Collegiate Student-Athletes Concussion Knowledge and Attitudes: What a Difference a Decade Makes

Abstract

Objective:

The purpose of this study was to assess changes in concussion knowledge and attitudes amongst incoming intercollegiate student-athletes over the course of a decade (2010-2012 vs 2021-2023).

Methods:

There were 592 student-athletes from 2 cohorts (2010 – 2012, 2021 – 2023) who completed the Rosenbaum Concussion Knowledge and Attitudes Survey (ROCKaS) questionnaire which is comprised of a concussion knowledge index (CKI, 0 – 24) and attitude index (CAI, 15- 75) with higher scores reflecting better performance. A three factor ANOVA (Group, Sex, Concussion History) compared performance on the CKI and CAI. Individual questions were compared between groups with a Chi-Square analysis.

Results:

For the CKI, there was a significant main effect for Group (2010-2012: 18.5 ± 2.6, 2021-2023: 19.4 ± 2.5, p<0.001, η² = 0.032). For the CAI, there was also a significant main effect for group (2010-2012: 52.9 ± 6.0, 2021-2023: 62.2 ± 6.5, p<0.001, η² = 0.359).

Conclusions:

The results of this study show a modest increase in concussion knowledge; however, large improvements in concussion attitudes were observed between groups. These results suggest a continued improvement in student-athlete concussion awareness and provides specific areas to continue addressing persistent misconceptions.

INTRODUCTION

Diagnosis of sports related concussions remains an ongoing challenge with diverse presentations affecting multiple systems (e.g., cognitive, motor, vestibular) along with numerous symptoms.(1-5) Thus, health care professionals use multifaceted concussion assessment batteries which are highly successful in accurately diagnosing a concussion.(6-8) However, most concussions do not present with loss of consciousness or obvious overt symptoms,(1) thus the athlete, or a teammate, must self-report in order to initiate the assessment. Unfortunately, concussion underreporting, either not recognizing the potential concussion or intentionally hiding, remains commonplace amongst athletes.(9-12) Thus, concussion education programs have been developed and implemented to increase the likelihood of self-report which may improve patient health outcomes and reduce medico-legal implications.(13-15)

Beginning in 2010, the National Collegiate Athletic Association (NCAA) mandated that all member institutions have a concussion management plan which included student-athletes receiving concussion education.(16) Some evidence supports the use of concussion education programs to improve the knowledge and attitudes which may increase the likelihood of self-reporting;(17) however, others have suggested mandatory concussion education to be an ineffective mandate.(18, 19) While knowledge of common symptoms and misconceptions is important, knowledge alone is not sufficient to determine if a student-athlete will self-report a potential concussion.(20) Indeed, concussion reporting is a complex process which includes both personal and sociological factors.(11, 14, 21) Student-athletes who recall receiving concussion education have a more favorable perceived social norm to concussion care seeking(17) and intent to disclose a suspected concussion is closely linked to organization culture and interpersonal relationships.(17, 22, 23) This is critical to student-athlete health and safety overall as earlier reporting is associated with better health outcomes and quicker return to sports participation.(15, 24)

Numerous approaches have been used to assess athlete’s knowledge and attitudes regarding concussions;(25, 26) however, the Rosenbaum Concussion Knowledge and Attitudes Survey (ROCKaS) was the first to successfully undergo rigorous psychometric assessment.(27) Specific to the ROCKaS, improvements in concussion knowledge and attitudes have been reported over the last decade with 5 – 10% improvements noted across diverse athletic populations.(28-32) Potential reasons for this improvement include the passage of state laws requiring concussion education for high school athletes,(33-35) substantial increases in both sports medicine and general media coverage (e.g., athletes in concussion “protocol”) of sports-related concussions,(36, 37) numerous high-profile lawsuits against professional and amateur sports organizations,(13) and concussions are now generally presented as a serious injury.(37, 38) These combined effects should lead to increased concussion knowledge and safer attitudes. Thus, the purpose of this study was to assess changes in concussion knowledge and attitudes amongst incoming intercollegiate student-athletes over the course of a decade (2010-2012 vs 2021-2023). We hypothesized that more recent incoming student-athletes would have better concussion knowledge as well as safer concussion attitudes.

