After sport-related concussion, a variety of signs and symptoms may be present, such as loss of consciousness, dizziness, headache, or mental status changes; thus, because of the heterogeneity of concussion symptom presentation after suspected injury, a multifaceted approach to assessment is recommended (1). Among the different domains assessed using the Sport Concussion Assessment Tool-3 (SCAT-3), or more recently the Sport Concussion Assessment Tool-5 (SCAT-5), it is recommended that a gait/balance evaluation should be conducted and that these evaluations add relevant clinical utility (2). Specifically, either tandem gait (TG) and/or the modified Balance Error Scoring System (mBESS) is recommended to assess balance after injury in the SCAT-3 and TG as part of the neurological screen in the SCAT-5 (2–4). The Balance Error Scoring System (BESS) is more commonly used by athletic trainers in concussion management; however, the BESS has consistently demonstrated low sensitivity when used as a diagnostic test acutely after concussion (0.34) (5–8). In addition, the psychometric properties of the test are poor with low interrater (0.57) and intrarater (0.74) reliability and high minimal detectable change (MDC) scores (7.3–9.4 errors), which exceed the normal postconcussion change of 3–6 errors (9,10). BESS interpretation is also limited by a practice effect secondary to repeat administration (11,12) and is negatively influenced by testing environment, ankle instability, and acute exertional fatigue (13–15). Therefore, several factors can influence BESS performance independent of a concussion.
Both the SCAT-3 and the SCAT-5 recommend the mBESS, which is scored from the firm stance trials of the BESS, for balance assessment (2). There is substantially less postconcussion literature on the mBESS compared with the BESS; however, Buckley et al. (16) recently identified a moderate sensitivity (0.71) for the mBESS acutely after concussion. Previously, the SCAT-3 evaluated TG on the basis of the best time of four trials, whereas the SCAT-5 now incorporates a normal TG with the neurological screen (2,3). Although still mentioned in the SCAT-5, TG is not discussed to the same degree as the BESS or mBESS in the concussion literature and represents an alternative method to evaluate postural control after a suspected concussion (16–19).
TG is a clinically feasible and highly reliable (intraclass correlation coefficient, 0.97) tool used to evaluate dynamic balance, speed, and coordination (16,20,21). The cerebellum plays a key role in regulating posture and balance during coordinated movements, as it receives afferent information from multiple sensory systems and sends it to the motor cortex and other control centers (22); thus, a highly coordinated movement, such as the heel-to-toe gait pattern of TG, may likely demonstrate impairments after a concussion (19). Normative parameters of TG have been established in children (23) collegiate student-athletes (18), and professional athletes (24); however, the effect of concussion on TG is still largely unexplored. Howell and colleagues (19) recently identified TG performance differences between a concussion group and a control group, particularly in a dual-task setting; however, the concussion participants lacked baseline data and were solely compared with a small group of healthy controls. Furthermore, alterations in postural control have been successfully identified through comparisons to premorbid data in other postconcussion gait tasks; therefore, baseline/postconcussion comparison of TG may lead to improved knowledge regarding its ability to assist with concussion management decisions (25,26).
Instrumented assessments of standard gait and transitional gait movements (i.e., gait initiation, gait termination) have successfully identified deficits after a concussion in the acute phase and beyond the resolution of symptoms (4,25–27). These dynamic tasks have elucidated impairments up to 2 months after concussion, which is beyond typical clinical recovery (i.e., 3–5 d for BESS to return to baseline) and further highlights the need for a more robust measure of postconcussion postural control (5,28–30). Much of the postconcussion postural control literature has identified acute and lingering deficits using sophisticated, instrumented motion capture and force plate systems (25–31). Unfortunately, these instrumented measures are not clinically feasible given the high equipment costs, technical challenges, and trained personnel required to perform laboratory studies. The use of a TG task may allow for a translational approach between the laboratory and the sideline, because it is applicable to all postconcussion testing environments and requires minimal cost, training, and time.
The MDC reflects the amount of time necessary to represent a change in dynamic postural control, as opposed to the normal variability of performance between raters (9). Although the BESS is a clinically feasible assessment to conduct, the high MDC scores (9) may reduce clinician confidence in appropriately determining that a change in score reflects actual performance change (e.g., postural instability) and not tester variability. There is currently no MDC established for TG, so the threshold for a significant change in performance for TG time is unknown. Determining the MDC for TG will allow clinicians to better interpret TG time and differentiate true change in performance from measurement variability.
