Neuromuscular training after concussion to improve motor and psychosocial outcomes: a feasibility trial

Abstract

Objective:

To determine the feasibility of an 8-week neuromuscular training program initiated upon return-to-play clearance following concussion.

Design:

Feasibility trial.

Setting:

A single sports medicine center.

Participants:

We approached n=54 patients; n=32 agreed to participate (59%). N=27 participants returned for their second visit at return-to-play clearance (84%) and were randomized to neuromuscular training (n=13) or standard-of-care (n=14).

Main outcome measures:

Participants completed three assessments: within 14 days post-concussion, immediately after return-to-play clearance, and 8-weeks following return-to-play clearance. The intervention aimed to achieve positive neuromuscular adaptations and occurred 2x/week for 8 weeks under supervision.

Results:

N=2 participants randomized to the intervention elected not to participate, both due to schedule conflicts (e.g., time required to meet with the study team). Participants began the intervention an average of 11 days after return-to-play clearance, the majority (91%) completed >75% of training sessions, and training sessions lasted an average of 18.2±4.8 minutes. One participant stopped the intervention after 7 training sessions due to time availability.

Conclusion:

It is feasible to initiate a neuromuscular training program for most athletes shortly after returning to play following concussion. Clinicians and researchers may consider this approach to mitigate the increased musculoskeletal injury risk for concussion patients returning to sports.

Introduction

While concussions can result in negative short- and long-term consequences, few established rehabilitation protocols have demonstrated efficacy to improve outcomes (1). Beyond recent trials evaluating early aerobic exercise after a concussion (2, 3), a relatively weak body of concussion rehabilitation literature exists, where studies have reported results without control groups, without randomization, or without control for confounders (1). There is a need to develop post-concussion interventions and examine their ability to improve short- and long-term post-concussion.

Over the past several years, authors of epidemiologic studies have reported an increased lower extremity musculoskeletal (MSK) injury risk during the year after concussion among adolescent, collegiate, and elite athletes, and among military populations (4). In fact, a meta-analysis indicated athletes who sustain a concussion have significantly higher odds of a subsequent MSK injury than non-concussed control athletes (36% vs. 18% injury rate, odds ratio=2.11), and this risk exists for both female and male athletes (5). Based on the sequence of prevention model (6), the incidence of a problem (concussion and subsequent sport-related injury risk) has been described (4, 5). Establishing mechanisms and introducing preventative measures are the necessary next steps.

Currently, we know very little about the mechanisms responsible for post-concussion MSK injury risk. One hypothesis is that despite clinical recovery, athletes cleared to return-to-play (RTP) following concussion, possess residual neuromuscular and/or attentional impairments that are undetectable via traditional concussion tests (e.g. symptoms, balance, neurocognitive tests) (7, 8). Thus, athletes return to sports prior to full neurologic recovery and are vulnerable to further injury. Dual-task gait (e.g. simultaneous motor and cognitive task execution) probes neuromuscular control and attention and can identify otherwise unnoticed post-concussion impairments (9, 10). In fact, dual-task performance has been associated with post-concussion subsequent sport-related injuries in two separate studies (11, 12). Therefore, the potential association between altered neuromuscular function and increased MSK injury risk after concussion (13) highlights neuromuscular training (NMT) as a matter necessitating further exploration, specifically given its efficacy within other injuries (e.g. anterior cruciate ligament tears) (14, 15). However, no preventive MSK injury risk mitigation strategies currently exist for clinicians to use after concussion.

To reduce sport-related injury risk independent of a concussion, researchers have implemented injury prevention programs (16) that integrate plyometrics, strengthening, technique training, and balance training to gain neuromuscular system adaptations (17). Specifically, researchers have reported NMT programs can reduce injury rates among young female athletes (18), and reduce ankle sprain injuries in youth soccer and basketball athletes (19). As such, we hypothesize a similar NMT approach may be effective in reducing post-concussion MSK injury rates. Yet, no evidence exploring the feasibility of this approach exists, and there is no uniform guidance related to MSK injury risk reduction after concussion.

