A very wide range of interventions to increase physical activity was appropriately considered. The authors grouped these interventions into six main categories, but the differences in costs and effects within these categories might have been relevant.
An appropriate approach was used to identify the relevant sources of evidence. The methods and conduct of the review were clearly stated and the inclusion of clinical trials should ensure the validity of the clinical evidence. Two systematic reviews were included in the analysis and this review was likely to have been comprehensive. The authors noted that trial quality was variable and most most of them used subjective measures of physical activity. The authors justified their selection of MET hours, which are specific to physical activity, as the summary benefit measure. They stated that quality-adjusted life-years (QALYs) might not have been appropriate because they involve assumptions about health benefits and do not adequately capture short-term benefits, such as mood improvement and improvements in musculoskeletal complaints. The methods used to translate other benefit measures into MET hours gained were reported in detail.
The categories of costs were consistent with the perspective stated. The key cost items were reported for each group of interventions, rather than for each intervention, meaning that the individual cost items were not stated and their unit costs and quantities of resources were not reported. This was probably due to the extremely high number of interventions compared. The price year was reported. Little information on the data sources was provided, but in general the authors used published estimates from implemented programmes.
Analysis and results:
The results were clearly presented. An average ratio rather than an incremental one was used, with each intervention presumably compared with no intervention rather than with each other. The authors estimated a threshold for cost-effectiveness to assess whether each group of interventions provided good value for money. The authors did not address the uncertainty and sensitivity analyses were not carried out. There was an online appendix, which might contain more cost-effectiveness results.
The clinical analysis was robust and was appropriately carried out and the authors’ conclusions appear to be valid.