analyses were limited to the effect of individuals who were enrolled in the intervention; we did not consider attempting to model the broader social networks that exist outside of the workplace. Additionally, we did not account for reverse causality. Innovative methods that incorporate new technologies can be used to develop new peer-based physical activity interventions. FundingNational Prevention Research Initiative (G0802045) and their funding partners (Alzheimer's Research Trust; Alzheimer's Society; Biotechnology and Biological Sciences Research Council; BHF; Cancer Research UK; Chief Scientist Office, Scottish Government Health Directorate; Department of Health; Diabetes UK; Economic and Social Research Council; Engineering and Physical Sciences Research Council; HSC R&D Division of the Public Health Agency; MRC; Stroke Association; Welsh Assembly Government; World Cancer Research Fund; Department for Employment and Learning, Northern Ireland [M6003CPH]).
Read full abstract