Skip to main content
. 2023 May 8;9(2):e001545. doi: 10.1136/bmjsem-2023-001545

Table 2.

Yearly BSI rates by study phase. Adjusted rates were determined used generalised estimating equation (GEE) models for Poisson regression adjusted for age and school and accounting for within-subject correlation, using data across all phases of the study at once to make years comparable

Phase N BSI events Persons with BSI Person-years Crude rate Adjusted rate 95% CI
Historical
 Year 1 35 18 18 29.9 0.60 0.59 (0.39 to 0.91)
 Year 2 36 14 11 32.1 0.44 0.42 (0.22 to 0.77)
 Year 3 31 16 10 28.1 0.57 0.57 (0.31 to 1.02)
Total 56 48 30 90.2 0.53 0.52 (0.38 to 0.71)
Pilot
 Year 1 30 5 4 28.6 0.18 0.19 (0.08 to 0.45)
 Year 2 34 19 14 29.8 0.64 0.59 (0.36 to 0.97)
 Year 3 44 19 17 36.8 0.52 0.48 (0.30 to 0.75)
Total 66 43 27 95.2 0.45 0.41 (0.30 to 0.57)
Intervention*
 Year 1 49 31 19 42.7 0.73 0.67 (0.43 to 1.05)
 Year 2 43 13 11 38.2 0.34 0.32 (0.17 to 0.57)
 Year 3 38 10 8 34.6 0.29 0.27 (0.13 to 0.55)
 Year 4 36 8 8 21.9 0.37 0.38 (0.20 to 0.73)
Total 78 62 34 137.3 0.45 0.43 (0.31 to 0.59)

*P value for historical versus intervention phase (derived from model that includes both phases plus age and school) = 0.34

BSI, bone stress injury .