Feasibility trials play a critical role in developing interventions that have large-scale impact in the future, and focus on the practicality of performing interventions (20, 21). Thus, the primary purpose of our trial was to determine the feasibility of conducting an 8-week NMT program initiated after athletes are cleared to RTP following a sport-related concussion. We aimed to identify access to participants, barriers to participation, feasibility of assessment procedures, time required, compliance with the intervention, and accessibility to treatment. Our secondary purpose was to describe the feasibility of obtaining a multimodal set of patient-reported and clinician-administered outcome measures, as well as describe the characteristics of the intervention group relative to a standard-of-care (non-intervention) group.

Methods

Trial Design

We conducted a feasibility trial of a NMT program compared to standard-of-care (allocation ratio, 1:1). Participants completed 3 assessments: Visit 1 (≤14 days post-concussion), Visit 2 (after return-to-play clearance was provided by their physician), and Visit 3 (8 weeks after Visit 2). During Visit 1, we recruited, enrolled, and initially assessed participants (Figure 1). Participants followed the recommendations of their physician independent of the study, and were cleared for RTP according to their individual progression. No study interventions were performed between the Visit 1 and Visit 2. Once RTP clearance was received from the treating physician, participants were asked to return for Visit 2. At this visit, participants were randomized to either a neuromuscular training (intervention) or standard-of-care (no intervention) group. As the intervention lasted 8 weeks, all participants regardless of group assignment returned for a final assessment 8 weeks after Visit 2 (Figure 1). Beginning at Visit 1, and in a consistent fashion at Visits 2 and 3, we collected patient-reported (Post-Concussion Symptom Inventory, PROMIS Global Pediatric v25, and Tampa Scale of Kinesiophobia) and clinician-administered (reaction time, single/dual-task gait, balance) outcomes, described below.

Figure 1.

Study flow diagram of the feasibility trial.

Participants

We recruited participants who were seen and cared for at a single sports medicine center between April 23, 2019–March 6, 2020. Inclusion criteria included enrollment within 14 days of injury, being 12–18 years of age, diagnosis of a concussion by a sports medicine physician, and a post-concussion symptom inventory score ≥9 during the first assessment to ensure all participants had not recovered by the time they enrolled in the study. Patients who had recovered by the time of the first assessment were not included so that our sample had a relatively homogenous profile upon enrollment, to the degree possible. We excluded participants if they had a co-existing lower extremity injury affecting balance, had a previous concussion within the past year, a pre-existing learning disability, documented structural brain injury via neuroimaging, sustained a concussion during a high velocity mechanism (e.g. motor vehicle accident), or if they did not intend to return to sports after recovering from their concussion. Board-certified sports medicine physicians made concussion diagnoses and RTP decisions independent of the study. Concussion diagnosis and return-to-play decision making decisions were made consistent with the international consensus statement on concussion in sport (22). For example, participants completed the 6-stage graduated return-to-sport progression as outlined in this statement prior to receiving RTP clearance. The study protocol was reviewed and approved by the institutional review board. All participants, and parents/guardians, if participants were under the age of 18, provided written informed consent/assent. The study protocol was pre-registered via clinicaltrials.gov (NCT03917290)

Intervention

Beginning after RTP clearance, participants randomized to the intervention group completed an 8-week NMT training program. They met 2x/week with an athletic trainer who supervised all exercises, provided instruction and feedback on form, and documented compliance. The intervention was based on existing NMT programs that use plyometric, strengthening, technique, and/or balance training to achieve neuromuscular system adaptations and reduce sports injury risk (18, 23). In addition, we added a dual-task element (single exercise at each visit) to the program, given the theorized relationship between dual-task function and musculoskeletal injury risk after concussion (11). The specific program elements and volumes are provided in Supplementary Table 1. Participants randomized to standard-of-care received no instruction beyond the RTP clearance instructions provided by their physician.

