Can a multi-level intervention approach, combining behavioural disciplines, novel technology and incentives increase physical activity at population-level?

Abstract

Background: Despite a global call for action and growing burden of non-communicable diseases (NCD) associated with physical inactivity, effective interventions to increase community-wide physical activity (PA) remain few. NCDs accounted for 80% of Singapore’s disease burden (2015) and yet 40% of Singaporeans did not meet minimum recommended weekly PA despite evidence of the benefits to cardiorespiratory health, diabetes and cancer prevention. Methods: A large-scale public health intervention was initiated in 2015 to increase population-level PA through incidental daily walking. Intervention components included fitness trackers, redeemable rewards and gamification, implemented in a mutually-reinforcing manner within an eco-system supportive of PA and informed by real-time data analytics. Mean daily step count at baseline and post-intervention were compared across periods, and the influence of participant sub-groups characteristics on overall results, using significance tests. Standards for Reporting on Implementation Studies (StaRI) were adhered to. Results: Intervention reach increased fourfold from 129,677 participants in wave 1 (2015–16) to 690,233 in wave 3 (2017–18) amounting to a total of 1,184,410 Step Challenge participations. Mean days of fitness tracker use increased from 2.4 to 5.0 days/week among participants completing the Challenge in wave 1 and from 5.3 to 6.0 days/week in wave 3. The mean number of daily steps between pre-Challenge and Challenge periods increased by 4163 (sd=1360; p< 0.001) in wave 1, by 2242 (sd=334; p< 0.001) in wave 2 and by 1645 steps/day (sd=54; p< 0.001) in wave 3. Mean daily step increases between wave 1 and 3 also suggest that incidental PA was maintained, a finding supported by a 2017 national population survey showing that incidental PA among adults increased from 5% in 2010 to 14% in 2017 while moderate-intensity PA increased from 5 to 10% over the same period. Conclusion: Population-level PA was effectively increased through multi-level interventions integrating technology, behavioural economics, gamification, marketing, communications and community linkages within a supportive context- and climate-appropriate environment. Responsive data analytics were instrumental to strengthen implementation by tailoring modalities that maximise effectiveness at population-level. Further analyses are needed to explore potential barriers, challenges or unmet needs in sub-groups with lower uptake to tailor future interventions for greater reach and impact.