Article Text

Download PDFPDF

Risk of injury and concussion associated with team performance and penalty minutes in competitive youth ice hockey
  1. Carolyn A Emery1,2,3,
  2. Jian Kang2,
  3. Kathryn J Schneider1,
  4. Willem H Meeuwisse1,3
  1. 1Sport Injury Prevention Research Centre, Roger Jackson Centre for Health and Wellness Research, Faculty of Kinesiology, University of Calgary, Calgary, Canada
  2. 2Alberta Children's Hospital, Department of Pediatrics, Faculty of Medicine, University of Calgary, Calgary, Canada
  3. 3Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Calgary, Canada
  1. Correspondence to Carolyn Emery, Sport Medicine Centre, University of Calgary, 2500 University Drive NW, Calgary, Alberta T2N 1N4, Canada; caemery{at}


Objective To determine if there is an association between the risk of all injury or concussion and win–loss records or penalty minutes in competitive youth ice hockey players (ages 11–14).

Design, setting, participants This is a secondary data analysis of a 2-year cohort study in Alberta and Quebec on the 2007/2008 (Pee Wee) and 2008/2009 (Bantam) seasons.

Main outcome measures Incidence rate ratios (IRRs) were estimated based on Poisson regression for game-related injury and concussion and adjusted for cluster by team.

Results A total of 140 teams from Alberta (n=2081) and 137 teams (n=2018) from Quebec were included in the analysis. There were 451 game-related injuries (121 concussions) from Alberta and 280 (62 concussions) from Quebec. For game-related injuries, the IRR between players from teams with more than 50% wins and players with less or equal to 50% wins was 0.78 (95% CI 0.64–0.95) for all injuries, 0.75 (95% CI 0.52–1.08) for concussions, 0.64 (95% CI 0.47–0.88) for injuries resulting in time loss of more than 7 days, and 0.74 (95% CI 0.39–1.40) for concussions resulting in time loss of more than 10 days; adjusting for clustering by team and other important risk factors (ie, province, age, level of play, previous injury, weight and position). There was no association found between the total penalty minutes per game and game-related injury or concussion.

Conclusions There was a significant association found between team performance (ie, win/loss/tie record) and injury risk with a 22% lower injury rate and 36% lower injury rate resulting in less than 7 days time loss in Pee Wee and Bantam ice hockey teams winning more than 50% of all season games. Total penalty minutes per game were not associated with injury or concussion rates.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Ice hockey is a popular North American winter sport with high participation rates.1,,3 However, ice hockey injuries in youth have drawn increasing concern in recent years. Canadian data suggest hockey injuries account for more than 10% of all youth sport injuries.3 4 Body checking has been associated with 45–86% of injuries among paediatric ice hockey players.5,,8 In a recent study, consistently greater risk of injury and concussion has been found in players exposed to policy that allows body checking, players in the lowest 25th percentile by weight, games compared with practices, older age groups and higher levels of competition.9 10 Penalty minutes often result from aggressive behaviours including illegal body checking in leagues where body checking or other illegal contact (ie, cross-checking, roughing or tripping) is not allowed, and are frequent mechanisms of injury in youth ice hockey.11,,16 In Quebec, a reward system based on penalty minutes called by referees was in place at the time of this cohort study.17 While there was an emphasis on fair play in Alberta, there was no universal reward system, rather local incentives. There is evidence, however, that injury rates and the observed number of transgressions do not differ in a Bantam League that rewards teams through a Fair Play point system for low penalty minutes compared with a Bantam League with no reward system.18 What remains unknown is if there is an association between team penalty minutes and injury or concussion risk specifically. In addition, a trend has been previously demonstrated in elite soccer examining the association between team performance and total team injury days.19 The purpose of this secondary analysis was to determine the association between risk of injury and concussion with team level performance based on win–loss record or total team penalty minutes in youth ice hockey (ages 11–14).


The primary objectives of this study were to examine whether injury risk was associated with team win–loss record and mean team penalty minutes per game. The authors estimated the incidence rate ratios (IRRs) comparing teams with more than 50% wins versus teams with less than or equal to 50% wins. IRRs associated with penalty minutes were also investigated by dichotomising teams into two groups based on penalty minutes per game (ie, >50th percentile or ≤50th percentile) within a given league.