The majority (n=9/11) of participants met with an athletic trainer in person to complete the NMT program. As n=2 participants were cleared to return-to-sports in March 2020 (start of cessation of in-person clinical research activities due to COVID-19), they completed interventions via telehealth. The program elements did not change, however, rather than meeting in person, instruction was provided via video feed and real-time interaction to maintain patient and researcher safety. If necessary, comparable household objects were substituted for equipment as needed for study visits conducted via telehealth.

Primary Endpoints

Our primary endpoint was approximately 8 weeks after RTP clearance for all participants. This timepoint coincided with completion of the intervention (Figure 1). During this visit, we obtained a multimodal set of outcome measures.

Feasibility Data

To determine the feasibility of our protocol, we obtained variables pertaining to treatment delivery, treatment acceptability, participant adherence, participant retention, and completeness of outcomes data. To examine treatment delivery and acceptability, we assessed the number of participants randomized to the intervention who agreed to participate and the amount of time spent with the study team member completing training. We examined participant adherence and retention by examining the duration of time they continued the intervention for, and the number of training sessions they completed. Furthermore, participants completed a battery of patient-reported and clinician-administered assessments at each study visit in order to understand completeness of outcomes data.

Patient-Reported Outcomes

To assess concussion symptoms, participants completed the Post-Concussion Symptom Inventory (PCSI). The PSCI has been developed and validated specifically within the youth and adolescent population, and contains strong internal consistency and test-retest reliability (24). Participants responded to 26 instrument items by rating each on a 7-point Guttman scale indicating severity from 0 (none) to 6 (severe). We calculated total PCSI score as the sum of all responses. We defined symptom resolution as a severity score of 0. During clinical visits, participants rated the severity of those symptoms that started at the time of injury and that they had experienced within the prior 24 hours of the clinical examination in order to account for non-concussion-related symptoms.

Participants completed the Tampa Scale of Kinesiophobia to identify fear of pain with movement. The Tampa Scale of Kinesiophobia (TSK) is a reliable and valid outcome measure capable of measuring attitudes regarding fear of pain with movement (25). Previously it has been used to examine changes in fear of re-injury throughout concussion recovery (26). We selected this assessment so that we could examine whether the intervention provided a beneficial effect regarding self-perceived readiness to return to sports. Participants responded to a battery of 17 questions, ranked on a scale from 1 (strongly disagree) to 4 (strongly agree). We calculated a total score as the sum of all items, using inversion scoring for four questions.

To quantify our psychosocial outcomes, we used the Patient-Reported Outcomes Measurement Information System (PROMIS) v 1.1 Pediatric Profile 25, previously validated among children ages 8 through 17 (27). We assessed the specific domains of mobility/physical function, anxiety, depressive symptoms, and peer relationships using the pen/paper short forms, which have shown clinically acceptable test-retest reliability (27). Each question was answered based on the athlete’s experience in the past seven days and consisted of a five point Likert scale, where a response of 0 indicated “never” and a response of 4 indicated “almost always.”

As described above, some participants completed interventions remotely. In addition, participants who were enrolled when in-person research restrictions went into place completed the patient-reported outcome questionnaires remotely as well. This was done via a Research Electronic Data Capture (REDCap) online survey (28). Outcomes obtained using this approach occurred at the same intervals as in-person.

Clinician-Administered outcomes

Participants completed three different postural control tests: an instrumented single/dual-task gait analysis, the modified Balance Error Scoring System (mBESS) test, and a single/dual-task tandem gait test. During the gait analysis, participants completed three single-task and three dual-task trials while wearing a smartphone (Samsung Galaxy S8) affixed to the lumbar spine with a running belt. Participants walked approximately 10 m toward a target placed in front of them, walked around it, and returned to the original starting position. The mean of the three walking trials in both conditions were used in further analyses. During single-task trials, participants walked down and back without any attention divided to other tasks. During dual-task trials, they concurrently completed a cognitive test: spelling a five-letter word backwards, reciting the months in reverse order, or serially subtracting by sixes or sevens (11, 29, 30). We used an Android application to calculate gait outcomes (Gait Analyzer, Control One LLC, Albuquerque, NM, USA, Version 0.9). This experimental setup was used previously to collect accelerometer data from the phone hardware sensors and analyze gait data in an automated process (29). Detailed methodology, reliability, and validity data for this application have been reported previously (30).

