Abstract
This study examines the association between concussion education and concussion history on reporting behavior in a novel simulated game experience. High school football players (n = 21,213) across 412 high schools located in North Carolina completed demographic questionnaires prior to watching a first-person, 12-min video in which a simulated concussion occurred, followed by a decision prompting respondents to continue to play or to report the concussion. Separate logistic regression models assessed the relationship that prior concussion education and history had with concussion reporting behavior within this simulated-game experience. Overall, 1396 (6.58%) participants chose to stay in the game after the simulated concussion. A majority of participants who chose to stay in the game had no previous concussion education (63.8%) and no history of concussion (82%). Athletes with a previous history of concussion significantly reduced odds of seeking medical attention by 35% (OR = 0.65, 95% CI = 0.57,0.76, p < 0.0001), as compared with not having had a previous concussion, after adjusting for gender. However, athletes who reported previous concussion education had 48% increased odds of seeking medical attention (OR = 1.48, 95% CI = 1.32,1.67, p < 0.0001), after adjusting for gender. Concussion-reporting behavior in a first-person simulated in-game experience is higher among individuals with previous concussion education, and lower among individuals with previous concussion history, in-line with previous literature pertaining to reporting behavior following injuries in real games. The use of simulated game experiences offers promise for evaluating concussion reporting behavior and identifying individuals at risk for underreporting concussion symptoms.
Introduction
Over the past decade, there has been increasing attention on concussions in sport, yet 50% of concussions go unreported.1,2 Athletes whose concussions are undiagnosed are at increased risk of negative sequelae, including prolonged symptoms and further injury. 3 Understanding concussion reporting decisions is therefore critical to ensuring athlete health. The Theory of Planned Behavior 4 has been applied to concussion reporting5,6 to describe how concussion reporting behavior is best predicted by a person’s intention to perform that behavior, with concussion reporting intention further influenced by concussion knowledge 7 and previous concussion experience. 8
Given the difficulties associated with measuring concussion reporting behavior prospectively, including variable concussion rates between sports and seasons, as well as the high rate of unreported concussions, no studies to date have examined factors associated with actual reporting behavior. Instead, researchers often rely on previous reporting history or reported intent. Studies have explored the efficacy of concussion education on improving injury reporting intention, with inconsistent results.2,7,9 Rawlins and colleagues found that more knowledge increased symptom reporting intention, 7 most likely due to the mediation of greater recognition of concussion signs and symptoms. 2 However, increased concussion knowledge may actually inform athletes on which symptoms to avoid reporting in order to continue to play. 9
Similarly, previous concussion experience has been shown to decrease an athlete’s intent to report a concussion,8,10,11 as assessed by retrospective reporting behavior,1,7,8 or stated intent to report concussions in the future.2,12 Assessing behavior in real-time should provide insights into decision-making surrounding concussions.
The study used a simulated game environment as a proxy to assess concussion reporting behavior in real-time. Participants experienced a simulated concussion in a first-person video, followed by a choice to either continue to play, or report the concussion. The primary aim of this study was to examine the association between concussion education, previous concussion experience, and simulated game concussion reporting behavior.
Methods
Study design
Participants
In January of 2020, the North Carolina High School Athletic Association (NCHSAA) Board of Directors passed regulations mandating that schools require parents and students to view the CrashCourse concussion education video prior to each season. All high school students who participated in athletic teams through their schools were required to view this training. All research activities were approved by the Stanford University Institutional Review Board.
Procedures
The CrashCourse concussion education videos were disseminated to NCHSAA-affiliated high school students at the start of each athletic season (fall, winter, and spring) during the 2021–22, 2022–23, and 2023–24 school years. Students were asked to complete the educational program regardless of past participation. Athletic Directors and Coaches were offered multiple options to present the CrashCourse education to their students, including integrating the standalone videos into existing educational programming or administering the content via an anonymous Qualtrics survey, which included the CrashCourse video and optional research surveys. Because the study was determined to be minimal risk and informed consent was not required, informed assent was obtained prior to participating in the research surveys. These pre-test and post-test surveys assessed participants’ concussion knowledge and intention to report concussion symptoms to their coaches. For education tracking purposes, at the end of the concussion education video, participants were linked to a separate Google Form to submit their names. Their responses were not linked to the Qualtrics survey, and the Google Form results were only accessible to coaches at respective schools for internal purposes to determine which students had completed the education. The Google Form results were not utilized by the research study team. In total, 57,653 survey responses were collected. Of these, 4756 athletes (8.25%) did not provide assent and were not included in the study. Only complete surveys were included in analyses (n = 21,213) (Fig. 1).
