V L Farmer, S M Williams, J I Mann, G Schofield, J C McPhee, R W Taylor

To investigate whether changing the play environment in primary schools to one that includes greater risk and challenge increases physical activity and reduces body mass index (BMI). A two-year cluster randomised controlled trial was undertaken in 16 New Zealand schools (Years 1-8). Intervention schools (n=8) redesigned their play environments to encourage imaginative and independent free play by increasing opportunities for risk and challenge (e.g. rough-and-tumble play), reducing rules, and adding new playground components (e.g. loose parts). Control schools (n=8) were asked to not change their play environment. A qualified playworker rated all school play environments at baseline and 1 year. Primary outcomes were moderate-to-vigorous physical activity (7-day accelerometry) and BMI z-score, collected in 840 children at baseline, one and two years. Data were analysed using generalised estimating equations. Multiple changes were made to the school play environments resulting in a significant difference in overall play evaluation score between intervention and control schools of 4.50 (95% CI: 1.82 to 7.18, P=0.005), which represents a substantial improvement from baseline values of 19.0 (s.d. 3.2). Overall, schools liked the intervention and reported many benefits, including increased physical activity. However, these beliefs did not translate into significant differences in objectively-measured physical activity, either as counts per minute (e.g. 35 (-51 to 120) during lunch break) or as minutes of moderate-to-vigorous physical activity (0.4, -1.1 to 2.0). Similarly, no significant differences were observed for BMI, BMI z-score, or waist circumference at one or two years (all P>0.321). Altering the school play environment to one that promoted greater risk and challenge for children did not increase physical activity, nor subsequently alter body weight. While schools embraced the concept of adding risk and challenge in the playground, our findings suggest that children may have been involved in different, rather than additional activities. Retrospectively registered with the Australia New Zealand Clinical Trial registry ID: ACTRN12612000675820.International Journal of Obesity accepted article preview online, 10 February 2017. doi:10.1038/ijo.2017.41.