The Ivy League–Big Ten Epidemiology of Concussion Study is a surveillance system and prospective cohort study with the goal to better understand SRC and produce evidence to benefit the health and well-being of student-athletes. Participating sites include the eight Ivy League and 12 of 14 universities in the Big Ten.

An informed consent procedure is used to recruit student-athletes who sustain a SRC or non-sport-related concussion as defined by the most recent CISG consensus statement at the time the SRC occurred.[3, 4] Athletic trainers and research coordinators abstract demographic and clinical information including concussion history from the medical record to document a new concussion and administer outstanding questions to the athlete directly. The present analysis includes data from 2013 to 2020 and is limited to athletes in contact/collision sports,[5] given that return to full-contact practice is a stage of the RTS progression in the 2017 CISG consensus statement.[3] The data include symptom burden, measured as whether at any time post-injury the athlete reported experiencing each of the 22 symptoms assessed in the Sport Concussion Assessment Tool (SCAT),[6] a standardized instrument endorsed by the CISG.[3]

Athletic trainers, paired with members of the multidisciplinary concussion management team, re-initiate and progress student-athletes through academic and athletic activities following concussion. This study does not directly evaluate details of the RTS protocols and the extent to which practices vary across study sites; however, each site must comply with the NCAA concussion management checklist.[7] Instead, each site records whether the athlete had an initial period of 24–48 hours physical rest and cognitive rest immediately post-injury.

Our data collection form asks “Did the student progress through the return to play protocol as expected?” with a binary (yes/no) response to indicate if any stage of the RTS was repeated or incomplete. Also, dates are recorded when each athlete 1) self-reported being asymptomatic (i.e., returned to baseline symptoms), and returned to 2) exertion activity, 3) limited sport and 4) full sport, and 5) full academics. Each of these milestones is relevant to one of the stages in the RTS and return-to-academics strategies in the 2017 CISG consensus statement and the 2014 NATA position statement. For example, our “exertion activity” milestones corresponds to either “light aerobic exercise” (stage 2) or “sport-specific exercise” (stage 3) in the 2017 CISG consensus statement; “limited sport” corresponds to either “sport-specific exercise” (stage 3) or “non-contact training drills” (stage 4). We refer to these as return-to-activity stages.

We determined the sequence in which each athlete reached the five return-to-activity stages. We then classified each athlete according to the “profile” describing their return.

The primary outcome was returning to full sport ≤14 days post-injury (i.e., yes/no), with prolonged delay to full sport >28 days as a secondary outcome. The timing of these outcomes is based on finding that 50% of student-athletes in the Ivy League and Big Ten return to full sport in 14 days, yet one-third require longer than 28 days, making them clinically relevant.[8] Further, these outcomes are easily translated into weekly increments, to aid in interpretation.

Adding covariates for athlete age, years competing in their sport, academic year, or time since last concussion in the adjusted logistic regression models did not improve model fit. Return to full sport ≤21 days post-injury was examined in a sensitivity analysis to assess whether the findings were sensitive to the time-period chosen as the outcome. We repeated the regressions using casewise deletion to gauge imprecision and the magnitude and direction of bias overcome using multiple imputation.

Table 1 reports that 1,715 student-athletes sustained a SRC. Over two-thirds (67.9%) had initial 24–48 hours of both physical and cognitive rest post-injury, 23.8% had 24–48 hours of only physical rest, and 2.6% had only cognitive rest. Athletes with 24–48 hours initial physical and cognitive rest endorsed more symptoms (median=10, IQR=7–15) than did other athletes (median=7, IQR=5–11, P<0.001) and had longer times to symptom resolution (median=10 days, IQR=5–22 versus 4, IQR=2–10, P<0.001). Having 24–48 hours physical and cognitive rest increased from 66.2% to 71.1% (P<0.05) after the 2017 CISG consensus statement.

In athletes who had symptom resolution ≤14 days post-injury, the incidence of symptom resolution was most common on day 2 post-injury and decreased steadily as days elapsed post-injury (Figure Supp1, online).

Figure 1 shows that three recovery profiles accounted for how the majority (88.4%) of athletes reached the stages. Approximately one-third (38.0%) had symptom resolution first, and then returned to full academics, physical exertion, limited sport, and full sport, respectively (Profile 1); which is most consistent with the 2017 CISG consensus recommendations. In Profile 2, 10.7% had symptom resolution first, followed by initiating physical exertion, then return to full academics, then limited sport, then full sport. The most common sequence, exhibited by 51.3% of athletes, was return to full academics first, then symptom resolution, then physical exertion, limited sport, and full sport (Profile 3).