The acute effects of concussion on gait parameters are well established (4,25–27); however, the translation of these effects to TG has received limited attention. Therefore, the primary purpose of this study was to investigate postconcussion TG times, as well as BESS and mBESS errors in a collegiate student-athlete population. Furthermore, we calculated the sensitivity and specificity of TG acutely after concussion. In addition, we sought to establish MDC values in a healthy collegiate student-athlete population to determine the meaningfulness of postconcussion changes in TG. We hypothesized that a concussion would result in a longer time to complete TG and postconcussion individuals would exceed the calculated MDC. Furthermore, we hypothesized that TG would demonstrate higher sensitivity and specificity than both the BESS and the mBESS.
We recruited 76 National Collegiate Athletic Association (NCAA) Division I student-athletes for this investigation: 38 in the concussion group and 38 in the control group (Table 1). The concussions were identified by a certified athletic trainer and subsequently diagnosed by a physician on the basis of the criteria from the 4th International Conference on Concussion in Sport consensus statement (32). The control participants were then identified to match those in the concussion group on the basis of sex, sport, and BESS test administrator. To establish the TG MDC, a separate group of 10 control student-athletes (Table 1) were recruited.
Participants were included in the study if they were active members on an NCAA team at the host institution and medically cleared for athletic participation. The exclusion criteria included any self-reported neurological disorder, current lower extremity orthopedic injury, and metabolic, vestibular, vision disorders, or other conditions, other than the acute concussion, that would impair gait performance. Each participant provided oral and written informed consent in accordance with the University’s institutional review board (IRB). The TG and BESS data were collected under separate IRB protocols, because the BESS and mBESS data were collected under the Concussion Assessment, Research, and Education (CARE) Consortium Clinical Study Core IRB.
The BESS and mBESS components of this investigation were collected in accordance with the CARE Consortium Clinical Study Core protocol, which has been previously described in detail (33). Briefly, in CARE, each participant completed a multifaceted concussion assessment battery, consisting of a graded symptom checklist, mental status examination, postural control task, and computerized neurocognitive assessment at both baseline and postinjury time points (33).
TG was collected in accordance with the SCAT-3 guidelines, which was the contemporary consensus statement at the time the study was performed (3). The participants were instructed to stand behind the starting line with feet together and, in response to a verbal cue, they walked with alternating heel-to-toe gait, in a forward direction, as quickly as possible, along a 3-m-long, 38-mm-wide line of sports tape. Once the participants reached the end of the line, they completed a 180° turn and returned to the starting line with the same heel-to-toe gait pattern. The total time required to successfully complete the test was recorded for each trial, and the best time was retained for further analysis (3). Successful trials required the participants to complete the task in ≤14 s at baseline and without stepping off the line or separating the heel and toe. According to the SCAT-3, unsuccessful baseline trials were to be repeated, if possible, but because of time constraints with the student-athletes in this investigation, unsuccessful trials were not repeated. All participants had at least three successful trials, which is in agreement with previous TG literature (16–18). All trials were timed using a handheld stopwatch (Champion Sports, Marlboro, NJ).
The BESS is a static assessment that involves standing in three different stance configurations (double, single, tandem) in two different conditions (firm and foam) for 20 s with the eyes closed (34). The BESS is scored by counting the total number of errors that occur over the 20-s period, and errors are defined as opening eyes, lifting hands off hips, falling out of position, lifting the forefront or heel, abducting the hip by more than 30°, or remaining out of the test position for longer than 5 s (34). The BESS was scored on the basis of the revised scoring system, with multiple errors occurring simultaneously being counted as one (7). The total number of errors possible in each condition is 10, resulting in 0–60 possible total BESS errors. The mBESS score is determined from the number of errors during the three stances of the firm condition only and herein was not performed as a separate test (2).
All of the participants completed baseline testing before the start of the athletic season (time 1). The concussion group was retested acutely (≤48 h) after concussion, and the controls were retested during the next year’s baseline (time 2).