METHODS

Participants

There were two cohorts of incoming, freshman or transfer, student-athlete participants (N=592) from two institutions that were recruited for this study. The first cohort was recruited from 2010 – 2012 from one institution (N=318) and the second cohort from 2021 – 2023 from a second institution (N=274). (Table 1) Both institutions were NCAA Division 1 Football Championship Subdivision members with comparably sized athletic training staffs at the time of testing. The inclusion criteria for both cohorts were incoming student-athletes who were medically cleared to participate in intercollegiate athletics. There were no exclusion criteria for enrollment. All participants provided written and oral informed consent prior to participation and consistent with each site’s Institutional Review Board (University of Delaware IRB #804454, Georgia Southern University IRB # H11323).

Table 1. Participants Demographics.

*denotes significant difference between groups.

	2010 – 2012	2021 – 2023	p-value
Number	318	274	N/A
Sex (Male/Female)	161/157	122/152	0.175
Prior Concussion*	59/318 (18.6%) Range: 0 – 8	71/274 (25.9%) Range: 0 – 6	0.031
Age*	18.3 ± 0.8 Range: 18 – 20	19.3 ± 1.4 Range: 18 – 21	0.001
Collision Sport*	93/318 (29.2%)	49/274 (17.9%)	0.020
Sports	Football: 29.2% (N: 93) Baseball: 12.6% (N: 40) Track & Field: 11.6% (N: 37) Cheerleading: 9.4% (N: 30) Women’s Soccer: 6.3% (N: 20) Men’s Soccer: 5.7% (N: 18) Swim & Dive: 5.0% (N: 16) Softball: 5.0% (N: 16) Women’s Basketball: 3.5% (N: 11) Women’s Volleyball: 3.1% (N: 10) Men’s Basketball: 3.1% (N: 2.8%) Tennis: 2.8% (N: 9) Rifle: 2.2% (N: 7) Golf: 0.6% (N: 2)	Rowing: 14.6% (N: 40) Football: 12.0% (N: 33) Swim & Dive: 9.9% (N: 27) Baseball: 8.4% (N: 23) Women’s Lacrosse: 6.6% (N: 18) Track & Field: 6.2% (N: 17) Men’s Soccer: 5.8% (N: 16) Men’s Lacrosse: 5.8% (N: 16) Field Hockey: 5.1% (N:14) Softball: 4.4% (N: 12) Women’s Soccer: 3.6% (N: 10) Volleyball: 2.9% (N: 2.9%) Women’s Tennis: 2.9% (N: 2.9%) Men’s Basketball: 2.9% (N: 2.9%) Men’s Tennis: 2.6% (N: 7) Women’s Basketball: 2.2% (N: 2) Golf: 1.5% (N: 4) Cross Country: 1.5% (N: 4) Cheerleading: 1.1% (N: 3)

Instrumentation

The ROCKaS consists of 55 questions which are divided into sections on concussion knowledge and attitude. The concussion knowledge index (CKI) is comprised of 25 questions in three sections; 1) 14 basic true or false questions, 2) three applied true or false questions, and 3) the recognition of eight common concussion symptoms from a list of 16 potential symptoms (distractors are not scored). One question was removed from the ROCKaS for all participants. Specifically, question #14 in the CKI asks, “After 10 days, symptoms of a concussion are usually completely gone” as a True/False question. While this was generally considered to be true in 2010,(39) recent evidence suggests the normal recovery time for a concussion may now be two to four weeks thus reducing the question’s validity.(1) In the 2010 – 2012 cohort, the question was generally answered correctly/“true” (73.6%) whereas the 2021 – 2023 cohort was lower (38.4%). The revised scale’s internal consistently was compared to the original scale (Cronbach’s Alpha: 0.76 original, 0.82 revised).(27) Thus, we removed this question from both groups score resulting in a score range of 0 – 24 with a higher score reflecting better concussion knowledge.

The concussion attitude index (CAI) contains 15 questions scored on a 1 – 5 Likert Scale with safer attitudes receiving five points and the least safe attitude receiving one point. The score range is 15 – 75 with a higher score reflecting a safer attitude. An internal validity index consisted of three unscored true/false questions and a score <2 resulted in the test being considered invalid.(27) The ROCKaS has undergone extensive psychometric review and has been established as reliable (CKI: r=0.67, CAI: r=0.79) and valid (validity scale: 2.2 – 4.8% error rate).(27) The complete ROCKaS questionnaire and outcomes by group are provided in Supplemental Files 1 - 4.

