Helmet Fit Assessment and Concussion Risk in Youth Ice Hockey Players: A Nested Case-Control Study

A total of 190 volunteers, consisting of Pee Wee (age = 11–12 years), Bantam (13–14 years), and Midget (15–17 years) ice hockey players in Calgary, Alberta, who registered as members of Hockey Calgary during the 2017–2018 season, were recruited to participate in this nested case-control substudy from 1 season of a 5-year longitudinal cohort study called Safe2Play (Principal investigator: C. Emery, Sport Injury Prevention Research Centre, University of Calgary). This study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (Ethics ID: REB14-0348). All participants and their parents or guardians were provided with written instruction and details of the study before they signed the consent form.

Based on a validated method of injury surveillance,³  concussion cases were players with a suspected (injury report form completed by the team safety designate or athletic therapist) concussion, a physician-diagnosed concussion, or both. The 2 control groups were (1) players who did not sustain a concussion but experienced a musculoskeletal injury during the season and (2) players who remained uninjured (ie, by either concussion or musculoskeletal injury) during the season. When any injury occurred, a team designate contacted a member of the research team to schedule a time for the player to have a helmet-fit assessment. At this time, the research team attempted to recruit a player who had remained uninjured to administer a helmet-fit assessment. When possible, players who sustained a concussion were matched with a nonconcussion-injured player, an uninjured player, or both based on age, sex, and level of play. All helmet-fit assessments were conducted within 10 days of injury, players were asked to wear and fasten all parts of their helmet as they normally would before ice time, and all players confirmed that no helmet-fit changes were made in the time between injury and helmet-fit assessment. When a participant had multiple suspected concussions (3 athletes had 2 suspected concussions each during the same season), each occurrence was treated as a new case, and the player was matched with a new control each time due to the opportunity to make helmet-fit changes after the first assessment.

We adapted the helmet-fit criteria from previous guidelines and literature that measured helmet fit in other settings.^11,22,25,26  The 16 criteria encompassed helmet specifications, age (based on manufacturing date and self-reported length of use by player), condition, certification, and fit characteristics of the helmet for each individual evaluated by the research team (Table 1).²³ 

All data were entered into the Research Electronic Data Capture platform hosted at the University of Calgary, Alberta, and analyzed using Stata (version 15; StataCorp).²⁷  Descriptive statistics were presented for participant characteristics, including age, sex, playing age group, helmet specifications, and self-reported items for all players who had their helmet fit assessed. Conditional logistic regression was the primary analysis for assessing the relationship between outcome (the odds of sustaining a concussion) and exposure (≥2 helmet-fit criteria missing) for concussed players matched with both nonconcussion-injured and uninjured players based on age, sex, and level of play. A secondary logistic regression analysis, clustered by team, was used to assess the relationship between the outcome (the odds of sustaining a concussion) and exposure (≥2 helmet-fit criteria missing) for the entire sample while adjusting for age and body-checking policy as possible modifiers or confounders (or both). We defined unexposed individuals as players with <2 helmet-fit criteria not met and exposed individuals as players with ≥2 criteria not met based on the mean (1.49) and median (1.00) of criteria not met for the entire sample. The results of these analyses were presented as odds ratios (ORs) and 95% CIs at a .05 level of significance.

Helmet-fit assessments were performed on 190 players (165 males and 25 females) from 42 Calgary Minor Hockey Association teams, consisting of 3 Pee Wee teams from playing divisions 4, 5, and 6 (all non–body-checking leagues), 19 Bantam teams representing 9 divisions (5 body-checking leagues), 17 Midget teams representing 9 divisions (6 body-checking leagues), and 3 female hockey teams from the Midget and Bantam levels (all non–body-checking leagues; Table 2). The age of the participants was 14.55 ± 1.47 years. Of the athletes who had their helmet fit assessed during the season, 72 sustained concussions, 41 sustained nonconcussion injuries, and 77 were uninjured. For the primary analysis, 54 concussed players were matched with uninjured players, and 25 concussed players were matched with nonconcussion-injured players based on age, sex, and level of play.

