Study Design. Only RCTs were included so as to limit the potential biasing effects of including studies with greater risks to internal validity. Studies that used either individual or cluster randomization were included. No restriction was placed on publication date or geographical location of the studies.