Injury Report: Undisclosed Health Events in the National Hockey League and Their Impact on Player Health Research

Abstract

Background:

Teams in the National Hockey League (NHL) are mandated to publicly disclose health events (inclusive of injury and illness) that result in time loss. If disclosure is deemed to negatively impact a player’s physical well-being, teams can withhold details from the public and report the event as undisclosed. As a result, research utilizing publicly available NHL health data may be underreporting specific health events of interest.

Purposes:

To (1) identify the incidence and severity of undisclosed health events that result in time loss in the NHL, (2) identify the magnitude of potential bias introduced by the exclusion of undisclosed health events, and (3) identify relevant factors that may influence a team’s decision to disclose health events.

Study Design:

Descriptive epidemiology study.

Methods:

Using a retrospective cohort inclusive of NHL players from 2016 to 2023, public access data related to time loss due to injury or illness were collected. Outcome measures included incidence and severity of undisclosed health events, and count/incidence of injury and illness categories with undisclosed health events reclassified. Statistical significance was tested using Poisson and negative binomial regression.

Results:

Of the 5397 publicly reported health events that resulted in time loss during the NHL regular season across 2016 and 2023, 12.10% (653/5397) were due to undisclosed health events. Specific injury and illness classifications were found to be underreported when adjusting for undisclosed health events (P < .001). Undisclosed health events were more likely to be reported in the last fifth of the season (P < .001) and more common for higher profile players (P < .001), regardless of the stage of the season. The severity of undisclosed health events was highly variable across seasons and did not show any significant trends over season duration.

Conclusion:

Studies utilizing publicly obtained NHL health data are at risk of underreporting occurrence and misreporting the severity if they do not adjust or account for health event disclosure in their analyses.

Keywords

bias disclosure policies epidemiology ice hockey injury National Hockey League

League-specific injury surveillance data—overseen by team and league-specific medical staff—is often considered the gold standard for sports medicine research focused on professional athlete populations.²¹ While these datasets can be made accessible for third-party researchers, strict review processes can take significant time and often require approval by both the respective League and Players Association, acting as notable barriers to accessing said data.²⁴ Given the nature of professional ice hockey, where athletes are exposed to a higher risk of injury relative to lower levels of ice hockey,² public interest in player health, and the robust data management infrastructures available, health event data (inclusive of injury and illness) from internet-based public sources (publicly obtained data [POD]) are an attractive option for sports-medicine-based research in professional athletes.^6,13,37 Peer-reviewed research can serve to fulfill a core role of professional sports by enhancing our understanding of athlete health and safety.⁷ These findings serve to benefit nonprofessional sports leagues, which may not have access to the same resources as professional leagues, in addition to promoting greater engagement among the general public.⁷ For example, a prospective study of concussions in National Hockey League (NHL) players was used to inform policy changes to reduce injury risk at the Pee Wee level by eliminating body checking (a common mechanism of injury for concussions).^5,15

Mandates at various levels (ie, league, state/provincial, or federal) often dictate public disclosure policies for injuries that occur to professional athletes while performing structured team activities (ie, games and practice).²⁷ The National Football League’s disclosure policy mandates the reporting of any injury that results in a player’s unavailability or limited availability to the club.³² In the National Basketball Association, injury designation detailing participation status and the specific injury is required by 5 PM local time on the day before a game.³¹ In their respective collective bargaining agreements (CBA), Major League Baseball and the NHL state that general details—such as nature, prognosis, and anticipated length of recovery from an injury—may be disclosed to the public.^26,33 Injury transparency is known to impact betting lines, and despite the NHL’s current CBA predating the legalization of sports betting in the United States,¹⁷ and similar legislation in Canada,¹⁸ there have been no changes to the NHL’s public injury disclosure policy.³⁸ In addition, NHL teams are given latitude regarding disclosure if they feel that disclosure may negatively impact a player’s physical well-being when they return to play.³⁸ In doing so, in compliance with reporting mandates, the specific nature of an injury or illness is withheld from the public and reported as undisclosed.