Participants completed the mBESS, where they stood for 20 s on a solid surface with their eyes closed in three positions: double-leg stance, single-leg stance, and tandem stance (31). During double-leg stance, participants stood on both feet placed side-by-side. During single-leg stance, participants stood on the foot that they identified as their non-dominant kicking leg, and during tandem stance they stood with their feet positioned where the non-dominant foot was placed directly behind the dominant foot. Participants were instructed to remain with their eyes closed and hands on their hips; errors were counted if the participant opened their eyes, lifted their hands off their hips, took a step, fell out of the testing position, lifted any portion the foot off the ground, abducted the hip greater than 30 degrees, or stayed out of the test position for 5 seconds or more.

We also gathered reaction time data using two methods: a smartphone simple reaction time test and a clinical-reaction time test (i.e. drop-stick test). During the smartphone reaction time test, a smartphone (Samsung Galaxy S8) was placed on a flat table at a 2 cm standardized distance in front of the participants’ dominant hand. A red dot appeared in the middle of the screen. At various points, the dot turned green, and the participant clicked the dot as fast as possible. The primary outcome variable was average response time, calculated from 30 successful trials. For the drop-stick clinical-reaction time test, participants placed their forearm on a table while seated, such that the stick portion of the device dropped between their thumb and finger, as previously described (32). Participants completed 8 successful drop-responses after two practice drops (32).

Sample size and randomization

Before our study began, we determined our sample size based on an injury prevention program to assess functional changes in youth athletes across a one month intervention (33). Within this study, the pre-post intervention change for drop jump performance was higher in the intervention group relative to the no intervention group (mean= 1.8±3.7 cm vs. −1.7±3.0 cm; Cohen’s d=1.05). Given an alpha of 0.05 and power of 0.80, these results suggest that enrolling 32 participants would be necessary to achieve our feasibility aims.

Participants were randomized using a block stratification procedure. We used a block size of 4, and stratification factors included age (12–15 vs. 16–18 years) and sex (34). Given the nature of the intervention, no blinding was possible or meaningful. The lead author generated the random allocation sequence, and investigative team members enrolled participants and assigned them to each group.

Statistical Analysis

Data are presented as means (standard deviations) or medians (interquartile range) for continuous outcome variables, and as the number included (corresponding percentage) within each group for categorical outcome variables. To ensure our randomization procedure achieved its intended effect, we compared the demographic and clinical characteristics of each group using independent samples t-tests or Fisher’s exact tests. We then examined the descriptive characteristics of the intervention and our ability to deliver treatment, treatment acceptability, adherence, retention, and completeness of outcomes data (20) using summary statistics. Given that statistical comparisons between intervention and standard-of-care groups are likely misleading in the context of a feasibility trial (20), we did not perform any statistical comparisons between groups aimed to evaluate trial efficacy on patient-reported or clinician-obtained variables.

Results

Participant characteristics

We approached n=54 patients to participate in the study during the recruitment period; n=32 consented to participate (59%). There were no differences in the proportion of girls (41% vs. 65%; p=0.10) or in the mean age (15.2±1.7 vs. 15.3±1.5; p=0.87) of those who enrolled and those who did not. There were no significant demographic or clinical characteristic differences between the intervention and standard-of-care groups (Table 1). Of the 32 participants (Figure 2), n=27 returned for post-injury test 2 (84%) and underwent randomization to the intervention NMT program (n=13) or standard-of-care (n=14). Of those randomized to intervention, n=2 did not participate in the program due to scheduling conflicts (i.e. the participant/family could not/did not want to dedicate the time needed to meet with an athletic trainer 2x/week for 8 weeks). We did not observe any adverse events in either participant groups during the intervention period.

Table 1.

Demographic and clinical characteristics between intervention and standard-of-care groups.

