Main findings for mouthguard use, type and helmet age association with concussion
| Variables | Prospective cohort analysis | Nested case–control analysis | |||
| Crude IR/1000 hours,* (95% CI) | MICE IRR,† (95% CI) | P value | MICE OR,‡ (95% CI) | P value | |
| Mouthguard use | |||||
| No | 2.15 (1.66 to 2.65) | 1 (reference) | 1 (reference) | ||
| Yes | 1.38 (1.18 to 1.59) |
0.72§
(0.56 to 0.93) | 0.014 |
0.43§
(0.27 to 0.70) | 0.001 |
| Off-the-shelf | 1.27 (1.05 to 1.49) |
0.67§
(0.51 to 0.88) | 0.004 |
0.58§
(0.35 to 0.98) | 0.042 |
| Dentist-fitted | 1.59 (1.22 to 1.96) | 0.76 (0.57 to 1.01) | 0.058 |
0.31§
(0.16 to 0.60) | 0.001 |
| Helmet age | |||||
| <2 years old | 1.58 (1.32 to 1.84) | 1 (reference) | 1 (reference) | ||
| ≥2 years old | 1.51 (1.25 to 1.77) | 0.94 (0.75 to 1.15) | 0.526 | 1.16 (0.73 to 1.86) | 0.530 |
| Weight (kg) | NA | 1.00 (0.99 to 1.01) | 0.564 | 0.99 (0.98 to 1.01) | 0.322 |
| Level of play | |||||
| Under 13 | 0.93 (0.70 to 1.15) | 1 (reference) | 1 (reference) | ||
| Under 15 | 1.76 (1.43 to 2.08) | 1.22 (0.83 to 1.81) | 0.305 | 0.83 (0.39 to 1.76) | 0.626 |
| Under 18 | 1.97 (1.52 to 2.41) | 1.35 (0.83 to 2.17) | 0.223 | 0.56 (0.23 to 1.34) | 0.191 |
| Policy permitting body checking | |||||
| No | 1.12 (0.89 to 1.34) | 1 (reference) | 1 (reference) | ||
| Yes | 1.99 (1.67 to 2.32) |
1.54§
(1.13 to 2.11) | 0.007 | 0.63 (0.35 to 1.13) | 0.120 |
| Previous concussion | |||||
| No | 1.14 (0.94 to 1.33) | 1 (reference) | 1 (reference) | ||
| Yes | 2.25 (1.87 to 2.62) |
1.79§
(1.39 to 2.32) | <0.001 |
1.46§
(1.02 to 2.08) | 0.037 |
| Position | |||||
| Forward | 1.56 (1.32 to 1.80) | 1 (reference) | 1 (reference) | ||
| Defence | 1.67 (1.35 to 1.99) | 1.10 (0.88 to 1.36) | 0.387 | 1.32 (0.90 to 1.92) | 0.153 |
| Goalie | 1.03 (0.56 to 1.50) |
0.58§
(0.38 to 0.90) | 0.015 | 1.07 (0.51 to 2.25) | 0.854 |
*Crude incidence rates for concussion included 3953 player-seasons (360 concussions suffered by 307 players over the study period in 331 player-seasons) with 1157 team clusters (random effect) and 240 008.9 player-hours.
†MICE mixed effects multivariable negative binomial regression model for concussion included 4541 player-seasons (426 concussions suffered by 369 players) with 1230 team clusters (random effect) and 271 148.7 player-hours.
‡MICE mixed effects multivariable logistic regression model included 382 concussions (cases) suffered by 330 players over the study period in 352 player-seasons and 333 acute musculoskeletal injuries (controls) suffered by 281 players over the study period in 303 player-seasons, with 314 team clusters (random effect).
§Statistically significant finding based on apriori α (p<0.05). Level of play included under 13 (formerly Pee Wee, ages 11–12), under 15 (formerly Bantam, ages 13–14) and under 18 (formerly Midget, ages 15–17).
IR, incidence rate; IRR, incidence rate ratio; MICE, multiple imputation by chained equations; NA, not applicable.