Methods diagram for statistical analysis. Fifty-seven thousand six hundred and fifty three surveys in total were collected across 2020–2024. Participants may have completed this survey more than once, therefore, only the first survey of each participant was used. Participants who did not assent and who had incomplete surveys were also removed from analyses. Overall, there were 21,213 complete surveys.
Demographics
All participants were asked to complete a demographics survey prior to the educational program. To determine concussion history, participants were asked, “How many concussions have you had in the past?” Answers were then categorically defined as either having no previous concussions (0) or at least 1 (1). Previous concussion education was ascertained from the question, “How many concussion education programs have you had in the past?” with the option to select 1, 2, 3, or 4+. Answers were categorically defined as either having no previous concussion education (0) or having had at least 1 previous concussion education program (1).
Participants were not specifically asked about gender. Rather, participants were asked, “What sport do you participate in?” with options to select: baseball, men’s golf, men’s lacrosse, men’s tennis, men’s track and field, softball, women’s track and field, women’s soccer, and other. For others, participants were then asked, “What other sport do you participate in?” and asked to fill in the blank. If participants selected baseball, men’s golf, men’s lacrosse, men’s tennis, or men’s track and field, gender was classified as male. If participants selected softball, women’s track & field, or women’s soccer, participant’s gender was classified as female. Fill in the blank answers also included women’s lacrosse, field hockey, cheerleading, and rugby. If a participant indicated participation in women’s lacrosse or field hockey, they were classified as female. Although field hockey may be played by males, to the best of our knowledge, only women’s field hockey is a high school sport in North Carolina. If a participant wrote cheerleading or rugby, their gender was left undefined as both females and males may participate in these school sports.
CrashCourse Video Simulation
After completing questions on demographics, participants were asked to complete the CrashCourse video simulation. The CrashCourse approach to concussion education was informed by user-centered formative-design research studies. The 12-min video-based curriculum synthesized first- and third-person experiential perspectives. Specifically, it began by placing learners in a simulated real-life, high-stakes sport scenario (Friday night football game). During the scenario, the learner received a concussion. Through a choose your own adventure approach, the learner decided between two options: Take a Knee or Stay in the Game. Each decision led the learner down a different path to reflect the consequences (both positive and negative). However, the educational content presented in the two pathways was identical, as the learner was also shown what would have happened if the alternate decision had been made. Throughout the educational intervention, the learner was guided by near-peer National Collegiate Athletic Association Division I collegiate football athletes. For the purpose of these analyses, Stay in the Game acted as the reference category.
Statistical analysis
Separate logistic regression models were used to determine the association between concussion history and concussion education on simulated game behavior. In both models, gender was adjusted for, as previous research has found that males often report a greater history of concussion, 13 and therefore may also have greater previous concussion education through individual experience.
Results
For the purpose of this study, if a participant had completed CrashCourse multiple times over the course of the study period, only the first survey was used in analyses. Overall, 1396 (6.58%) participants chose to stay in the game after the simulated concussion. Of these participants, a majority were male (74.2%). Similarly, a majority of participants who chose to stay in the game had no previous concussion education (63.8%) and no history of concussion (82%) (Table 1).
Demographics, including gender, previous concussion education, and previous concussion history stratified by whether the participant did or did not “Stay in the Game” after experiencing the simulated concussion
Athletes with a prior history of concussion had significantly reduced odds of seeking medical attention (OR = 0.65, 95% CI = 0.57, 0.76, p < 0.0001) as compared with those who did not have a previous concussion (Table 2). Athletes who reported previous concussion education sought medical attention with increased odds (OR = 1.48, 95% CI = 1.32, 1.67, p < 0.0001) (Table 2). Across both models, being male compared with female reduced the odds of seeking medical attention by approximately 45% (Table 2).