Data and statistical analysis
This was a prospective, longitudinal study and the independent variables were group (concussion, control) and time (time 1, time 2). The dependent variables were the TG best time, total BESS score, and mBESS score. Independent t-tests were conducted to determine if there were any baseline group differences for TG time or BESS score between groups. The dependent variables of interest were compared with a 2 (group) × 2 (time) mixed-design ANOVA. The MDC was calculated using the standard formula (MDC = 1.96 × SEM × √2), where SEM is the standard error of measurement (9).
Sensitivity (true positive) and specificity (true negative) were calculated on the basis of the clinically diagnosed classification of “concussion/no concussion,” with control participants representing the “no concussion” group. A receiver operating characteristic (ROC) analysis was performed to identify the trade-off between sensitivity and specificity for both TG and BESS. The diagnostic accuracy of the dependent variables was calculated using the area under the curve (AUC) for the ROC. To determine the accuracy of the AUC for TG and BESS to distinguish between groups acutely after concussion, we used the following criteria values: 1.00–0.90, excellent; 0.89–0.80, good; 0.79–0.70, acceptable; 0.69–0.60, poor; and 0.59–0.50, failure (36). Statistical significance was set at P < 0.05, and all statistical analyses were performed with SPSS (version 22; IBM Inc, Armonk, NY).
All participants completed all trials of both TG and BESS without falls or other incidents. The mean time between time 1 and time 2 for the concussion group was 167.6 ± 105.1 d (range, 6–515 d), and the mean time between time 1 and time 2 for the controls was 335.2 ± 57.1 d (range, 214–444 d).
There was a significant interaction (F = 8.757; P = 0.004, = 0.106) for TG performance (Fig. 1A). We observed a significant increase (worsening) for acute concussion participants (time 1, −10.59 ± 1.53 s; time 2, 11.80 ± 2.67 s; P = 0.038; Cohen d = 0.81), but no differences were noted in the control group (time 1, −10.13 ± 1.72 s; time 2, −9.93 ± 1.85 s). Acutely after concussion, TG had a sensitivity of 0.632 and specificity of 0.605 (Table 2). The AUC on the ROC curve was 0.704 (Fig. 2). In the concussion group, 24 (63%) of 38 participants performed worse (i.e., took longer to complete) on TG acutely after injury. The interrater MDC score for TG was 0.38 s, and 23 (96%) of 24 concussion participants who performed worse on TG acutely after concussion exceeded the MDC. For the control participants, 15 (39%) of 38 performed worse (took longer to complete) on TG.
There were no significant differences in BESS performance between time points for either group (concussion: −14.00 ± 7.10 errors for time 1 and 13.71 ± 6.29 errors for time 2; controls: −13.08 ± 6.66 errors for time 1 and −13.47 ± 6.12 errors for time 2; F = 0.235; P = 0.630; = 0.003; Cohen d = 0.04) (Fig. 1B). The sensitivity of the BESS acutely after concussion was 0.447, with a specificity of 0.500 (Table 2). The AUC on the ROC curve was 0.508 (Fig. 2). Among the postconcussion participants, only 45% (17/38) had BESS scores worse than baseline and only 18% (3/17) of those individuals had postconcussion BESS scores that exceeded the established MDC values (9.3 errors) (9). Fifty percent of the control participants (19/38) performed worse (i.e., had a greater number of errors) on the BESS at time 2.
In addition, there were no significant differences in mBESS performance between time points for either group (concussion: 3.32 ± 3.37 errors for time 1 and 3.50 ± 3.19 errors for time 2; controls: −2.82 ± 2.45 errors for time 1 and −3.05 ± 2.87 errors for time 2; F = 0.007; P = 0.935; = 0.000; Cohen d = 0.15) (Fig. 1C). The sensitivity of the mBESS acutely after concussion was 0.474, with a specificity of 0.632 (Table 2). The AUC on the ROC curve was 0.535 (Fig. 2). Only 47% (18/38) of the postconcussion participants demonstrated a worse score on the mBESS acutely after injury. For the control participants, 14 (37%) of 38 had a worse mBESS performance (i.e., a greater number of errors) at time 2.