After completing helmet-fit assessments, we determined that 177 players (93%) reported their helmet fit as excellent or good, and 182 players (96%) described their helmet comfort as excellent or good (Table 3). The mean of helmet-fit criteria not met (ranging from all criteria met to 6 criteria not met) was 1.49 (95% CI = 1.30, 1.70), and only 27% (32% concussed, 27% nonconcussion-injured, and 22% uninjured) of the athletes wore helmets that met all helmet-fit criteria (Table 3). After comparing the 72 concussed players with the 41 nonconcussion-injured (OR = 0.98 [95% CI = 0.43, 2.24], P = .961) and 72 uninjured (OR = 1.66 [95% CI = 0.90, 3.05], P = .103) players, no significant associations were found between helmet fit and concussion (unexposed = 0 or 1 criteria not met, exposed = >1 criteria not met). Age and body-checking policy did not modify or confound these relationships. The most common helmet-fit criteria not met by players were chin strap not loose (44% of 190 players), helmet cannot be removed without undoing chin strap (19%), crown of helmet is 1–2 fingers above eyebrows (19%), and helmet does not have ‘cage hang' (17%; Table 4). The most common helmet-fit criterion not met in all 3 groups (43% to 46% of players) was chin strap not loose.

When we matched the concussed players with nonconcussion-injured players (n = 25 matched pairs) based on age, sex, and level of play, no significant association was present between players who had ≥2 helmet-fit criteria not met (unexposed = 0 or 1 criteria not met, exposed = ≥2 criteria not met) and the odds of sustaining a concussion (OR = 1.60 [95% CI = 0.52, 4.89], P = .410). However, when concussed players were matched with uninjured players based on age, sex, and level of play, the odds of sustaining a concussion were 2.67 (95% CI = 1.04, 6.81) greater for players who had ≥2 helmet-fit criteria not met (n = 54 matched pairs, P = .040).

The average age of the players' helmets was 2.98 years (95% CI = 2.45, 3.50) for concussed players, 2.94 years (95% CI = 2.37, 3.51) for nonconcussion-injured players, and 3.04 years (95% CI = 2.63, 3.46) for uninjured players. The self-reported average numbers of years that the helmets were used were 1.98 (95% CI = 1.63, 2.32) in concussed, 1.58 (95% CI = 1.19, 1.98) in nonconcussion-injured, and 2.12 (95% CI = 1.81, 2.42) in uninjured players (Table 5). Only 8 of 190 helmets (4 concussed, 1 nonconcussion-injured, 3 uninjured players) were determined to be expired; however, 68 helmets were missing the Hockey Equipment Certification Council (HECC) sticker that displays the expiration date. Of these helmets, 54 were unexpired based on the manufacturing sticker date (<6.5 years old), which left 14 helmet expiration dates undetermined.

The most common brands of helmets and helmet masks were Bauer (124 helmets, 117 masks), CCM (41 helmets, 54 masks), and Reebok (16 helmets, 9 masks). In the 190 helmets assessed, 20 materials were used. Although some helmets included more than 1 material, the most common materials found in helmets were Bauer PORON Performance Cushioning XRD Extreme Impact Protection foam (63 helmets), Bauer Suspend-Tech liner (28 helmets), and CCM Impact Pods with a Rotational Energy Dampening System (14 helmets). All players wore a full facemask (187 full cages and 3 full plastic shields).

The purpose of our study was to determine the association between helmet fit and the odds of sustaining a concussion in youth ice hockey players. We demonstrated (1) a significant association between inadequate helmet fit and players who sustained concussions when we compared concussed players with uninjured players using a conditional logistic regression analysis matched for age, sex, and level of play; (2) no association between player helmet fit and the odds of sustaining a concussion when comparing concussed players with nonconcussion-injured or uninjured players using an unmatched logistic regression analysis for the entire sample; and (3) that the most common helmet-fit criterion not met by players involved the chin strap being too loose.

Most players (93%, 177 of 190) reported that the fit and comfort of their helmet was excellent or good, yet 78 helmets had ≥2 criteria that were not met, and 19 helmets had ≥4 criteria that were not met. This suggests that the players' perceptions of helmet fit differed from those of the helmet assessors using these helmet-fit criteria.

Given the limited research on helmet fit in ice hockey players, we compared our findings with those in sports such as football and lacrosse. McGuine and Nass¹⁹  reported that 5% of high school football players' helmets had 1 fitting error and 62% had multiple fitting errors. Similarly, 62% of 67 helmets failed to meet all criteria for youth lacrosse and ice hockey players.²²  Although the criteria were obviously different, our results were consistent, with 73% (139 of 190) of the helmets having ≥1 helmet-fit criterion not met. The mean helmet-fit criteria not met were 1.49 (median = 1). Because most helmets did not meet all criteria, players were considered exposed to inadequate helmet fit if ≥2 helmet-fit criteria were not met. Future researchers should strive for a larger sample size to examine different approaches (eg, different numbers of criteria not met, individual criteria, criteria weighted differently) when assessing the relationship between helmet fit and players who sustained concussions.