Procedures

Incoming student-athletes were required by their institution and NCAA policy(13) to complete a pre-participation concussion baseline assessment. Upon completion of this assessment, perspective participants were recruited to complete additional assessments including the RoCKAS. If they expressed interest, a separate informed consent was completed thus separating the required from voluntary components. While not specifically tracked, many participants declined due to lack of time or interest as the comprehensive baseline concussion battery was time consuming.(40) There was no financial or other incentive to complete the assessment.

Participants sex was self-reported and concussion history was ascertained via a reliable (ICC: 0.92) questionnaire.(9, 41) The ROCKaS was completed either via a pen and paper (2010 – 2012) or electronic version (2021-2023) as approved by each sites IRB. A previous study amongst collegiate ROTC students showed no differences between assessment modalities.(31) All participants completed the questionnaires independently and were allowed to ask procedural questions to the research staff, but no other assistance was provided in answering the questions. The questionnaires were completed prior to any institutional specific, or NCAA required, concussion educational sessions.

Statistical Analysis

Participant demographics were reported as means, standard deviations, frequencies, and percentages as appropriate and compared with independent samples t-tests. A three factor ANOVA with interactions compared CKI (0 – 24) and CAI (15 – 75) performance with Group (2010 – 2012, 2021 – 2023), Sex, and Concussion History (Yes, No) as predictors. Model adequacy was confirmed through residual analysis and residual outliers (>3 SD from the mean) were removed. Each model was then reduced to remove non-significant predictors and interactions in accordance with the hierarchy principle in statistical modeling. Partial eta squared (η²) values are reported for effect size classified as Small (0.01), Medium (0.06), Large (0.14), 95% confidence intervals are reported, and the alpha value was set at 0.05.(42) Finally, each individual question and symptom were compared between groups using a Chi-Square analysis with either a dichotomous (CKI) or ordinal (CAI) response. Given the 39 individual assessments (CKI: 24, CAI: 15), a Bonferroni correction was applied and the alpha value was adjusted to 0.0013. All analyses were performed with JMP (Version 17, Cary, NC.).

RESULTS

Concussion Knowledge Index

Three observations resulting in residual outliers were removed (2 from 2010 – 2012; 1 from 2021 – 2023). There was a significant main effect for Group (2010-2012: 18.5 ± 2.6, 2021-2023: 19.4 ± 2.5, p=0.011, η² = 0.011) and Concussion History (p=0.033, η² = 0.008), with an overall adjusted R² = 0.041. (Figure 1) In a reduced model with these two factors, only Group (p<0.001, η² = 0.032) remained significant with an adjusted R² = 0.034. (Table 2, Supplemental Table 1)

Figure 1: Concussion Knowledge Index (CKI).

There was a significant difference between 2010 – 2012 (18.5 ± 2.6, Median: 19, Range: 9 – 24) and 2021 – 2023 (19.4 ± 2.5, Median: 20, Range: 10 – 25) groups (F=4.745, P<0.001, η² = 0.011).

Table 2. Group Outcomes.

*There was a significant group difference for both CKI and CAI Symptom recognition is part of the CKI score and thus not analyzed separately.

	2010 – 2012 Mean Score (95% CI)	2021 – 2023 Mean Score (95% CI)	P-Value	η 2
*Concussion Knowledge Index (CKI) (0- 24)	18.5 ± 2.6 (18.2 – 18.8)	19.4 + 2.5 (19.1 – 19.7)	<0.001	0.032
Symptoms (0 – 8)	6.8 ± 1.6 (6.6 – 6.9)	6.5 ± 2.1 (6.2 – 6.7)	NA	NA
*Concussion Attitude Index (CAI) (15 – 75)	52.9 ± 6.0 (52.2 – 53.6)	62.2 ± 6.5 (61.4 – 63.0)	<0.001	0.359

The symptom recognition score (0 – 8) is a component of the overall CKI and not analyzed independently. Descriptively, the 2010 – 2012 group had a higher score (6.8 ± 1.6) than the 2021 – 2023 group (6.5 ± 2.1). The 2010 – 2012 group had a significantly higher recognition on Difficulty Remembering (92.4 ± 26.5% vs 85.9 ± 37.8%, p<0.001) and Dizziness (95.6 ± 20.1% vs 84.3 ± 36.4%, p<0.001). (Table 2, Supplemental Table 2)

Concussion Attitude Index

Nine observations resulting in residual outliers were removed (2 from 2010-2012; 7 from 2021 – 2023). There was a significant main effect only for group (2010-2012: 52.9 ± 6.0, 2021-2023: 62.2 ± 6.5, p<0.001, η² = 0.359) with an overall adjusted R² = 0.381. (Figure 2) Group remained significant (p<0.001, η² = 0.359) in the single factor reduced model. (Table 2, Supplemental Table 3 and 4)

Figure 2: Concussion Attitude Index (CAI).