Misinformation bias refers to a distortion in the measure of effect caused by a lack of accurate measurements of exposure or outcome (ie, injury or illness) status.^1,25 By excluding, not adjusting, or not acknowledging the unknowns surrounding the presence undisclosed health events, studies utilizing publicly obtained NHL health data may be underestimating or misrepresenting the true magnitude of a specific injury/illness’s occurrence or severity.^13,37 Given that event disclosure is up to the team’s discretion, it is also important to consider any relevant covariates that may systematically influence the bias introduced through the presence of undisclosed health events in public NHL health data. For example, higher-profile players have biased coverage in the media over non-high-profile players, resulting in greater availability of performance information to the public.¹⁹

The burden (inclusive of occurrence and severity) of undisclosed health events on time loss in the NHL has yet to be empirically investigated. In addition, the impact of excluding undisclosed health events in epidemiological studies utilizing publicly available NHL health data is largely unknown. The purpose of this study was to (1) identify the incidence and severity of undisclosed health events that result in time loss in the NHL, (2) identify the magnitude of potential bias introduced by the exclusion of undisclosed health events, and (3) identify relevant factors that may influence a team’s decision to disclose health events.

Methods

Data Extraction and Health Event Definition

Data were extracted from multiple online sources. The Pro Sports Transaction Archive (PSTA; https://www.prosportstransactions.com), a commonly used data repository for sports medicine research that utilizes POD, was accessed for health data; basic descriptive player information and schedule data were extracted at www.NHL.com, and player-hour exposure data were extracted at www.moneypuck.com (online data repository for NHL performance and participation data). Extraction was conducted through reproducible data scraping methods,²² written by A.M.P. in Python Version 3.10.4 (Python Software Foundation) with the Pandas and BeautifulSoup libraries. Only pages under the Hockey and Missed games due to injuries filters were extracted between September 1, 2016, and September 1, 2024, from the PSTA (Figure 1). Health events were classified into upper body, lower body, head, or core injuries in addition to illness. An undisclosed health event was defined as any event that contained the term undisclosed or non-regional specific descriptive terms such as “general soreness,” “body maintenance,” or “strained muscle.” Any health events that did not directly result from gameplay (ie, “blood clots” or “heart arrhythmia”) or were related to conditioning were coded as nongame-related. The initial dataset contained some data that did not meet the inclusion criteria. Events not associated with player health (ie, “trade,” “player not with team,” and “General Manager”) were excluded from the study (5 events). Any health event that was logged for a player who did not have any playing time in the specified season range was further excluded from the analysis, as the health event could not have occurred during regular-season gameplay (5 events). To ensure replicable and valid methods and results, all code for data collection has been included as an Appendix.⁹ Before filtering, the database was tested for reliability by 2 external reviewers (L.T. and J.R.) by testing independent random samples of 100 events (allocated by a random number generator).⁸ Each entry was evaluated for (1) the validity of the player and date entries and (2) the appropriateness of injury classification. Given the independent sampling between reviewers, inter-rater agreement was not calculated.

Figure 1.

Extraction, filtering, and inclusion criteria for data acquisition and handling for analysis. PSTA, Pro Sports Transaction Archive.

Statistical Analyses

Exposure was calculated using player-hours with data obtained from moneypuck.com.³⁰ Incidence rate, reported per 1000 player-hours, was calculated as the quotient of the number of health events (numerator) and player-hours for each team on the specified game date. Of the 8 regular seasons included, season length and start/end dates varied significantly, largely due to disruptions related to the coronavirus disease 2019 pandemic in the 2019-2020 season. To account for variable season lengths across the 2016 and 2023 seasons, the seasonal incidence was calculated using a standardized season length, with each season divided into 10 equal sections, as defined by the first and last game of the respective season. The severity was calculated as the number of days between the date of the reported health event and the date of return. If a player did not return to play that season, the return date was set to 1 day after the last game of the regular season.

Several strategic factors may influence a team’s decision to disclose a health event. Teams may be less inclined to disclose the reasoning behind time loss at later stages of the season or if they are involved in the playoffs, so as not to provide a potential competitive advantage to an opponent. Reporting of public injuries is often biased toward higher-profile players. Therefore, a profile was assigned to a player based on the mean time on ice (TOI) per game for a given season (mean ice time among 2016-2023: 17.55 minutes). In addition, disclosure policies were modified during the modified 2019-2020 season playoffs,⁴ indicating a potential change in a team’s interpretation of the disclosure policy in post-pandemic seasons. Covariates included in the analysis were as follows: (1) >80% of season completion (1 = yes; else = 0); (2) pre- or post-pandemic season (1 = post; else = 0); (3) made the playoffs in a given season (1 = yes; else = 0); and (5) high-profile player (1= >17.55 minutes; else = 0). Normality was checked using the Shapiro-Wilk test for normality, and the dataset was not normally distributed (W = 0.38; P < .001). Overdispersion was checked for both incidence (5.52) and severity (20.79) outcomes, which warranted the usage of Negative Binomial regression for both models.