Based on interviews with emergency department patients, Romanow et al¹²  reported that inadequate helmet fit and position could affect the risk of head and facial injuries in cyclists. In American high school football players, McGuine et al²⁰  suggested that incorrect helmet fit could be associated with a risk of concussion, and Greenhill et al²¹  noted that inadequate helmet fit in football players increased concussion severity based on the number and length of symptoms. However, no investigators have studied youth ice hockey players. We could not address previous concussion and injury severity, among other potentially relevant covariates, but were able to look at age, sex, and level of play in the matched analysis. In this scenario, inadequate helmet fit appeared to be associated with players who sustained concussions when we compared concussed and uninjured players (OR = 2.67). The direction and magnitude of this effect was similar to that in a recent nested case-control study²⁸  of mouthguard use in a youth ice hockey population (OR = 2.78). When considering the entire sample to compare helmet fit in concussed players with helmet fit in nonconcussed-injured and uninjured players, we identified no association. Age and body-checking policy were assessed as possible modifiers and confounders, but the limited and different sample sizes between groups did not allow for the adjustment of other potentially relevant covariates (eg, sex, level of play, previous concussion, injury severity).

The HECC sticker on the helmet specifies when the helmet expires (6.5 years after the manufacturing date). Only 8 of the 190 helmets were expired; however, 14 helmets were missing the manufacturer's date and the HECC sticker and date, making it impossible to determine the expiration date for those helmets. Of the expired helmets, 4 were worn by concussed players, 1 by a nonconcussion-injured player, and 3 by uninjured players. An investigation of the effect of bicycle helmet age on impact performance showed limited differences in helmets ranging from new to 26 years old.²⁹  Another group³⁰  determined that most ice hockey helmet impact-attenuation characteristics were maintained after 10 years in storage, yet no researchers have examined the effect of helmet age or expiration on the risk of injury in ice hockey or other sports.

In contrast to Patton et al,²³  who observed no difference between actual and estimated helmet ages, 69% of our players stated that they had worn their helmets for <3 years, even though the manufacturing stickers indicated that only 51% (78 of 152) of helmets were <3 years old. This suggested 2 possible scenarios: (1) players started using their helmets a considerable length of time after the manufacturing date or (2) players were not aware of the age of their helmets. This difference appeared to translate to the entire sample in that the proportions were similar for both helmet age (50% to 55% <3 years old) and reported length of time helmets were worn (79% to 86% worn for <3 years) for all 3 groups.

We are among the first to consider the association between helmet fit and players who sustained concussions in youth ice hockey; however, the limitations of our work are worth noting. The different results when analyzing the entire sample compared with only the participants capable of being matched raises concern for selection bias. A larger sample size would help avoid the limitations surrounding the lack of potential covariates that could be considered when we analyzed these relationships (eg, player position, game situation, concussion severity, concussion history). The size of the sample also restricted a wide range of possibilities for examining the most appropriate threshold for categorizing exposed and unexposed individuals. Also, a larger sample size would allow for a sensitivity analysis to investigate this outcome specifically to ensure that the criteria do not overlap too much and that the most accurate threshold is selected for determining exposed individuals. Second, we were able to collect data in a timely manner from all player helmets and all players with an injury who contacted the Sport Injury Prevention Research Centre, but teams that were less responsive restricted some of the data that could be collected. Lastly, considerably fewer players were in the female and Pee Wee groups. Therefore, the generalizability of our results to players in those categories may be limited.

Although more than 93% (177 of 190) of players reported that their helmet fit was excellent or good, only 27% (51 of 190) of players met all helmet-fit criteria. No association was found between helmet fit and players who sustained concussions among the entire sample, but inadequate helmet fit resulted in greater odds of concussion when we compared concussed and uninjured players who were matched on age, sex, and level of play. Despite these different results, inadequate helmet fit may affect the odds of sustaining a concussion in youth ice hockey. A larger sample size is needed in future research to evaluate these helmet-fit criteria with certainty, confirm this relationship, and inform helmet-fit recommendations and potential policy changes in youth ice hockey.

The Sport Injury Prevention Research Centre is one of the International Research Centres for Prevention of Injury and Protection of Athlete Health supported by the International Olympic Committee. We acknowledge the funding from Canadian Institutes of Health Research, Alberta Innovates Health Solutions, Hotchkiss Brain Institute, and Alberta Children's Hospital Foundation. We also recognize Hockey Canada, Hockey Calgary, and all of the involved coaches, players, and parents for their time and support in completing this research project.