The profiles did not differ significantly in the percent of athletes with ≥1 previous concussion (χ²=1.52, p=0.47), proportion female (χ²=3.42, p=0.18), nor percent from each athletic conference (χ²=2.38, p=0.31). The profiles did significantly differ in having an initial 24–48 hour period of rest, however, which was exhibited by 75.0% of athletes in profile 1 and 83.3% of athletes in profile 2, but 47.8% of athletes in profile 3 (χ²=70.64, p<0.01). The median symptom count was 8 (IQR=5–12) in Profile 1, 9 (IQR=5–12) in Profile 2, and 9 (IQR=6–14, χ²=5.11, p=0.08) in Profile 3.

An additional 16 sequences were exhibited in the remaining 6.6% of athletes; each was uncommon, accounting for <2% of athletes. In the sample overall, very few (5.0%) athletes initiated physical exertion as a first event.

A similar proportion of athletes followed Profile 3 before (52.2%) and after (49.6%) the 2017 CISG consensus statement. The prevalence of Profile 1, where full academics occurred before exertion activities, decreased from 40.5% to 33.3% after the 2017 statement, and the prevalence of Profile 2 where exertion occurred before full academics increased from 7.3% to 17.1% (P<0.001).

Figure 3 shows scatterplots of the timing when (i.e., days post-injury) each athlete reached each of the stages relative to when they experienced symptom resolution. One quarter (25.5%) initiated physical exertion on the same day symptoms resolved, two-thirds (66.8%) initiated exertion only after symptoms resolved, and few (7.7%) initiated exertion before symptoms resolved (Figure 3B). In contrast, 19.3% of athletes returned to full academics on the day symptoms resolved, one quarter (28.8%) returned to full academics only after symptoms resolved, whereas half the athletes (51.9%) returned to full academics before symptoms resolved (Figure 3A). The great majority of athletes initiated limited sport (Figure 3C) and full sport (Figure 3D) only after symptoms resolved.

Having 24–48 hours initial physical and cognitive rest was associated with lower likelihood of return to full sport ≤14 days (ARD −25.1%, 95%CI −30.7% to −19.5%) (Table 2D). While controlling for 24–48 hours initial rest and the additional potential confounders, the likelihood of returning to full sport ≤14 days was 64.3% in athletes who initiated exertion on the same day of symptom resolution (Table 2A and Figure 5A) and was not statistically different in athletes who initiated exertion before symptom resolution (Table 2D and Figure 5A). In contrast, in athletes who initiated exertion after symptom resolution, the likelihood of returning to full sport ≤14 days was progressively lower as days additional days elapsed between the two stages (Table 2D and Figure 5A).

Modeling the academic stage of the return sequence revealed 4+ days prior to symptom resolution as a relative threshold for early academic return. Specifically, initiating academics 4+ days before symptom resolution was associated with a 21.5% lower likelihood (95%CI −27.4% to −15.5%) of returning to full sport ≤14 days (Table 2D) after controlling for initial rest, timing of exertion relative to symptom resolution, and the additional confounders.

Our tests for effect modification identified that athletes who had initial 24–48 hours physical and cognitive rest were less likely to return to full sport ≤14 days regardless of whether they initiated exertion before, on, or after the day they experienced symptom resolution (Table Supp1 and Figure 5B). Analogously, athletes who returned to academics 4+ days prior to symptom resolution were less likely to return to full sport ≤14 regardless of whether they initiated exertion before, on, or after they day they experienced symptom resolution (Table Supp1 and Figure 5C). Results were similar when using return to full sport ≤21 days as the outcome as a sensitivity analysis (Table 2B and 2E, Table Supp1, Figure Supp2).

Having 24–48 hours initial physical and cognitive rest was associated with a higher likelihood that return to full sport would be prolonged >28 days (ARD 12.9%, 95%CI 8.3% to 19.5%) (Table 2F), and initiating academics 4+ days prior to symptom resolution was associated with a higher likelihood of prolonged delay to full sport (ARD 19.1%, 95%CI 13.4% to 24.7%) (Table 2F). While taking these into account, the likelihood that athletes’ return to full sport would be prolonged >28 days was 17.7% (95%CI 13.4% to 22.0%) for those who returned to exertion on the same day as symptom resolution (Table 2C and Figure 5D). Compared to these athletes, initiating exertion considerably before (i.e., 7+ days) symptom resolution was associated with a 19.3% higher likelihood (95%CI 6.7% to 31.9%) of prolonged delay to full sport (Table 2F and Figure 5D). In contrast, initiating exertion only 3–4 days before symptom resolution was associated with a 14.1% lower likelihood (95%CI −20.7% to −7.5%) of prolonged delay to full sport (Table 2F and Figure 5D).