Study design and subjects

This is a secondary data analysis including data from two consecutive independent cohort studies examining the risk of injury associated with body checking in Pee Wee ice hockey players9 (aged 11–12 years, top 60% of divisions of play in the 2007/2008 season) and the risk of injury associated with body checking experience in Bantam ice hockey players (aged 13–14 years, top 30% of divisions of play in the 2008/2009 season).10 The details of the cohort study methodology have been previously published.9 10 In Pee Wee, body checking was allowed in Alberta, Canada, at all levels of competitive play and was not permitted in Quebec, Canada. In Bantam, body checking was allowed in both provinces in the top 30% of divisions of play in Quebec and in all levels of competitive play in Alberta. In total, 4125 players from 280 teams (2154 Pee Wee players from 150 teams and 1991 Bantam players from 130 teams) were recruited. Individuals who had sustained an injury or had chronic illness that prevented full participation in ice hockey at the beginning of the season of the study were excluded. Level of play was defined within age group. In Pee Wee, the cohort included only the upper 60% of divisions of play overall; thus, the study upper level refers to top 20% and lower level referred to mid-40%. In Bantam, the cohort included only the upper 30% of divisions of play overall; thus, the study upper level includes the top 10% and lower level includes the next 20%.

Data collection was the same in both age groups, Pee Wee and Bantam. Preseason questionnaires included height, weight, date of birth, years of hockey participation, division of play, player position, and injury and concussion history. The weekly exposure sheet was used to collect the daily participation data on each consenting player. Team performance including the outcome of each game and the total team game penalty minutes was also recorded.


The injury surveillance system used in both cohorts was developed for use in university varsity sport and adapted and validated for use in community youth ice hockey.11 20 Baseline questionnaires were collected prior to the start of the season. A team designate assigned on each team (ie, safety coach) was responsible for collecting weekly exposure data and initiating an injury report form which would be completed by the team therapist (a physiotherapist, an athletic therapist or a senior therapy student) who followed the players weekly during the season and assessed all injuries. Injury report forms were completed by the team therapist following assessment and follow-up occurred until the player returned to play. The injury definition included any injury occurring in ice hockey that resulted in the inability to complete a session, missing a subsequent session and/or required medical attention. Every attempt was made for the team therapist to see the player as soon as possible following the injury. Any player with a suspected concussion or significant time-loss injury (>1 week) was referred to the study sport medicine physician for follow-up.


Percentage of wins for each team was calculated as the total number of games won divided by the total number of games that the team played throughout the season. Win–loss record was calculated as the percentage of wins and dichotomised as greater than 50% versus less or equal to 50% wins. Teams with no record of win–loss tie were excluded. Penalty minutes per game were estimated by taking the total penalty minutes throughout the year for that team divided by the total number of games the team participated in. Teams were dichotomised into low- or high-penalty minute groups by the median penalty minute per game in each age group (ie, Pee Wee or Bantam). IRRs for win–loss record and penalty minutes for each of game injury, game concussion, game injury resulting in >7 days time loss and game concussion resulting in >10 days time loss with 95% CI were estimated using Poisson regression. Time loss >7 days was based on the consensus definition for moderate injury severity.21 Time loss >10 days for concussion was based on the Zurich (2006) Concussion in Sport Consensus statement for complex concussion available at the time of the cohort study.22,,26 Other risk factors including the level of play (upper versus lower in each age group), player size (<25th percentile by weight), year of play (first or second year in an age group), age group (Pee Wee or Bantam), previous history of injury in the past year, previous concussion (lifetime) and position of play (forward, defence, goalie) were adjusted in the model to assess confounding. Each player was classified as upper or lower level player based on the cut-off established previously for Pee Wee and Bantam.9 10 IRRs between Alberta and Quebec were different in Pee Wee and Bantam from previous studies based on body checking policy differences (ie, no body checking in Alberta Pee Wee cohort only).9 10 Therefore, the effect modification of age group (ie, Pee Wee or Bantam) and province (ie, Alberta or Quebec) was included in all models. Joint effect modifications by age group and province for each of win–loss record and penalties were examined. In all models, the game hours were included as an offset and cluster adjustment by the team was performed. Significance was set at 0.05 and all hypothesis tests were two-sided. All statistical analyses were performed using Stata version 10.027.

