Methodological quality and risk of bias
The assessment of bias is presented in Figure 2. There was a low risk of bias for amount or handling of incomplete outcome data and other bias in five RCTs (100%). Separately, the risk of bias was associated with inadequate generation of randomised sequences and inadequate concealment of allocations prior to assignments. There was also inadequate outcome reporting selection of participants in four studies (80%) and unclear outcome reporting in one study. The risk of bias associated with assessor blinding was unclear in three studies (60%) and low in two studies (20%). In addition, the risk of bias associated with participants and personnel blinding was low in one study (20%) and unclear in others. Although the risk of bias in blinding is generally unclear, this is considered to be due to the nature of the intervention. Overall, most of the included studies were classified as low risk for bias.