Athletes who had initial 24–48 hours physical and cognitive rest were more likely to have prolonged return to full sport >28 days regardless of whether they initiated exertion before, on, or after the day they experienced symptom resolution (Table Supp1 and Figure 5E). Analogously, athletes who returned to academics 4+ days prior to symptom resolution were more likely to have prolonged return to full sport >28 regardless of whether they initiated exertion before, on, or after they day they experienced symptom resolution (Table Supp1 and Figure 5F).

Two of every three athletes (67.9%) had an initial 24–48 hours of physical and cognitive rest before starting a RTS progression. These athletes endorsed more symptoms than did others, and took longer for symptoms to resolve, suggesting that having 24–48 hours initial rest may correspond to SRC severity. Initial physical and cognitive rest became more common over the study period, consistent with the possibility that athletic trainers were adopting the recommendation for brief initial rest included in 2017 CISG consensus statement.

Note that initial physical and cognitive rest was associated with lower likelihood of return to full sport ≤14 days, and higher likelihood of prolonged full sport >28 days. This may be in part due to acute rest setting back the date when the athlete could begin a RTS progression. As such, this finding should not be interpreted to mean that rest is detrimental. Another factor is athletes’ symptom experiences, as those who completed 24–48 hours of initial rest endorsed more symptoms. This could explain why an initial 24–48 hours of rest was associated with a delay to symptom resolution and full sport, but cannot be understood definitively here given the observational nature of the study.

The sequential progression exhibited by student-athletes with SRC indicates that athletic trainers are providing active management consistent with best practice recommendations of the CISG and NATA. This study did not set out to explicitly measure RTS stages as defined by either the CISG or NATA. Rather, the data include dates of key milestones that map to the stages defined in the CISG and NATA guidelines. Importantly, whereas the CISG consensus statement describes a return-to-academics strategy that is separate from their RTS (i.e., return-to-sport) progression, we captured time (i.e., days post-injury) to return to full academics. This was one of five return-to-activity stages we used to investigate their timing relative to each other, and their relation to the outcome of interest, return to full sport.

Using these five stages let us detect that the most common progression, exhibited by 51.3% of the athletes, involved returning to full academics first, then experiencing symptom resolution, followed by initiating exertion, limited sport, and full sport (Profile 3). However, 38.0% of our sample exhibited Profile 1: symptom resolution, return to full academics, initiating exertion followed by a gradual reintroduction to sport. Profile 1, in general, most closely corresponds to the 2017 CISG consensus statement recommending symptom-limited activities, including a gradual re-introduction of school activities, that do not provoke symptoms early during the return-to-sport strategy. Importantly, we found that athletes in all three profiles adhered to CISG guidelines, in that sport-specific return stages occurred after symptom resolution; however, we identified variability as to when return-to-academics occurred in the sequential stages of return. Notably, unadjusted analyses identified that athletes within Profile 3, with academic return as their first stage, exhibited delayed RTS, as the median time to full sport was 14 days compared to 11 days in athletes in Profiles 1 and 2.

An early intervention in concussion management that is increasingly accepted includes sub-symptom threshold activity, which provides evidence in decreasing symptom burden,[13] and additional research supporting this notion was published since the 2017 CISG consensus statement was reported.[14–17] In the present study, we were not able to detect whether or when athletes may have pursued sub-threshold exercise as treatment for concussion. Instead, we measured the initiation of exertion, where the median time was considerably after the injury date (i.e., 9 days). Our profiles indicated that the majority of athletes began exertion activities after symptom resolution, and therefore, future research should further investigate sub-symptom threshold exercise in early stages of a RTS protocol.

