these schools served as a natural control group for background injury rates during the study period. Injury rates per 1000 students per month, relative risks (RRs) and 95% confidence intervals (CIs) were calculated, adjusting for clustering within schools.