Variable	Neuromuscular training intervention (n=11)	Standard-of-care (n=14)	P value
Age (years)	14.7 (1.7)	14.8 (1.5)	0.88
Sex (female)	4 (36%)	6 (43%)	> 0.99
Concussion history	5 (45%)	10 (71%)	0.24
Time of Study visit 1 (days post-injury)	9 (3)	7 (3)	0.11
Time of Study visit 2 (days post-injury)	35 (25)	41 (43)	0.33
Time of study visit 3 (days post-injury)	107 (25)	108 (44)	0.96
Symptom resolution time (days post-injury)	28 (22)	32 (30)	0.70
Height (cm)	166 (11)	167 (10)	0.82
Weight (kg)	65 (24)	59 (12)	0.43
Race	White: 6 (55%) Hispanic or Latino: 3 (27%) American Indian or Alaska Native: 1 (9%) Unknown/not reported: 1 (9%)	White: 9 (64%) Black or African American: 2 (14%) More than one race: 2 (14%) Hispanic or Latino: 1 (7%)	0.18
History of headaches	3 (27%)	4 (29%)	> 0.99
History of migraines	3 (27%)	1 (7%)	0.29
ADD or ADHD diagnosis	0 (0%)	1 (7%)	> 0.99
Pre-injury anxiety diagnosis	2 (18%)	2 (14%)	> 0.99
Pre-injury depression diagnosis	2 (18%)	3 (21%)	> 0.99
LOC at time of injury	3 (27%)	2 (14%)	0.62
Amnesia at time of injury	6 (55%)	3 (21%)	0.12
Adverse events	0 (0%)	0 (0%)	> 0.99

Figure 2.

CONSORT (Consolidated Standards of Reporting Trials) diagram.

The median time to symptom resolution was 17 days [interquartile range (IQR)= 12, 43] for the intervention group and 19 days [IQR=11, 49] for the standard-of-care group. The median time to medical clearance to RTP for the intervention group was 22 [IQR= 15, 49] days, and 35 [IQR=15, 62] days for the standard-of-care group.

Intervention Compliance and Retention

The majority of intervention group participants (55%) were able to complete all 16 training sessions (n=4 in-person, n=2 via remote instruction). The remaining intervention participants were able to complete ≥75% of the NMT sessions, with the exception of one participant who only completed 44% of the sessions (Table 2). This participant stopped participation due to time availability to meet with the athletic trainer 2x/week after completing 7 of the 16 training sessions. Within each session, all participants completed all exercises for that session, and each session required approximately 18 minutes to complete (excluding warm-up and cool-down; Table 2). On average, participants began the NMT intervention 11 days following RTP clearance from their physician.

Table 2.

Feasibility outcomes and intervention characteristics for participants randomized to the neuromuscular training intervention.

Intervention characteristics (n=11)	Mean (SD)	Range
Time from RTP clearance to first intervention session (days)	11.0 (7.6)	1 – 28
Mean duration of the intervention (days elapsed first to last session)	58.8 (13.4)	44 – 93
Mean sessions completed (maximum = 16)	14.1 (2.8)	7 – 16
Average time per NMT session (minutes)	18.2 (4.8)	14 – 30
	Sessions completed
Percentage of total intervention sessions completed (maximum=16)	100%: n=6 88%: n=1 81%: n=2 75%: n=1 44%: n=1

Outcome data compliance

All participants completed all patient-reported and clinician-administered outcomes in-person at the initial study visit (Table 3). 100% of participants completed the patient-reported outcome assessment battery at the study visit 2 (pre-intervention) and study visit 3 (post-intervention) assessment points (Table 3). Due to the restriction of in-person clinical research beginning in March 2020 as a result of the COVID-19 pandemic, 4% (n=1) of participants responded to the patient-reported outcome questionnaires via online survey rather than pen/paper at study visit 2, and 26% at study visit 3 (n=7).

Table 3.

Outcomes obtained at each time point in intervention and standard-of-care groups. Data are presented as the number of participants (% within group) who completed each assessment.