Reclassification Analysis

The impact of misclassification errors in epidemiological studies is rarely explicitly addressed.¹⁶ A deterministic sensitivity analysis was implemented to account for potential misclassification bias due to health events reported as undisclosed. Specifically, a series of scenarios was simulated in which 0% to 100% of undisclosed health events were reclassified into existing health event categories (head, core, upper body, lower body, and illness). The chosen approach assumes misclassification is systematic (as teams intentionally decide it) and explores the range of possible outcome values under plausible reclassification conditions. No random sampling or probabilistic distributions were applied; thus, the analysis represents the upper and lower bounds of possible bias, rather than quantifying the probability of specific outcomes. Statistical significance testing under the null hypothesis of there being no difference between baseline incidence (0% of reclassification) and any percentage of reclassification was conducted. Normality was checked using the Shapiro-Wilk test for normality, and the dataset was not normally distributed (W = 0.56; P < .001). Overdispersion was checked for incidence (0.63), which warranted the usage of Poisson regression. A Poisson regression mixed-effect model was fit with incidence as the outcome variable, reclassification percent as the predictor variable, and injury category as a covariate. Season was also included in the model as a random effect to account for inter-seasonal variability. Significance for all statistical analyses was set at P < .05.

Results

There were 5397 reported health events that resulted in time loss during the NHL regular season across the 2016 and 2023 seasons (Table 1). Also, 12.10% (653/5397) of all regular-season health events resulting in time loss in that window were undisclosed. For any given season across 2016-2023, undisclosed health events were more prevalent than head, core, and non-game-related injuries (Table 1). During reliability testing, reviewer No. 1 found 94% of entries valid for player/date and 95% for health event classification. Reviewer No. 2 found 90% and 95%, respectively. Although inter-rater reliability could not be calculated due to nonoverlapping samples, the consistency of proportions supports the overall validity of the dataset. The most common reason for lack of consistency between sources was a mismatch in the date of reporting (within 3 days of the injury/illness occurring).

Table 1

Data on Health Event Frequency for Each Season and Health Event Classification ^a

	Season
	2016	2017	2018	2019	2020	2021	2022	2023
Total No. of health events	679	658	659	539	627	1119	453	663
Upper body	219 (32.3)	204 (31)	200 (30.4)	140 (26)	175 (27.9)	239 (21.4)	126 (27.8)	208 (31.4)
Lower body	248 (36.5)	207 (31.5)	212 (32.2)	187 (34.7)	188 (30)	227 (20.3)	147 (32.5)	182 (27.5)
Head	30 (4.4)	45 (6.8)	50 (7.6)	35 (6.5)	26 (4.2)	35 (3.1)	14 (3.1)	29 (4.4)
Core	16 (2.4)	16 (2.4)	12 (2.9)	17 (3.2)	11 (1.8)	11 (1)	5 (1.1)	7 (1.1)
Nongame-related	0 (0)	2 (0.3)	0 (0)	0 (0)	1 (0.2)	2 (0.2)	0 (0)	0 (0)
Illness	105 (15.5)	95 (14.4)	115 (17.5)	101 (18.7)	127 (20.3)	485 (43.3)	82 (18.1)	154 (23.2)
Undisclosed	61 (9)	89 (13.5)	63 (9.6)	59 (11)	99 (15.8)	120 (10.7)	79 (17.4)	83 (12.5)

Data are presented as n (%). TOI reported the total ice time for all players in the season per 1000 player-hours. TOI, time on ice.