Ethical considerations

All players and their parents provided a written consent to participate in the study. This study is a secondary data analysis of the studies that have previously obtained ethics approval from the Office of Medical Bioethics at the University of Calgary, University of Alberta, McGill University, Université de Montreal and Laval University.9 10


Twenty-six players from three teams that did not report win–loss tie record were excluded from the analysis. The analysis thus included a total of 140 teams from Alberta (n=2081 players) and 137 teams (n=2018 players) from Quebec. There were 451 game-related injuries (142 injuries resulting in >7 days time loss, 121 concussions, 25 concussions resulting in >10 days time loss) in Alberta and 280 game-related injuries (116 injuries resulting in >7 days time loss, 62 concussions, 18 concussions resulting in >10 days time loss) in Quebec. The proportion of games missed by the win–loss tie record was 9.74% (95% CI 8.83–10.65%). Descriptive statistics for all risk factors are presented in table 1. In Pee Wee, a median of 34 (range 19–57) team games were played in Alberta and 41 (range 19–59) team games were played in Quebec. In Bantam, the median number of team games played in Alberta was 41 (range 26–70) and that in Quebec was 45 (range 25–73). Seventeen players from Bantam and 31 players from Pee Wee had different percentages of wins and penalty minutes than the team, likely due to affiliation with other teams. Team records for wins and penalty minutes were assigned to these players. Penalty minutes per game were dichotomised based on the median as the cut-off in each cohort (Bantam: 11.51 min (IQR 9.46–13.75); Pee Wee: 7.71 min (IQR 5.85–9.63)). Weight was dichotomised based on the 25th percentile as the cut-off in each cohort (52 kg for Bantam and 38 kg for Pee Wee).

Table 1

Descriptive statistics for all risk factors (Pee Wee cohort 2007–2008 and Bantam cohort 2008–2009 season)

Univariate analysis for win–loss and penalty record is presented in table 2. The game-related injury rate for teams with greater than 50% wins was 25% lower than the injury rate for teams with less than or equal to 50% wins. For time-loss injuries >7 days, the injury rate was 36% lower for teams winning more than 50% of games compared with the rate for teams with less than or equal to 50% wins. The evidence was insufficient to support a higher injury risk in teams with high penalty minutes per game compared with teams with low penalty minutes.

Table 2

Incidence rate ratios of game-related injury, concussion, severe injury and severe concussion by win–loss record and penalty record based on univariate analyses

To adjust for other risk factors, the effect modification between cohort and province was included in all models. Pee Wee players in Quebec were chosen as the referent group because it was the only province and age group (ie, Pee Wee and Bantam) where body checking was not allowed. Estimated IRRs from the other three groups (Pee Wee Alberta, Bantam Quebec, Bantam Alberta) compared with the referent group are presented in table 3. For each of win–loss and penalty record, no significant effect modifications by age group or province were detected.

Table 3

Multivariate risk factor analyses for game-related injury, injury resulting in >1 week time loss, concussion and concussion resulting in > 10 days time loss in Pee Wee (2007–2008 season) and Bantam (2008–2009 season)

Table 3 summarises the results for risk factor analysis based on multivariate Poisson regression models for injury, concussion, injury >7 days time loss and concussion >10 days time loss during games. In this model, the game hours were included as an offset; clustering by the team was accounted for with adjustment for all included covariates. Teams winning more than 50% of all games throughout the season had a lower injury rate than teams with less than or equal to 50% wins (IRR = 0.78 (95% CI 0.64–0.95)). Teams with more than 50% wins in all games had significantly lower injury rate (including time loss >7 days) than those with less than or equal to 50% wins (IRR = 0.64 (95% CI 0.47–0.88)). The point estimate also suggests a clinically relevant difference in concussion rate in teams with more than 50% wins (IRR = 0.75 (95% CI 0.52–1.08)). There was insufficient evidence to suggest that teams with >50th percentile penalty minutes had higher injury risk than teams with short penalty minutes. First year players had higher injury risk than second year players. Previous injury was a risk factor for injury and injury resulting in more than 7 days time loss. Previous concussion was a risk factor for concussion and concussion with more than 10 days time loss. Playing at the goalie position was associated with lower risk of injury, concussion and severe injury than playing at the forward position.