Patient-Reported Outcome	Intervention Group (n=11)				Standard-of-Care Group (n=14)
	Visit 1	Visit 2	Visit 3		Visit 1	Visit 2	Visit 3
Post-Concussion Symptom Inventory	11 (100%)	11 (100%)	11 (100%)		14 (100%)	14 (100%)	14 (100%)
Tampa Scale of Kinesiophobia	11 (100%)	11 (100%)	11 (100%)		14 (100%)	14 (100%)	14 (100%)
Patient-Reported Outcomes Measurement Information System Pediatric Profile 25	11 (100%)	11 (100%)	11 (100%)		14 (100%)	14 (100%)	14 (100%)
Clinician-administered assessment	Visit 1	Visit 2	Visit 3		Visit 1	Visit 2	Visit 3
Single/dual-task gait analysis	11 (100%)	11 (100%)	9 (82%)		14 (100%)	13 (93%)	9 (64%)
Modified Balance Error Scoring System	11 (100%)	11 (100%)	11 (100%)		14 (100%)	14 (100%)	10 (71%)
Single/dual-task tandem gait test	11 (100%)	11 (100%)	11 (100%)		14 (100%)	14 (100%)	10 (71%)
Simple reaction time	11 (100%)	11 (100%)	10 (91%)		14 (100%)	13 (93%)	10 (71%)
Clinical reaction time	11 (100%)	11 (100%)	9 (82%)		14 (100%)	13 (93%)	10 (71%)

In a similar fashion as the patient-reported outcomes, some participants could not complete the clinician-administered assessment battery at study visits 2 and 3 due to in-person research restrictions. For single-task gait, dual-task gait, and both reaction time assessments, 4% (n=1) of participants were unable to complete at study visit 2. We were able to collect mBESS and tandem gait data at study visit 2 for all participants, as the missing data was collected as a part of routine standard-of-care clinical assessment for the mBESS at the RTP visit. At study visit 3, a small proportion of participants in each group were unable to complete the clinician-administered assessments (Table 3).

Discussion

The results of our study indicate that incorporating a NMT program under supervision of an athletic trainer after post-concussion RTP clearance is feasible, with minimal loss to follow-up. The premise of this approach is to incorporate sport-like exercises with a previously demonstrated prophylactic effect to reduce sport-related injuries (18) among a population vulnerable to lower extremity musculoskeletal injuries (5). Currently, these sorts of exercises are not included as a typical RTP protocol component following concussion (22). Given the feasibility of this approach observed within our investigation, efficacy trials are the next logical step in determining the role of NMT exercises within standard-of-care treatment for athletes who have sustained a concussion and wish to return to sports. Beyond enrolling athletes into these programs and monitoring intervention compliance, further prospective monitoring studies are needed to determine if this approach reduces the proportion of those who are injured, or the severity of subsequent injuries sustained in the year after a concussion compared to those who did not undergo any sort of formal NMT program.

There is a need to investigate treatment strategies for concussion sequelae using high-quality designs that account for potential confounding factors and use randomization procedures to best determine treatment effects (22). Our feasibility trial demonstrates that our randomization procedure was effective as intended and expected, as no group demographic or medical injury history characteristics were identified. In addition, among those randomized to the intervention, compliance was higher than previous studies (35–37). This may have been due to the requirement to meet with an athletic trainer, rather than performing the exercises at a self-guided pace like these previous studies. While beneficial for compliance, this also creates an additional time burden on both the patient and clinician. Within the framework of sports medicine clinical trials, compliance is not routinely reported (38) and thus should be an area of focus among researchers. While the majority of intervention group participants completed all training sessions in our trial, 45% did not. The reason for lack of perfect compliance is likely multifactorial. The time burden placed on patients within this approach (~40 minutes/week for 8 weeks) is less than other rehabilitation interventions (39), although patient populations, intent, and healthcare provider delivery vary substantially in this past work compared to our study. Furthermore, time demands and lack of transportation resulted in two participants electing to drop out of the study upon randomization.