Incidence

There was no significant difference in the incidence of undisclosed health events among seasons (P = .172). Undisclosed health events were more frequent after 80% of the regular NHL season (~65 games played of an 82-game season) (incidence rate ratio [IRR], 1.69 [95% CI, 1.27-2.25]; P < .001) compared with before the 80% mark of the regular season (Figure 2). The occurrence of undisclosed health events was 94% higher in high-profile players compared with non-high-profile players (IRR, 1.94 [95% CI, 1.48-2.56]; P < .001). No significant difference was observed between playoff and nonplayoff teams in the reporting of undisclosed health events (IRR, 1 [95% CI, 0.76-1.32]; P = .995). In addition, no significant difference was found in the reporting of undisclosed health events between pre- and postpandemic seasons (IRR, 1.27 [95% CI, 0.96-1.67]; P = 0.09).

Figure 2.

Seasonal trends for reporting undisclosed health events across 2016-2023. The mean (black dashed line) and 95% CI (gray shaded area) for each season.

Severity

The mean total time loss (mean ± SD) due to undisclosed health events across the NHL regular season was 1640 ± 538 days missed. The 2022 season had the least time lost, with 965 total days missed, while the 2021 season had the most, with 2632 total days missed (Figure 3). For any given season, undisclosed health events account for more time loss than head, core, and non-game-related injuries (Table 2). The severity of undisclosed injuries was not significantly affected by season (P = .421), profile of the player (P = .829), qualification for playoffs in a given season (P = .884), or post-pandemic seasons (P = .475).

Figure 3.

Cumulative days missed due to undisclosed health events over the season duration from 2016 to 2023.

Table 2

Severity for Each Season and Health Event Classification Shown via Total Days Lost and Relative Distribution for Total Days Lost in the Given Season ^a

	Season
	2016	2017	2018	2019	2020	2021	2022	2023
Total No. of days lost	17702	17893	18643	13976	15358	38227	8123	14206
Upper body	6140 (34.7)	6367 (35.6)	6739 (36.1)	3901 (27.9)	4126 (26.9)	9172 (24)	3151 (38.8)	5409 (38.1)
Lower body	7527 (42.5)	6772 (37.8)	7288 (39.1)	6151 (44)	4393 (28.6)	8020 (21)	3140 (38.7)	4797 (33.8)
Head	1032 (5.8)	1217 (6.8)	1826 (9.8)	728 (5.2)	601 (3.9)	1105 (2.9)	340 (4.2)	776 (5.5)
Core	1084 (6.1)	579 (3.2)	455 (2.4)	911 (6.5)	367 (2.4)	264 (0.7)	35 (0.4)	442 (3.1)
Nongame-related	0 (0)	172 (1)	0 (0)	0 (0)	4 (0)	6 (0)	0 (0)	0 (0)
Illness	813 (4.6)	919 (5.1)	849 (4.6)	951 (6.8)	4128 (26.9)	17028 (44.5)	492 (6.1)	795 (5.6)
Undisclosed	1106 (6.2)	1867 (10.4)	1486 (8)	1334 (9.5)	1739 (11.3)	2632 (6.9)	965 (11.9)	1987 (14)

Data are presented as n (%).

Reclassification Analysis

When adjusting for season as a random effect, baseline incidence (0% of reclassification) for upper body injuries was 12.60 injuries per 1000 player-hours (95% CI,12.29 to 12.92), lower body was 13.50 injuries per 1000 player-hours (95% CI, 13.18 to 13.8), head was 1.48 injuries per 1000 player-hours (95% CI, 1.17 to 1.80), core was 0.86 injuries per 1000 player-hours (95% CI, –1.13 to 2.85), and illness was 10.17 illnesses per 1000 player-hours (95% CI, 9.86 to 10.49) (Figure 4). On average, the incidence of a given injury category increased by 0.059 health events per 1000 player-hours for every 1% increase in reclassification (P < .001) (Figure 4). The incidence for every health event category was statistically different after reclassification relative to baseline values (ie, 0% reclassification) (P < .001). The random effect for the season was 8.06 ± 2.84. The count data of each event category after reclassification can be seen in Table 3 for every 10% step in reclassification from 0% to 100%.

Figure 4.

Incidence of each health event category at a given percentage of undisclosed health events reclassified as each injury category. The mean (black line) and 95% CI (gray shaded area) for each season.