This investigation evaluated TG, BESS, and mBESS at baseline and acutely after concussion to determine how a concussion affects performance on each of these tests. Overall, TG was significantly worse (i.e., took significantly longer to complete the task) within 48 h of a concussion, with a 1.21-s mean increase in time compared with baseline, whereas controls maintained a relatively constant performance across time. Conversely, there were no significant differences in the number of postconcussion BESS or mBESS errors relative to baseline for the concussion or control group. Furthermore, TG demonstrated a higher AUC and more individuals exceeded the MDC compared with the BESS or mBESS. These findings suggest that TG provides a postural control test that is a clinically valuable assessment in identifying postconcussion impairments, whereas the BESS and mBESS herein displayed poorer sensitivity (0.447–0.474) and specificity (0.500–0.632), potentially limiting the clinical usefulness of the examinations (Table 2).
After concussion, student-athletes took significantly longer (1.21 ± 1.1 s) to successfully complete TG relative to their preinjury performance. Conversely, the healthy controls stayed relatively consistent across testing points (10.13–9.93 s) and the improvement was well below the TG MDC (0.38 s). The performance by healthy controls extends previous research that has reported that gait-related tasks are highly stable in college-age student-athletes (4,25), and baseline TG times for both groups were within the established normative values for TG (18). There is minimal literature on postconcussion TG; however, our findings were consistent with Howell et al. (19), who demonstrated that concussed individuals (17.7 ± 5.9 s) took longer to complete single-task TG than did healthy controls (9.9 ± 2.7 s) when tested within 72 h after injury. The similar TG times in the control groups of both investigations further support the consistency of the TG task. The apparent differences in TG time between the study by Howell et al. (19) and this investigation may be due to an older sample of participants in the current study. Furthermore, the average return to play time in the investigation by Howell et al. (19) was 26.8 d, suggesting that participants might have suffered from more severe postconcussion impairments. The larger sample size herein, as well as the inclusion of baseline data, strengthens and extends the postconcussion TG knowledge. These investigations provide further evidence that TG is a useful and clinically feasible assessment after concussion.
There was no significant group–time interaction for the number of BESS errors. The performance on the BESS was equivalent between the concussion and control groups. It is important to note the acute time point herein was <48 h after concussion, and these results are fairly similar to earlier larger studies. For example, McCrea et al. (5,36) reported approximately a 1.0-error increase at 48 h after concussion in NCAA collegiate athletes. The limitations of the BESS are well documented, particularly related to repeat administration (11,15,37), which could explain our lack of substantial findings after injury. Furthermore, the BESS is a novel assessment that examines static postural control in different stances, and it has been suggested that dynamic motor activities, such as gait tasks, are better at identifying deficits after a concussion (25,26,31,37). Unlike quiet stance, gait tasks are less reliant on sensory feedback, because they rely more on feedforward control (23). Thus, TG may be a suitable, alternative task for clinicians to elucidate impairments in dynamic postural control.
The overall psychometric properties of TG herein exceeded those of both BESS and mBESS. This is the first study to determine an MDC threshold (0.38 s) for TG. Roughly two thirds (24/38) of concussion participants performed worse on TG when tested within 48 h of concussion, and of these 24, 96% (22/24) exceeded the MDC. Ninety-six percent of the student-athletes who performed worse on TG when tested within 48 h of concussion exceeded this threshold, demonstrating that postinjury change is likely not due to scorer variability. This is in opposition to the MDC threshold for the BESS (9), which was exceed by only 18% of participants who performed worse on BESS within 48 h of concussion. Low sensitivity, even acutely after injury, is a substantial limitation of the BESS (5). Herein, we reported the BESS to have a sensitivity of 0.45 acutely (≤48 h) after concussion, which was marginally higher than 0.34 at the time of injury or 0.24 at 48 h after injury as reported by McCrea et al. (5). Both of our BESS and mBESS sensitivities (0.45 and 0.47) were lower acutely after injury than those reported by Buckley et al. (16) (0.60 and 0.71); however, Buckley et al. defined acute as within 24 h after concussion, whereas we extended our acute window to 48 h after injury. The BESS and mBESS sensitivities were both still lower than the sensitivity of the TG (0.63). In the ROC analysis, the AUC values for TG, BESS, and mBESS were 0.704, 0.508, and 0.535, respectively, which correlate with an acceptable accuracy for TG and a failure for both BESS and mBESS (35). Furthermore, TG demonstrates similar but higher sensitivity than the Sensory Organization Test (0.61), which is an instrumented measure of static postural control (38). Unlike TG, the Sensory Organization Test has limited clinically feasible given the high costs, time, and expertise to administer.