To our knowledge, this is the first study in elite youth ice hockey to investigate the association between injury risk and team performance based on game win–loss record and penalty record. In this study, the teams winning more than 50% of their games had a significantly lower injury rate than other teams or all injuries and injuries resulting in >7 days time loss. Point estimates suggest a similar trend for concussion but these findings were not statistically significant. There was no evidence to support any association between high team game penalty minutes and injury rate. Comparing results based on other individual level risk factors in this analysis and those from the analysis of Pee Wee,9 weight and the level of play were no longer significant predictors for injury. All other risk factor findings were consistent with those from the Pee Wee and Bantam studies.9 10

A previous study in adult elite soccer showed that physical fitness of the team players was associated with team success and a trend was found to support a better standing for teams with lower injury rates.19 For players with higher physical fitness, it was not surprising to fewer injuries. The team record in games was also related to the physical fitness of its players. This study supports the notion that there is an association between performance and injury rates in that teams that had lower injury rates were more likely to have a win–loss record >50%. There was no measure of physical fitness in this study.

Penalty minutes in this study were consistent with findings in other studies.17 18 To the authors' knowledge, this study was the first to examine the association of game penalty minutes with injury rates in youth ice hockey.


With a therapist present at only one session each week, it is possible that minor injuries may have been underestimated if the team designate was not aware of the injury. However, in the weekly follow-up, the therapist was to communicate with the team designate and players to reduce the number of missed injuries. All players with a suspected concussion were referred directly to the study physician. In addition, it is unlikely that this reporting issue differed by province.

Win–loss record and penalty minutes were dichotomised from continuous scales. When converting a continuous variable into a categorical variable, information can be lost. Clinically speaking, however, the interpretation of results from dichotomous covariate was arguably more meaningful than that based on a continuous scale. Furthermore, dichotomising percentage of wins reduced the inconsistency between player and the team on win–loss record.

Although win–loss record was an assumed team level covariate, the players on some teams had different win–loss percentages from the team due to lack of participation for various reasons (ie, illness, injury and penalty outcome of game suspension) or affiliation with another team. One may consider the mean percentage of wins throughout the year as a remedy. Interpretation would become difficult, however, and not as meaningful as that for the penalty minute covariate.

A limitation in acquiring penalty minute data reported on the weekly exposure sheet included incomplete detail to differentiate penalty types. Future research should examine infraction type to differentiate between intentional and non-intentional penalties assessed.

Cut-offs of the level of play in Pee Wee and Bantam were different. The authors argue that for a given player, they were classified into high or low level of play based on the division of play within his age group. As all are concerned about the relative IRR between high versus low level of play within an age group cohort, the authors feel this is an acceptable dichotomisation. Therefore, the difference in cut-offs between the two age groups did not affect this conclusion.

Conclusions from this study can only be drawn on associations. Causal effect cannot be determined based on the design of the study. We did not measure the win–loss record as a time-varying covariate. Thus, a causal effect between changes in win–loss record and timing of injuries cannot be identified in this study. This is also true for penalty minutes.


Team performance, in particular team win–loss record, was associated with injury rates in competitive youth ice hockey players. Teams with better season game performance (winning more than 50% of all games in the season) had fewer injuries and fewer injuries resulting in >7 days time loss than teams with lower win records. Future research should examine the impact of injury prevention strategies in reducing injury risk and improving team performance in competitive youth ice hockey.


This study was funded by the Canadian Institutes of Health Research and Max Bell Foundation. This research was possible due to the support from Hockey Calgary, Hockey Edmonton, Hockey Quebec, Hockey Alberta, Hockey Canada and the Québec Ministry of Education, Leisure and Sport. This research could not have been completed without the support of many Pee Wee players, parents, coaches, team designates, physiotherapists, athletic therapists, student therapists and study physicians. Dr Carolyn Emery is supported by a Population Health Investigator Award from the Alberta Heritage Foundation, a New Investigator Award from the Canadian Institutes of Health Research and a Professorship in Pediatric Rehabilitation from the Department of Pediatrics (Alberta Children's Hospital Foundation), Faculty of Medicine.



  • Competing interests None.

  • Ethics approval Office of Medical Bioethics, University of Calgary.

  • Provenance and peer review Not commissioned; externally peer reviewed.