Table 3

Data for Each Injury Category at a Given Level of Reclassification of Undisclosed Injuries ^a

Percent of Reclassification	Upper Body	Lower Body	Head	Core	Illness
0	189 ± 39	200 ± 31	33 ± 11	12 ± 4	158 ± 134
10	197 ± 40	208 ± 31	41 ± 11	20 ± 4	166 ± 136
20	205 ± 41	216 ± 31	49 ± 11	28 ± 5	174 ± 137
30	213 ± 42	224 ± 32	57 ± 12	36 ± 6	182 ± 139
40	222 ± 43	232 ± 32	66 ± 13	45 ± 8	191 ± 140
50	230 ± 44	241 ± 33	74 ± 14	53 ± 10	199 ± 142
60	238 ± 46	249 ± 34	82 ± 16	61 ± 12	207 ± 144
70	246 ± 47	257 ± 34	90 ± 17	69 ± 14	215 ± 145
80	254 ± 48	265 ± 35	98 ± 19	77 ± 16	223 ± 147
90	262 ± 50	273 ± 37	106 ± 21	85 ± 18	231 ± 149
100	270 ± 51	281 ± 38	115 ± 23	94 ± 20	240 ± 150

Seasonal mean ± seasonal standard deviation. For practical utility, values have been rounded to the nearest whole health event.

Discussion

This study aimed to investigate the burden of undisclosed health events that resulted in time loss in the NHL and relevant covariates that may affect event disclosure. This study also aimed to quantify the magnitude of potential bias introduced by the exclusion of undisclosed health events. Undisclosed health events were more likely to be reported in the final fifth of the season and more common for high-profile players, regardless of the stage of the season. The severity of undisclosed health events was highly variable across seasons and showed no significant trends over season duration. The exclusion of undisclosed health events from NHL player health research may significantly underreport the incidence of injury and illness.

In compliance with reporting mandates, teams can withhold injury or illness details from the public if they deem it in the best interest of the player’s physical well-being.³⁸ Wording is left general and should be treated as a dynamic definition to reflect the sensitive nature of health data relating to an individual’s physical well-being. Professional sports teams are notoriously secretive about performance and health data to avoid providing opponents with favorable (ie, winning) scenarios. The number of undisclosed injuries tended to increase at the tail end of the season (> ~65/82 games played), regardless of a team’s participation in the playoffs. Teams may be less likely to disclose specific injury details to prevent disclosure of information to opponents that may create leverage in the final stages of playoff qualification or for future playoff matchups. If teams are no longer in playoff contention, they may be less inclined to disclose health events (in detail) that result in time loss, as it is no longer of public interest. Player availability is relevant to public interest when it can potentially influence results, and when the relevance of results declines, interest in availability does as well.³⁹ Teams tended not to disclose the injury status of high-profile players compared with non-high-profile players. The lack of disclosure for high-profile players may be due to a potential fear of being at a higher risk of opponents targeting them (over non-high-profile players),³⁸ thereby creating competitive leverage by diminishing the impact of key performers.

Through the reclassification analysis, the random effect of season indicated a large interseasonal variance that could not be explained by reclassification or health event category.³⁴ The high variance in time loss among seasons due to undisclosed events is likely reflective of the variable nature of what defines an undisclosed health event. For example, an undisclosed injury could potentially encapsulate a day-to-day injury, such as a low-grade quad contusion (1 to 2 days lost) or an anterior cruciate ligament rupture (>8 months lost).²³ Positing the presence of a potential unmeasured confounder that may reduce the interseasonal variance during reclassification was beyond the scope of the present study.¹² It is recommended that future studies investigating injury/illness trends in the NHL cover multiple seasons to account for any interseasonal variation.

A distrust exists between third parties conducting sports medicine research with POD, and end-users who work with the equivalent gold standard, such as the NHL’s athlete health management system (AHMS),²⁰ reducing the overall clinical utility of said research. Despite the distrust, a relative paucity exists in literature utilizing gold standard NHL injury data, while peer-reviewed studies utilizing POD are increasing in popularity.⁶ Computing and artificial intelligence advancements will continue to break down previous barriers to data availability, which will continue to accelerate the popularity of POD in sports medicine research.³⁶ At the moment, there is no standardized method for handling undisclosed health events, although the exclusion of an undisclosed injury by cross-referencing with another media source has been used.^10,35 Reliability testing by 2 external reviewers showed that 95% of randomly selected records were confirmed in other public reports, with the most common mismatch corresponding to reporting of the injury/illness in the dataset within 3 days between sources. Injury checks generally occur within 24 hours after games, and reporting to the media is likely done once formal medical assessments and diagnoses are completed, which may take upwards of 3 days (if not longer) from the date of actual injury. These factors likely contribute to the variable severity observed across seasons for a given injury category in publicly obtained player health data (Table 2).