The TG test was performed in a quiet environment within 48 h of concussion. Thus, the sideline applicability should not be extrapolated from this study. In addition, the participants herein were collegiate student-athletes and extrapolation to other populations should be limited. We could not control for any injury occurrence between preseason baseline testing and postconcussion testing. Furthermore, the addition of a cognitive task has been recently explored and we believe should continue to be explored. The testing time points differed between the control and concussion groups; however, we do not believe that testing time influenced testing performance, because it was dictated by the concussions, and gait in a healthy college student should be consistent across time. The participants’ foot size, which was not included in this investigation, could be a potential influencing factor in the time to complete TG; however, our within-subjects approach allowed for consistency across anthropometric metrics between time points.
Despite receiving limited attention throughout the concussion literature, TG times were significantly slower after concussion, relative to baseline, with psychometric outcomes (AUC = 0.704) exceeding those of the BESS and mBESS. There were no significant interactions for either the BESS or mBESS acutely after concussion, and the AUC values (0.51 and 0.54, respectively) were classified as “failure” (35). Importantly, the control participants’ performance remained highly stable across time, indicating the lack of a significant practice effect. Although the BESS and mBESS remain more common assessments of postconcussion postural control, the psychometric properties, including MDC, sensitivity, and specificity, were poor compared with that of the TG within our sample of athletes tested acutely after concussion. Thus, TG should be viewed as a clinically useful assessment of postural control after concussion among collegiate student-athletes.
This study is supported by the Grand Alliance Concussion Assessment, Research, and Education Consortium Clinical Study Core, which is jointly funded by the National Collegiate Athletic Association and the Department of Defense. There are no conflicts of interest associated with the authors of this study. The results of the present study do not constitute endorsement by the American College of Sports Medicine and are presented without fabrication, falsification, or data manipulation.
1. McCrory P, Meeuwisse W, Dvorak J, et al. Consensus statement on concussion in sport—the 5th International Conference on Concussion In Sport held in Berlin, October 2016. Br J Sport Med
2. Echemendia RJ, Meeuwisse W, McCrory P, et al. The Sport Concussion Assessment Tool 5th Edition (SCAT5). Br J Sport Med
3. Sport Concussion Assessment Tool—3rd Edition. Br J Sports Med
4. Buckley T, Oldham J, Caccese J. Postural control deficits identify lingering post-concussion neurological deficits. J Sport Heal Sci
5. McCrea M, Barr WB, Guskiewicz K, et al. Standard regression-based methods for measuring recovery after sport-related concussion. J Int Neuropsychol Soc
6. Kelly KC, Jordan EM, Joyner AB, Burdette GT, Buckley TA. National Collegiate Athletic Association Division I athletic trainers’ concussion-management practice patterns. J Athl Train
7. Buckley T, Burdette G, Kelly K. Concussion-management practice patterns of National Collegiate Athletic Association Division II and III athletic trainers: how the other half lives. J Athl Train
8. Buckley T, Munkasy B, Clouse B. Sensitivity and specificity of the modified balance error scoring system in concussed collegiate student athletes. Clin J Sport Med
9. Finnoff JT, Peterson VJ, Hollman JH, Smith J. Intrarater and interrater reliability of the Balance Error Scoring System (BESS). PM R
10. Valovich McLeod T, Barr WB, McCrea M, Guskiewicz KM. Psychometric and measurement properties of concussion assessment tools in youth sports. J Athl Train
11. Burk J, Joyner A, Munkasy B, Buckley T. Balance Error Scoring System performance changes after a competitive athletic season. Clin J Sport Med