The shortest time to return to full sport ≤14 days occurred in athletes who initiated sport-related exertion on the same day their symptoms resolved, which was a median 2 days longer in athletes who initiated exertion after symptoms resolved, and 4 days longer in athletes who initiated exertion before symptoms resolved. Further, initiating exertion just 3–4 days prior to symptom resolution was associated with decreased likelihood of a prolonged return to full sport >28 days. In contrast, initiating exertion considerably before (7+ days) symptom resolution corresponded to a higher likelihood of delayed recovery beyond 28 days. These are novel findings, suggesting there may be prolonged-delay consequences of returning an athlete to sport too quickly; yet, there may be benefits of starting exertion at an appropriate time shortly before symptom resolution.[17] Often, an athlete’s progress on the timeline to symptom resolution is only known in retrospect, emphasizing the importance of continual monitoring of athletes’ symptom recovery. This observational study could motivate future research using approaches including real-time monitoring[18] to identify whether an optimal time to introduce exertional activity exists. In addition, those with prolonged recovery >28 days generally had a greater number of symptoms. We controlled for this as this may be a potential reason for the longer initial rest period or delay in initiation of a return protocol. Accordingly, the results of this study should motivate further research to identify appropriate timing and intensities for sport-related exertion activities as part of a RTS progression in collegiate athletes, and the benefits or consequences of initiating exertion activities relative to symptomology.

Initiating full academics 4+ days before symptom resolution was associated with lower likelihood of return to full sport ≤14 days, and higher likelihood of prolonged recovery >28 days. This is consistent with studies finding that athletes who return to academics before their symptoms resolve may be overexerting themselves cognitively or prioritizing academics results in delayed recovery, but is the first evidence from a large collegiate-athlete population. In a study of patients presenting to a sports concussion clinic, higher self-reported cognitive activity was associated with a longer duration of symptoms.[19] That study included patients who presented after a considerable delay (i.e., up to 3 weeks), however, and the median age was 15 years and a large percentage of patients (19%) had loss of consciousness. Moreover, a chart review found that almost half (44.7%) of the student-athletes treated at a sport medicine practice returned to school too soon, as evidenced by a relapse or recurrence of symptoms.[20] Only 20% percent of those students were in college, however, with the remainder in elementary and high school. Taken together, this is evidence that consensus guidelines should consider having cognitive or academic activities explicitly in a RTS protocol. Also, the CISG and NATA should consider updating their guidance related to academics to include students participating at the collegiate level.

The sequence by which athletes complete stages of a RTS protocol may be the result of many factors in addition to oversight by an athletic trainer and their integration with consensus guidelines,[1] including rationale of different guidelines[21] as well as pressure and priorities of key stakeholders involved in concussion management.[22] For example, in a survey of college athletic trainers, 36% indicated they have not felt pressured to return an athlete to the classroom after a concussion, whereas only 18% reported no pressure to return an athlete to physical activity.[22] Another factor may be student-athletes’ or clinicians’ priorities for returning collegiate athletes to sport prior to or after full classroom participation without restrictions, as previous research identified 13% of athletic trainers were neutral and 77% agreed that classroom participation must come first.[22]

In the collegiate setting members of the interdisciplinary concussion management team must adhere to the NCAA concussion management checklist.[7] Therefore, although our results suggest some variation in return to activity progressions, mandated homogeneity in concussion management exists at this level. The differences may exist as result of the ability of, and way in which, sites carry out guidelines, involvement or inclusion of members of the multidisciplinary care team, variability in resources, and individuality in concussion management. Importantly, the findings in the current study demonstrate the utility of individualized concussion management plans,[23, 24] as clinical impairments and needs of student-athletes following SRC are not uniform. Accordingly, this study should motivate future work to identify particular reasons for the different progression profiles that we identified in this collegiate student-athlete population.

The results may not be generalizable to non-contact or non-collision sport athletes, or athletes participating in levels of competition other than college. Also, we defined each athletes’ experiences after SRC based on dates when they reached a relevant outcome, rather than a more nuanced accounting of levels of exertion, both physical and cognitive. In addition, the data lack a number of comorbid factors that may influence recovery after SRC, including migraine, psychiatric condition, or learning disability. Further, it is unknown whether athletes or members of their concussion management team prioritized one outcome over another (e.g., academics versus sport). Future work could prospectively monitor physical and cognitive activity that athletes exert through a RTS progression, and determine whether duration and intensity affects recovery. Also, while we were able to control for symptom burden as measured by symptom count, we were not able to control for symptom severity which has been suggested be a consistent predictor of delayed recovery outcomes.[22][25] The CISG[3] and NATA[1] both recommend academic accommodations as part of a return-to-school strategy. Our data include only a yes/no indicator of whether a student-athlete received academic accommodations, which prevented us from studying their role in recovery. Finally, we cannot establish whether an athlete’s condition due to SRC affected when they reached each outcome, versus whether their progression through RTS stages affected their time to recovery given the observational nature of the study. Being a prospective cohort study, however, along with other study design features that are strong including a large sample and detailed, date-specific data from many sites, do elevate the strength of our evidence that the timing of early or late exertion and full academics are associated with prognosis in terms of time to full sport participation.