Future sports medicine research utilizing publicly available NHL health data should take appropriate steps to ensure the influence of a potential misclassification bias due to a lack of specificity in reported health event classification is minimized, or at the very least, acknowledged. For example, a differential misclassification exists that disproportionately affects high-profile players over non-high-profile players. To account for potential misclassification biases, end-users can incorporate previously reported NHL injury proportions to inform reclassification or to be used within Bayesian models to predict likely injury distributions.¹⁴ In a study of injury and illness trends across 6 NHL seasons, head and neck injuries from the AHMS accounted for ~20% of all time loss, while upper body, lower body, and core injuries accounted for 24%, 45% and 11%, respectively.²⁹ Assuming undisclosed health events are a reflective sample of the above distribution for all health events resulting in time loss, for the average NHL season, all health event categories in the present study would be underestimated by >10% before reclassification. End-users utilizing POD sources should always report reclassified and raw estimates in tandem to ensure readers understand the range of plausible outcomes and uncertainty due to misclassification. Standardized reporting of methods and acknowledgement of biases that are introduced using POD in NHL player health research, such as those mentioned above, are imperative to provide clinical utility at the NHL and lower levels of hockey. The responsibility to improve knowledge translation to inform best practice does not solely lie with the researcher. Research will always involve some degree of uncertainty, and reviewers are responsible for critique if the unknown outweighs the value of the known in the analysis. Moreover, given that 90.9% of peer-reviewed NHL orthopaedic research utilizes publicly obtained data,⁶ improvements to AHMS data accessibility would ultimately render much of the listed recommendations for researchers redundant.

A notable limitation of this study is the health event categorization. Health events were categorized into 6 categories, and 5 were used in the reclassification analysis (with non-game-related health issues removed due to extremely rare occurrences). A low level of granularity, such as this, provides little clinical significance, whereas specific injuries are vastly more applicable to sports medicine practitioners.²⁸ Because of the lack of consistency in the granularity of reported health events due to the flexibility in the NHL’s injury disclosure policy, it is difficult to say with any certainty that specific injury categories, inclusive of those presented in these studies, are true reflections of the actual burden of injury on time loss. For instance, the same injury for 4 individual players could be reported as an upper body injury, shoulder injury, acromioclavicular sprain injury, or undisclosed. Studies that focus on specific injuries, such as eye³⁵ or craniomaxillofacial,¹⁰ should be aware of the potential underreporting due to misclassification/exclusion of undisclosed injuries in the filtering stages of data collection (of which eye and craniomaxillofacial may fall into). While there may be unknowns in the reporting of injury or illness to the public, health events may not be reported to medical staff in the first place.¹¹ Athletes regularly play through health-related issues that may impact their quality of performance or reduce their ability to participate in team activities fully.⁴⁰ They may contribute to reports with undisclosed injuries and illness, as they may intentionally downplay the severity of health events to avoid extending potential time loss through mandatory injury protocols such as the concussion protocol.³ A notable omission from the raw data set was any event with the term mental or related to mental health/wellness. While it is a distinct possibility that these issues are not present in NHL players, it is far more likely that mental health issues are either labelled by medical staff as illness when reported to the public or not reported by players at all (resulting in no time loss). It is likely that these factors further influence the true burden of injuries and illness in professional hockey, regardless of whether the data are accessed from the NHL’s injury surveillance system or publicly obtained sources.

Conclusion

Undisclosed health events were more likely to be reported in the last 20% of the season and were more common for high-profile players than non-high-profile players, regardless of the stage of the season. The severity of undisclosed health events was highly variable across seasons and showed no significant trends over season duration. Exclusion of undisclosed health events has the potential to significantly underreport the incidence of injury and illness, especially if reporting the incidence or severity with a higher granularity, like upper body, lower body, head, core, and or illness. By quantifying the potential information bias introduced by the exclusion of undisclosed health events, future sports medicine studies utilizing publicly obtained NHL injury data can adjust their analyses accordingly to provide greater clinical utility in their outcomes.