There was a significant difference between 2010 – 2012 (52.9 ± 6.0, Median: 53, Range: 30 – 67) and 2021 – 2023 (62.2 ± 6.5, Median: 62, Range 42 – 75) groups (F: 324.8, p<0.001, η² = 0.359).

DISCUSSION

The last decade has seen substantial increases in the scientific knowledge and general media coverage of sports related concussion(36-38) which are typically no longer considered a minor transient injury.(1) However, the changing perceptions of student-athletes over time has received limited attention; therefore we compared concussion knowledge and attitudes of incoming NCAA student athletes approximately a decade apart. The primary finding of this study was a small (4.7%) increase in concussion knowledge; however, a substantially safer attitude (~15% increased CAI score) was reported by student-athletes in the most recent cohort (2021-2023). Interesting, neither concussion history nor sex were significantly associated with concussion attitude.(43, 44) It is hopeful these safer self-reported attitudes will translate to increased intent to seek care resulting in improved student-athlete health outcomes. These results also suggest that health care providers should likely focus their education sessions on attitudes as opposed to basic concussion knowledge and symptom recognition.

The overall goal of concussion education programs to increase athletes knowledge of concussion symptoms which would hopefully reduce concussion underreporting.(11) The Theory of Planned Behavior suggests that knowledge is only one component of an individual’s intended behavior as attitude, subjective norms, and perceived behavioral control all likely influence an athlete’s intention to report a suspected concussion.(11, 45) Herein, we identified a significant 4.7% improvement in concussion knowledge across a decade; however, the effect size (η² = 0.033) and the adjusted R² (0.034) were both quite low suggesting this improvement was likely not clinically meaningful. Within individual questions, differences between groups were limited (2 significant differences from 16 questions). (Supplementary Table 1) There was an increased recognition in the 2021 – 2023 cohort of concussions being associated with emotional disruptions which is not surprising as the last decade has seen increased awareness of socio-psychological consequences of concussion.(4, 46, 47) Individual questions typically had >75% accurate responses suggesting good overall knowledge and a potential ceiling effect on the questionnaire; however, two questions were clear outliers. First, a low percentage identified that being “knocked out” from a concussion was experiencing a coma and this may reflect over medicalization of the term “coma”. Secondly, and likely clinically more relevant, high percentages of both groups incorrectly responded (19.2 – 27.6% correct) that a concussion can be identified by “brain imaging”. This presents potential challenges to clinicians as negative imaging outcomes may lead to athletes believing they don’t have a concussion and potentially not complying with follow-up instructions.(48) Both of these questions have been consistently answered incorrectly on the RoCKAS(29, 31, 49-51) with the notable exception of professional women’s soccer/football athletes (>90%)(52) on the coma question. Overall, incoming collegiate student-athletes have good overall concussion knowledge which suggests clinicians may be able focus their education sessions on improving attitudes and reporting considerations with the notable exception of diagnostic imaging.

The participants symptom recognition performance was included in the overall CKI score, as per guidelines,(27) and both cohorts correctly identified over 80% of the symptoms overall. The performance of the two groups was largely comparable to prior symptom recognition studies using the RoCKAS (6.1 – 7.5 correct) questionnaire.(29-31, 49, 52) There were group differences on two symptoms with the 2010-2012 cohort surprisingly recognizing both “difficulty remembering” (92.4% vs 82.9%,, p<0.001)) and “dizziness” (95.6% vs 84.3%, p<0.001) more than the 2021-2023 cohort. Herein, three of the eight concussion symptoms were correctly identified less than 80% which were “feeling slowed down” (71.6 – 76.9%), “feeling in a fog” (69.6 – 78.%), and drowsiness (63.4 – 69.0%) which have typically been the less frequently correctly identified symptoms in some but not all, studies.(31, 49) Similar to the CKI findings, the results of the symptom recognition score also support clinicians focusing education sessions on reporting considerations more so than basic factual knowledge.

