Table 3 Possible scenarios (‘best’ and ‘worst’ case) for scaling up PACE

A. Best-case scenario: Scale-up removing in-person strategies with best-case effect

Strategy 3b is removed. Strategy 5a and 5b are removed as a health service delivered strategy, and assumed to be performed by the in-school champion as part of their program role.

The effect size has been reduced using a scale-up penalty based on a recent review. In this analysis, it is assumed that 60% of the original treatment effect is maintained. Specifically, the estimated average treatment effect and lower and upper intervals were each reduced by 40% to obtain a re-calculated treatment effect based on scale-up penalty that maintains 60% of the original effect size.

It is assumed that delivery of the in-person training is unfeasible for wide scale-up. An alternate version of PACE is proposed whereby in-school champions and teachers receive training via a self-directed online course. This course is envisaged to be integrated as part of teachers continuing professional education, thus no additional costs are assumed. The team are currently developing this optimised iteration of PACE.

Systematic review evidence determined that the effect estimates from scale-up of effective interventions are reduced. However, the extent to which the estimated effects are reduced is variable. For this scenario, we adjusted the effect size based on a recent review finding that physical activity interventions delivered in community settings (e.g. schools) maintain a median 60% of the original effect size [65].

B. Worst-case scenario: scale-up removing in-person strategies with worst case effect

A replication of Version A assuming a worst-case scale-up penalty of only 25% of the original effect size is maintained.

Specifically, in this scenario the estimated average treatment effect and lower and upper intervals were each reduced by 75% to obtain a re-calculated treatment effect based on a scale-up penalty that maintains 25% of the original effect size.

Systematic review evidence determined that the effect estimates from scale-up of effective interventions are reduced. However, the extent to which the estimated effects are reduced is variable. This scenario provides the worst case scenario in terms of possible scale-up penalty, with the effect size adjusted based on the lower end of the scale-up penalty values found in the recent review [65].