Footnotes

Final revision submitted July 9, 2025; accepted August 19, 2025.

The authors have declared that there are no conflicts of interest in the authorship and publication of this contribution. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.

Ethical approval was not sought for the present study.

References

Alexander

Lopes

Ricchetti-Masterson

Yeatts

KB.

Sources of systematic error or bias: information bias. Eric Notebook. 2015;2(14):1-5.

Anderson

Melugin

Stuart

MJ.

Epidemiology of injuries in ice hockey. Sports Health. 2019;11(6):514-519.

Andrews

Jildeh

Abbas

, et al. Concussions in the National Hockey League: analysis of incidence, return to play, and performance. Orthop J Sports Med. 2022;10(1):23259671211052069.

Associated Press. NHL gags teams on injury disclosure for player privacy. AP News. July 11, 2020. Accessed September 2024. https://apnews.com/shh-nhl-gags-teams-on-injury-disclosure-for-player-privacy-0baca274d14d9b21ad799bd31bd570cc

Benson

Meeuwisse

Rizos

Kang

Burke

CJ.

A prospective study of concussions among National Hockey League players during regular season games: the NHL-NHLPA Concussion Program. CMAJ. 2011;183(8):905-911. doi:10.1503/cmaj.092190

Berube

Lopez

Trofa

Popkin

CA.

A systematic review of the orthopedic literature involving National Hockey League players. Open Access J Sports Med. 2020;11:145-160. doi:10.2147/OAJSM.S263260

Brocherie

Beard

All alone we go faster, together we go further: The necessary evolution of professional and elite sporting environment to bridge the gap between research and practice. Front Sports Act Living. 2021;2:631147. doi:10.3389/fspor.2020.631147

Bullock

Murray

Vaughan

Kluzek

Temporal trends in incidence of time-loss injuries in four male professional North American sports over 13 seasons. Sci Rep. 2021;11(1):8278.

Bullock

Ward

Kluzek

, et al. Paving the way for greater open science in sports and exercise medicine: navigating the barriers to adopting open and accessible data practices. Br J Sports Med. 2024;58(6):293-295. doi:10.1136/bjsports-2023-107225

10.

Cimba

Labe

Dazen

, et al. Do mandatory face shields reduce the number of craniomaxillofacial injuries in the National Hockey League? J Oral Maxillofac Surg. 2021;79(6):1313-1318.

11.

Clarsen

Bahr

Matching the choice of injury/illness definition to study setting, purpose, and design: One size does not fit all!

Br J Sports Med. 2014;48(7):510-512.

12.

Delaney

Seeger

JD.

Sensitivity analysis. In: Developing a Protocol for Observational Comparative Effectiveness Research: A User's Guide. Agency for Healthcare Research and Quality; 2013.

13.

Ehlen

Costello

Wagner

, et al. The effects of modified game schedules on injury rates in the National Hockey League (NHL). Cureus. 2023;15(10):e46898.

14.

Elliott

Margulies

Maltese

Arbogast

KB.

Accounting for sampling variability, injury under-reporting, and sensor error in concussion injury risk curves. J Biomech. 2015;48(12):3059-3065. doi:10.1016/j.jbiomech.2015.07.025

15.

Emery

Kang

Shrier

, et al. Risk of injury associated with body checking among youth ice hockey players. JAMA. 2010;303(22):2265-2272. doi:10.1001/jama.2010.755

16.

Fox

MacLehose

Lash

. Applying quantitative bias analysis to epidemiologic data. Springer; 2021:105-139.

17.

Holden

JT.

Prohibitive failure: The demise of the ban on sports betting. Ga St UL Rev. 2018;35:329.

18.

Holden

JT.

The emergence of single-game sports betting in Canada. J Legal Aspects Sport. 2022;32:198.

19.

Inclan

Chang

Mack

, et al. Validity of research based on public data in sports medicine: a quantitative assessment of anterior cruciate ligament injuries in the National Football League. Am J Sports Med. 2022;50(6):1717-1726.

20.

Inclan

Kuhn

Troyer

, et al. Use of publicly obtained data in sports medicine research: a systematic review and bibliometric analysis. Am J Sports Med. 2024;52(5):1367-1373.

21.