There was a significant and likely meaningful improvement in the CAI within the study with the 2021 – 2023 cohort endorsing ~15% safer attitudes than the 2010 – 2012 cohort (2010-2012: 52.9 ± 6.0, 2021-2023: 62.2 ± 6.5). The 2021 – 2023 cohort displayed very safe attitude response with a mean score of 62.2 of out a potential 75 which comparable to a recent (2021) study on women’s English professional football/soccer players (63.3).(52) Travis recently reported that safer attitudes on the ROCKAS were associated with increased likelihood of reporting a potential concussion to a medical provider or coach.(49) This is an encouraging finding as the highest ROCKAS safe attitudes scores (>60) have occurred herein and on recent studies.(31, 52) Overall, a positive trend is emerging over time as the earlier group herein (2010-2012) had the lowest CAI (52.9), which generally increased to 57 in 2014,(51) 58 in 2017(50) and now scores routinely exceed 60.(52) While these results are encouraging, it is important to note that questionnaire responses may not reflect actual decisions made by athletes during participation.(29)

There were numerous significant differences within specific questions (10 out of 15 questions); however, the group median and mode (appropriate reporting statistics for ordinal data) tended to be the same and the findings were likely the result of differences in the distributions. For example, CAI question #1 (continuing participation with a mild concussion) had the same median (2, somewhat disagree) and mode (1, strongly disagree), but were significantly different (p<0.001). This occurred because when combining disagree (2) and strongly disagree (1) values, this reflected 60.8% (192/316) of the 2010 – 2012 respondents as opposed to 81.6% (213/267) of the 2021 – 2023 cohort and similar trends were seen in many of the questions. Overall, these results are encouraging and show both that the 2021-2023 group had a safer concussion attitude and over 70% of respondents selected the safer answers. (53) While these results are encouraging, it is important to note that questionnaire responses may not reflect actual decisions made by athletes during participation.(29) The one CAI question whereby the 2010-2012 cohort performed better was suggesting the Athletic Trainer, rather than the athlete, should decide on their RTP timeline. The two groups had the same median (4) and same percentage with SA/A (82.0%), but the difference is likely because the 2010-12 had ~3x more participants select Strongly Agree than Agree (191 vs 68) whereas the 2021-2023 group was more evenly split (103 and 116). There is no clear explanation for this finding and is worthy of continued investigation and likely an important area for health care providers to reinforce during education sessions. Overall, these results suggest that while safer attitudes are being reported, there are still areas for clinicians to target in educational sessions.

The questionnaire responses herein, like all questionnaires, are limited by participant honesty in their responses and potentially subject to a social desirably bias; when participants intentionally underreport socially undesirable attitudes and over report more perceived desirable attitudes.(54) Despite the large enrollment (N=592), a potential non-responder bias may have also be present amongst participants who elected not to enroll in the study and they may have different knowledge or attitudes than the respondents. The quantity and quality of prior concussion education sessions (e.g., high school, private leagues, etc) that participants attended was unknown and state laws have differing educational requirement levels. Both institutions in this study were NCAA Division I FCS at the time of data collection and therefore results may not be translatable to other groups. The effect of the COVID pandemic on concussion attitudes is unknown and differences were previously noted on mental health measures.(55) Finally, the RoCKAS underwent extensive psychometric testing during the questionnaire’s development in the mid-2000’s; however ongoing research has expanded concussion knowledge and affects the questionnaire (i.e., we removed the CKI question on recovery within 10 days). Thus, more contemporary questionnaires may be timelier, but to measures changes over a decade the same instrument was utilized.

The results of this study suggest that over the last decade concussion knowledge and symptom recognition have shown small improvements (<5%); however, the current cohort (2021 – 2023) showed a significant ~15% safer attitude with a large effect size (η² = 0.36). The overall scores on the RoCKAS show continued improvement in athletes concussion knowledge and attitudes across various athletic populations as the results herein were fairly consistent with other recent studies and better than reports 5 – 10 year prior. Furthermore, these results provide health care providers with several specific symptoms (“feeling in a fog”, “feeling slowed down” and drowsiness) and specific procedural considerations (e.g., health care providers make RTP decisions, imaging is not effective for diagnosis concussion) to address during NCAA mandated concussion education sessions. Continued efforts to improve concussion recognition and reporting by student-athletes may reduce the rate of unreported concussions and improve student-athlete healthcare.