Table A2.
A Range of Exposure and Risk Measures Derived From Injury Surveillance Dataa
| Statistic | Value | Calculation | Explanation | Comment |
|---|---|---|---|---|
| Injury statistics | ||||
| No. of injuries (carrier injury replacements in 434 matches) | 53 | Nil | Count of the number of tackler injuries requiring the injured player to be replaced observed in 434 matches | The “numerator” used for calculating the rate of tackler replacement injuries per unit of time or per tackle. Absolute numbers and costs of injuries are of interest to risk managers, especially when provided in parallel with rates. |
| No. of injured players (some were injured more than once) | 48 | Nil | This is the numerator for calculating injury risk. | |
| Exposure measures | ||||
| Player-hours in 434 matches | 17,360 | 30 × 579 | 30 players (15 from each team) multiplied by 579 (hours of play in 434 matches of 80 minutes’ duration) | This number provides a “time-window” denominator. Usually, it is assumed that time lost for yellow and red cards, or time gained for “extra time,” is negligible and is ignored. |
| No. of single-tackler tackle events in 434 matches | 43,366 | Nil | All tackles in 434 matches were coded, regardless of whether they resulted in injury | This number forms an “event-based” denominator. |
| No. of players who appeared in the 434 matches | 1403 | Nil | This is a count of the size of the cohort across the entire study period. It is used as the denominator for calculating injury risk. | |
| No. of full player matches | 13,020 | 30 × 434 | 30 players (15 from each team) multiplied by 434 matches | This number provides a “per-match” denominator. |
| No. of athlete-exposures (athlete-participations) | 17,685 | Nil | Count of the number of players who took the field over 434 matches (players can be substituted for tactical purposes or replaced due to injury) | The similarity to the number of player-hours is coincidental. There are 40 hours of player time per match, and the average number of athlete-exposures per match over this series of matches was 40.8. |
| Risk measures | ||||
| Period prevalence (percentage of cohort injured) | 3% | (48/1403) × 100 | Percentage of people who appeared in matches who were replaced | Often reported as the “risk per season” or “risk per year.” It cannot be easily used to compare between activities if the duration of surveillance varies from activity to activity. The longer the surveillance period, the higher the risk will appear to be for closed cohorts. |
| Injuries per 1000 player-hours | 3.1 | (53/17,360) × 1000 | Number of injuries is divided by the number of hours of player exposure and multiplied by a scaling factor (eg, 1000, 10,000) to provide a rate that is convenient to work with (eg, numbers in the range of 1 to 1000 rather than numbers <0 or >1000) | The most commonly reported metric of injury rates in studies of rugby injury epidemiology has been the rate of injuries per 1000 player-hours. This convention is endorsed in the consensus statement by Fuller et al.52 It is relatively simple to estimate based on the number of matches played. Comparisons of incidence rates between activities or within activities over time based on this denominator require the assumption that the number and characteristics of energy transfers to which participants are exposed remain relatively constant per unit of exposure time. |
| Injuries per 1000 matches | 122 | (53/434) × 1000 | Rate of tackler replacements per rugby union match multiplied by 1000; rate per match multiplied by a factor that provides a convenient interpretation (0.12 carrier replacement injuries per match, 12.2 per 100 matches, 122 per 1000 matches, etc) | Ignores the number of players and match duration and provides an estimate of the number of injuries that an observer would expect to see if they watched 1000 matches. Not useful for comparing incidence rates between activities of differing durations or numbers of participants. |
| Injuries per 1000 hours of play (ignoring number of players) | 92 | (53/579) × 1000 | Rate per hour is multiplied by a factor that provides a convenient interpretation (0.9 carrier replacement injuries per hour, 9.2 per 100 hours, 92 per 1000 hours, etc) | Ignores number of players and provides an estimate of the number of injuries an observer would expect to see if they watched 1000 hours of play. Not useful for comparing between activities with differing numbers of participants because the sizes of the populations at risk differ. |
| Injuries per 1000 athlete-exposures (athlete-participations) | 3.0 | (53/17,685) × 1000 | Carrier injury replacements per 1000 athlete-exposures | Injuries per 1000 athlete-exposures are commonly reported in injury surveillance in the United States. Problematic for comparing between activities that have different numbers of typical athlete-exposures per match or when the average exposure time per player changes over time. |
| Injuries per 1000 full player matches | 4.1 | (53/13,020) × 1000 | Not commonly used. It ignores the duration of the match and, as such, has similar drawbacks to reporting injuries per athlete-exposure because the time window of exposure varies between activities of different durations. | |
| Injuries per 1000 “ball in play” player-hours | 6.8 | (53/7740) × 1000 | Not commonly used but technically a more accurate measure of exposure than injuries per 1000 player-hours because players are only exposed to tackles when the ball is “in play.” | |
| Injuries per 1000 “ball in play and ball carrier’s team in possession” player-hours | 13.5 | (53/3819) × 1000 | Again, not commonly used but an even closer approximation of the actual time exposed to the risk of ball carrier injuries. Players are only tackled when the ball is in play and their team is in possession. | |
| Injuries per 1000 tackle events | 1.2 | (53/43,366) × 1000 | Ball carrier injury replacements per 1000 times tackled | Provides an accurate assessment of per-event injury rates but in isolation ignores the frequency of occurrence of the event of interest. Injury rates per event have sometimes been termed “injury propensity.”48 |
| Injuries per 1000 players per year | 24 | ([23 + 17 + 13] × 1000) / (983 + 589 + 627) | Sometimes provided as a gross estimate of the injury risk when participant numbers and injury numbers are available but no measure of exposure for players is available (eg, data derived from insurance claims combined with registers of participants). Of limited use when exposure varies by subgroup or across sports. | |
aExamples from a study of rugby tackle injuries.87