International Olympic Committee Injury and Illness Epidemiology Consensus Group, Bahr

Clarsen

, et al. International Olympic Committee consensus statement: methods for recording and reporting of epidemiological data on injury and illness in sports 2020 (including the STROBE extension for sports injury and illness surveillance [STROBE-SIIS]). Orthop J Sports Med. 2020;8(2):2325967120902908.

22.

Landers

Brusso

Cavanaugh

Collmus

AB.

A primer on theory-driven web scraping: automatic extraction of big data from the Internet for use in psychological research. Psychol Methods. 2016;21(4):475-492. doi:10.1037/met0000081

23.

Longstaffe

Leiter

MacDonald

Anterior cruciate ligament injuries in the National Hockey League: epidemiology and performance impact. Clin J Sport Med. 2020;30(3):224-230.

24.

Mack

Matava

Zeidler

Sills

Solomon

Clinical research in the National Football League: the player scientific and medical research protocol. Curr Sports Med Rep. 2020;19(5):168-174.

25.

Maclure

Schneeweiss

Causation of bias: the episcope. Epidemiology. 2001;12(1):114-122. doi:10.1097/00001648-200101000-00019

26.

Major League Baseball Players Association. 2022-2026 Major League Baseball collective bargaining agreement. Accessed September 2024. https://www.mlbplayers.com/_files/ugd/4d23dc_d6dfc2344d2042de973e37de62484da5.pdf

27.

Makhni

Buza

Byram

Ahmad

CS.

Sports reporting: a comprehensive review of the medical literature regarding North American professional sports. Phys Sportsmed. 2014;42(2):154-162. doi:10.3810/psm.2014.05.2067

28.

Martin

Lieb

Tokish

, et al. Initial versus subsequent injury and illness and temporal trends among professional hockey players. Int J Sports Phys Ther. 2024;19(2):215.

29.

McKay

Tufts

Shaffer

Meeuwisse

WH.

The epidemiology of professional ice hockey injuries: a prospective report of six NHL seasons. Br J Sports Med. 2014;48(1): 57-62.

30.

MoneyPuck. Data. MoneyPuck. Updated October 16, 2024. Accessed September 2024. https://moneypuck.com/data.htm

31.

National Basketball Association. NBA injury report 2023-24 season. NBA Official. Accessed September 2024. 2023. https://official.nba.com/nba-injury-report-2023-24-season/#:~:text=By%205%20p.m.%20local%20time,by%20such%20injury%2C%20illness%20or.

32.

National Football League. NFL injury report policy. NFL Operations. 2017. Accessed September 2024. https://operations.nfl.com/media/2683/2017-nfl-injury-report-policy.pdf

33.

National Hockey League; National Hockey League Players' Association. Collective Bargaining Agreement between National Hockey League and National Hockey League Players' Association. February 15, 2013. Article 34, Section 34.3.

34.

Oberg

Mahoney

. Linear mixed effects models. Methods Mol Biol. 2007;404:213-234.

35.

Pradeep

Arun

Ravipati

, et al. Eye injuries in the National Hockey League from 2010 to 2018: an analysis of injury rates, mechanisms, and the National Hockey League visor policy. Can J Ophthalmol. 2021;56(1):17-23.

36.

Ramkumar

Luu

Haeberle

, et al. Sports medicine and artificial intelligence: a primer. Am J Sports Med. 2022;50(4):1166-1174.

37.

Rosenberg

Baker

Rizzi

Terry

Tjong

VK.

Frequency of overtime and consecutive games and their impact on injury rate in the National Hockey League (NHL). Phys Sportsmed. 2021;49(1):100-105.

38.

Strong

. When it comes to NHL player injury transparency, mum’s still the word. The Canadian Press. April 30, 2024. Accessed September 2024. https://www.thecanadianpressnews.ca/sports/when-it-comes-to-nhl-player-injury-transparency-mums-still-the-word/article_dfd9b3b3-d136-53f2-9ddf-0844f6a416dc.html

39.

Tainsky

Spectator age and periodic changes in game viewership: an application of positivity and life-span theory in the National Football League. J Appl Sport Manage. 2019;11(4):8.

40.

Weinberg

Vernau

Horn

Playing through pain and injury: psychosocial considerations. J Clin Sport Psychol. 2013;7(1):41-59.