12. Valovich TC, Perrin DH, Gansneder BM. Repeat administration elicits a practice effect with the Balance Error Scoring System but not with the Standardized Assessment of Concussion in high school athletes. J Athl Train
13. Docherty C, Valovich T, Shultz S. Postural control deficits in participants with functional ankle instability as measured by the Balance Error Scoring System. Clin J Sport Med
14. Fox ZG, Mihalik JP, Blackburn JT, Battaglini CL, Guskiewicz KM. Return of postural control to baseline after anaerobic and aerobic exercise protocols. J Athl Train
15. Rahn C, Munkasy B, Joyner B, Buckley T. Sideline performance of the Balance Error Scoring System during a live sporting event. Clin J Sport Med
16. Schneiders AG, Sullivan SJ, Gray AR, Hammond-Tooke GD, McCrory PR. Normative values for three clinical measures of motor performance used in the neurological assessment of sports concussion. J Sci Med Sport
17. Schneiders AG, Sullivan SJ, McCrory PR, et al. The effect of exercise on motor performance tasks used in the neurological assessment of sports-related concussion. Br J Sports Med
18. Oldham JR, DiFabio MS, Kaminski TW, DeWolf RM, Buckley TA. Normative tandem gait in collegiate student-athletes: implications for clinical concussion assessment. Sports Health
19. Howell D, Osternig L, Chou L. Single-task and dual-task tandem gait test performance after concussion. J Sci Med Sport
20. Giorgetti M, Harris B, Jette A. Reliability of clinical balance outcome measures in the elderly. Physiother Res Int
21. Kammerlind A, Larsson P, Ledin T, Skargren E. Reliability of clinical balance tests and subjective ratings in dizziness and disequilibrium. Adv Physiol Educ
22. Shumway-Cook A, Woollacott M. Motor Control Theory and Practical Applications
. 2nd ed. Biblis M, editor. Lippincott Williams & Wilkins; 2007. p. 74–5.
23. Nelson L, Loman M, LaRoche A, Furger R, McCrea M. Baseline performance and psychometric properties of the child sport concussion assessment tool 3 in 5 to 13 year old athletes. Clin J Sport Med
24. Hänninen T, Tuominen M, Parkkari J, et al. Sport Concussion Assessment Tool—3rd edition—normative reference values for professional ice hockey players. J Sci Med Sport
25. Oldham JR, Munkasy BA, Evans KM, Wikstrom EA, Buckley TA. Altered dynamic postural control during gait termination following concussion. Gait Posture
26. Buckley TA, Oldham JR, Munkasy BA, Evans KM. Decreased anticipatory postural adjustments during gait initiation acutely postconcussion. Arch Phys Med Rehabil
27. Howell D, Osternig L, Chou L. Monitoring recovery of gait balance control following concussion using an accelerometer. J Biomech
28. Howell DR, Osternig LR, Koester MC, Chou L-S. The effect of cognitive task complexity on gait stability in adolescents following concussion. Exp Brain Res
29. Howell D, Osternig L, Chou L. Dual-task effect on gait balance control in adolescents with concussion. Arch Phys Med Rehabil
30. Howell DR, Osternig LR, Chou L-S. Adolescents demonstrate greater gait balance control deficits after concussion than young adults. Am J Sports Med
31. Parker TM, Osternig LR, Van Donkelaar P, Chou LS. Gait stability following concussion. Med Sci Sports Exerc
32. McCrory P, Meeuwisse WH, Aubry M, Cantu B, Dvorak J, Echemendia RJ. Consensus statement on concussion in sport: the 4th International Conference on Concussion in Sport held in Zurich, November 2012. Clin J Sport Med
33. Broglio SP, McCrea M, McAllister T, et al. A national study on the effects of concussion in collegiate athletes and US military service academy members: the NCAA–DoD Concussion Assessment, Research and Education (CARE) Consortium Structure and Methods. Sports Med
34. Riemann BL, Guskiewicz KM, Shields EW. Relationship between clinical and forceplate measures of postural stability. J Sport Rehabil
35. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology
36. McCrea M, Guskiewicz KM, Marshall SW, et al. Acute effects and recovery time following concussion in collegiate football players: the NCAA concussion study. JAMA
37. Chou L, Kaufman KR, Walker-Rabatin AE, Brey RH, Basford JR. Dynamic instability during obstacle crossing following traumatic brain injury. Gait Posture
38. Broglio SP, Macciocchi SN, Ferrara MS. Sensitivity of the concussion assessment battery. Neurosurgery