Odds ratios, 95% confidence intervals, and p-values for logistic regression models in which the outcome was simulated game behavior (Take a Knee or Stay in the Game). Model 1, the predictors were concussion history and hender. Model 2, the predictors sere previous concussion education and gender
Discussion
Athletes who reported prior concussions had decreased odds of “seeking medical attention” within the virtual environment after experiencing the simulated in-game concussion, while those who reported previous concussion education had significantly increased odds of “seeking medical attention.” Males, in general, also reported reduced odds of “seeking medical attention,” regardless of their concussion history or previous concussion education, when compared with females. This finding is consonant with other research indicating that men seek medical and psychological help less frequently than women. 14 Playing sports can also reinforce strict adherence to emotional stoicism and self-reliance, sending the message to athletes that help-seeking is not valued or acceptable.15,16
Previous research examining the effects of concussion education2,7,9 on concussion reporting behavior has found inconsistent results, with multiple prior studies showing similar relationships between concussion education and reporting behavior in real environments, as those observed in this study of behavior in a virtual setting.2,7 In addition, our findings of athlete behavior following a simulated concussion also align with previous research showing previous concussion history is associated with decreased odds of intent to report a concussion.8,10
Prior studies have examined associations between concussion education and either intent to report concussion or retrospective assessment of concussion reporting, rather than true reporting behavior itself. However, the relationship between these factors and true reporting behavior has not been assessed, most likely because of the difficult nature of this research. Examining the relationship between reporting intent and behavior would require a longitudinal study with thousands of participants to ensure an adequate incidence of concussion to ascertain relationships between a priori reporting intent and actual reporting behavior. A simulated environment provides a much more efficient way to study relevant behaviors.
Previous research has found that simulation-based learning influences real-life attitudes. 17 When participants experience an immersive virtual reality (IVR), they experience “place illusion” (PI), or, the sense of “being there,” 18 and will respond to the reality as if it were real. 18 Therefore, we believe that the behavior athletes expressed in the virtual game experience is a reasonable proxy for their actual reporting behavior. Although prior concussion education was positively associated with reporting behavior, neither previous concussion experience nor being male showed a significant association. Therefore, male athletes and athletes with previous concussion history may be at increased risk of continuing play while injured. Importantly, as these groups may not disclose their concussion symptoms, educating coaches on early recognition of concussion in their players may aid medical staff by having an additional trained set of eyes looking out for athlete safety. 19
There are limitations to this study. Sport names were used to identify participant gender, rather than self-identified gender. Therefore, some participants may have been misclassified or had an unknown gender, and thus gender differences in the results may be confounded. Furthermore, information regarding substantiating prior concussion diagnosis and prior concussion education relied on self-report, which has inherent limitations. Additional demographic information was not collected due to time constraints of the educational intervention, so relationships between additional factors could not be explored.
Conclusion
Measuring concussion reporting behavior is challenging and video-simulation behavior offers a novel approach to understanding this phenomenon. Utilization of video-simulated behavior may provide a rapid means to assess the effect of different interventions in real-time, while mitigating some of the limitations and biases associated with widely used retrospective methods to assesses behavior retrospectively. Based on the results of this study, extra attention, and perhaps targeted education, should also be given to male athletes and athletes with previous concussion history to reduce concussion underreporting.
Transparency, Rigor, and Reproducibility
This study was approved by the Stanford University, Institutional Review Board (IRB). A total of 21,213 participants completed the demographics questionnaire and video simulation. Demographics information was self-reported by the participant prior to testing. All tests were completed prior to the start of the season, and if a participant completed this survey multiple times over the course of the study-period, only the first survey was used in analyses. All information collected was confidential, and patients had a unique study ID. Data may be provided upon reasonable request. Logistic regression models were used to determine the association between concussion education, concussion history, and concussion reporting behavior within the novel simulated game environment. Having a concussion history as compared with not having a previous concussion significantly reduces the odds of seeking medical attention. Previous concussion education, when controlling for gender, significantly increases the odds of seeking medical attention. All the analyses were performed by N.K. using R.4.3. 20 All authors have agreed to publish the article.
Authors’ Contributions
N.K. conducted formal analysis and led writing of the article. A.B., L.O., C.M.B., J.B., P.S., R.D. P., and D.H.D. all contributed to conceptualization of the study, and assisted with reviewing and editing the article.
Footnotes
Acknowledgments
The authors are deeply grateful to the North Carolina High School Athletic Association (NCHSAA) leadership team for their unwavering support of this research. In particular, the authors thank Commissioner Que Tucker, Ken Brown, and Janna Fonseca for their exceptional dedication and partnership. They also wish to acknowledge Coach Dick Gould for his invaluable insights and thoughtful contributions, which were instrumental in shaping this work.
Author Disclosure Statement
P.S. serves on the Education Advisory Council for USA Football and is an Adjunct Professor in the Department of Pediatrics at the Stanford University School of Medicine, as well as an Adjunct Lecturer in Stanford’s Graduate School of Education. D.H.D. reported personal fees from expert testimony related to TBIs and being a medical advisor and options holder for StataDx outside the submitted work.
Funding Information
There was no funding provided for this research.