Rehabilitation interventions after concussion require proper supervision to mitigate poor movement patterns during exercise, and to ensure compliance with the protocol (32). Given patient compliance variability, unknown effects within an investigation may occur if not documented (33). In addition, within a clinical trial framework, compliance variability may dilute the actual intervention effect (34). As such, feasibility trials can help inform researchers and clinicians about the difficulties encountered within such an endeavor and serve to provide the basis for larger scale RCTs (20). Related to compliance, other external factors may affect the feasibility of a trial. In our case, two participants had to switch from in-person to a remote-based intervention given the inability to interact with patients in person beginning in March 2020. While not ideal, this suggests telehealth intervention approaches may allow for adherence observation, real-time visual and verbal feedback, and an avenue of future protocol scalability, although our feasibility trial was not designed to make inferences regarding pre-post intervention changes or differences between intervention delivery type. Furthermore, clinicians should consider the timing of intervention initiation. We began after return-to-play clearance was provided so that there were no activity restrictions for study participants. However, initiating NMT exercises earlier post-concussion may yield positive effects on recovery, and should be considered in future trials.

In order to understand effective outcome measures for RCTs within the scope of MSK injury after concussion, there is a need to move to outcomes beyond current concussion clinical assessment. Specifically, commonly used tests are not able to successfully identify those who may or may not be at risk for MSK injury after concussion (40). Given the lack of robust experimental data regarding the mechanisms driving this increased injury risk, we do not currently understand what outcome measures can best assess intervention efficacy. However, several theories have been postulated, in particular that measures of dual-task function (11), neuromechanical responsiveness (13), or perception-action coupling (41) may provide useful information related to post-concussion MSK injury risk. As such, our multimodal evaluation included elements of dual-task function, perception of symptoms, psychosocial characteristics, and reaction time. Our feasibility data suggest there were few missing data points, and reasons were often outside the control of the investigative team (i.e. a global pandemic). Due to the nature of a feasibility trial, we are unable to make comparisons between group at pre- and post-intervention time points (20), however, the ability to collect these outcomes is feasible within the context of our study with minimal missing data. Other outcomes for consideration in future trials include measures of attention (42) or more sport-specific tasks, such as jump landing (43).

In the long-term, this work serves as the basis for future trials and potential incorporation into clinical practice given the feasibility of integrating relatively simple training and supervision for athletes who have demonstrated clinical recovery after a concussion. Beyond the time and potential transportation limitations, we encountered few risks or problems with this added training done in addition to organized athletic activities. The efficacy of this approach, however, requires further investigation. Specifically, the dosage required to induce a change within an 8-week program or if a dose-response relationship exists require further study. However, we exceeded the volume noted by past meta- and sub-group analyses, which suggested that if NMT programs are performed for more than 30 minutes per week during an athletic season, approximately 70% of ACL injuries may be avoided (14).

Study Limitations

The nature of our feasibility trial inherently resulted in several limitations. Namely, we cannot infer meaning from any observed differences between intervention and standard-of-care groups, or across time. In addition, the participants in our study were recruited from a single specialty care sports medicine clinic. Thus, they may represent a more severely injured sample than the general population. Further, the feasibility of our approach is specific to this setting and may not be generalizable to other settings. Finally, we included adolescent athletes in our cohort only, and as such, we cannot extrapolate the feasibility of this approach to other age groups, such as collegiate athletes. Further research is required to understand the relative effect of this approach in actually reducing the number of injuries sustained after post-concussion RTP clearance. Longitudinal monitoring for up to one year after injury among intervention and standard-of-care groups will enhance the ability to examine the utility of this approach beyond its feasibility. Finally, we did not record patient or clinician satisfaction with our approach.

Conclusion

It is feasible to identify, recruit, enroll, and perform a neuromuscular training program for athletes who have returned to play following a concussion. Further work to determine efficacy in improving outcomes and reducing risk of subsequent MSK injury after concussion RTP is required. However, given the low-risk nature of the training after clinical concussion recovery, clinicians may consider neuromuscular training as a method to incorporate as an element of concussion rehabilitation.