Day of injury assessment of sport-related concussion
- 1Departments of Neurosurgery and Neurology, Medical College of Wisconsin, Milwaukee, Wisconsin, USA
- 2Department of Psychiatry, University of British Columbia, Vancouver, British Columbia, Canada
- 3University of Missouri-Kansas City, Psychological and Neurobehavioral Associates, Inc., State College, Pennsylvania, USA
- 4The Florey Institute of Neuroscience and Mental Health, Melbourne Brain Centre, Melbourne, Australia
- 5International Rugby Board, New South Wales, Australia
- Correspondence to Dr Michael McCrea, Departments of Neurosurgery and Neurology, Medical College of Wisconsin, Milwaukee, WI 53226, USA;
- Received 3 January 2013
- Accepted 19 January 2013
Objective To conduct a critical review of the literature on instruments currently used in the assessment of sport-related concussion on the day of injury.
Data sources Computerised searches of the literature posted to MEDLINE, PubMed, CINAHL, PsychInfo and Cochrane Library from 1 January 1982 through 21 August 2012. Key words and medical embedded subheadings (MeSH) terms relevant to sport-related concussion were applied, which identified 577 articles.
Study selection In addition to MeSH term and key word criteria, a study was included in the analysis if the article: (1) was published in English, (2) represented original research, (3) pertained to sport-related concussion (ie, not non-sports traumatic brain injury), (4) included assessment or diagnostic data collected within 24 h of the injury event and (5) involved human research. A total of 41 studies qualified for review.
Data extraction All articles were examined to determine if the study met the additional requirements for inclusion. A standardised method was used to document critical elements of the study design, population, tests employed and key findings.
Data synthesis A large number of studies were analysed that reported data from testing conducted within 24 h of injury. These studies collectively demonstrated that a number of instruments are capable of measuring the acute effects of concussion across several domains, such as symptoms, cognition and balance.
Results Relating to specific assessment domains are compiled in separate tables and an interpretive summary of the findings is provided.
Conclusions Several well-validated tests are appropriate for use in the assessment of acute concussion in the competitive sporting environment. These tests provide important data on the symptoms and functional impairments that clinicians can incorporate into their diagnostic formulation, but they should not solely be used to diagnose concussion.
Concussion is often considered by clinicians to be among the most complex injuries in sports medicine to diagnose, assess and manage. There is no perfect diagnostic test or marker for an immediate diagnosis of concussion in the sporting environment. Rather, it is a clinical diagnosis based largely on the observed injury mechanism, signs and symptoms. A vast majority of sport-related concussions (hereafter referred to as concussion) occur without loss of consciousness or frank neurological signs.1–4 In milder forms of concussion, the athlete might be slightly confused, without clearly identifiable retrograde or post-traumatic amnesia. In addition, most concussions cannot be identified or diagnosed by neuroimaging techniques (eg, CT orMRI).
Over the past two decades, there has been a concentrated, international effort towards improved methods for diagnosing concussion. Subject matter experts have attempted to build consensus on the key signs and symptoms critical to the diagnosis, particularly during the acute postinjury period.5 ,6 Sport governing bodies have sponsored global educational campaigns to make athletes, parents, clinicians and other key stakeholders better able to recognise the signs and symptoms that indicate possible concussion.
At the same time, a ‘standardisation movement’ has facilitated the development of several tests designed to aid in concussion assessment and diagnosis. A basic premise of this movement has been that performance-based assessment measures may be superior to an athlete's reporting of symptoms, which might be unreliable due to the athlete's tendency to under-report or fail to recognise their symptoms. Standardised tests are intended to provide a more objective, performance-based method of measuring postinjury recovery and determining an athlete's fitness to return to play. In most cases, these tests assess single domains of concussion effects (eg, postconcussion symptoms, neurocognitive functioning and postural stability). The recently developed Sports Concussion Assessment Tool—Second Edition (SCAT2)7 involves an integrated, multimodal assessment model. In addition to brief screening tests, there has been a proliferation in recent years of computerised neurocognitive tests proposed for concussion assessment.
Standardised tests and measures can be used on the day of injury to assess the acute effects of injury. The purpose of this article is to describe the methods and results of a critical review intended to specifically address whether the existing tests and measures currently in use are sensitive and reliable enough on the day of injury to assist clinicians in accurately evaluating sport-related concussion.
We conducted a systematic survey of the published literature in several computerised databases, including: MEDLINE, PubMed, CINAHL, PsychInfo and Cochrane Library. Based on the focus of this critical review, we limited our search to the published literature over a 30-year period from 1 January 1982 through 21 August 2012. We used a combination of the following key words and medical embedded subheadings (MeSH) terms as parameters for our search:
1. Concuss* (concussion, concussed, concussive, etc), Brain Inj* (brain injury, injuries, injured, etc), Head Inj* (head injury, injuries, injured, etc);
2. Sport*, athlete*, football, soccer, ice hockey, field hockey, boxing, rugby, lacrosse, wrestling;
3. Diagnos*, assess*, test*.
Terms within each group were combined using the Boolean operator ‘OR’; then the two groups were combined using the Boolean ‘AND’ operator. Search statistics were documented with respect to the number of articles generated by the MeSH term and key word search. In addition, the reference lists of acquired articles of relevance were reviewed for other articles not found in the electronic databases.
Study selection criteria
In addition to meeting the MeSH term and key word search criteria, the basic requirements for a study to be included were that the article: (1) was published in English, (2) represented original research, (3) pertained to sport-related concussion (ie, not non-sports traumatic brain injury), (4) included assessment or diagnostic data collected on the day of injury (eg, within 24 h of the injury event) and (5) involved human research (no animal studies). Duplicate articles from more than one electronic database were eliminated.
Data extraction and analysis
Each article identified by the electronic literature search was reviewed to determine if it met the inclusion criteria described above. In some instances, this determination could be made from a review of the abstract. In other cases, it required a review of the full-text article. For the included studies, a standardised data extraction method was used to document the study design, population and assessment techniques (see Tables 1⇓⇓⇓⇓–6). For each study, the reviewer provided a summary of the key findings, statistics and a critical assessment of the evidence relevant to the day of injury assessment. We did not conduct a detailed or systematic analysis of the psychometric properties of the instruments included in this review.
In total, the electronic literature database search identified 577 articles. After applying additional requirements and eliminating duplicate articles, a total of 44 qualified for further review. A majority of articles were rejected because data were not collected within 24 h of the injury. Several studies included assessments across multiple domains (eg, symptoms, cognition and balance) and were included in those individual domains accordingly. Tables 1⇓⇓⇓⇓–6 below provide a summary of key findings from studies in each assessment domain.
Symptom rating scales
Several studies have documented symptoms of concussion immediately following, or over the first 24 h after, injury. These studies usually use a symptom rating scale completed by the athlete. The most commonly and consistently reported acute symptoms, across studies, are headaches, dizziness and some form of mental status disturbance, such as mental clouding, confusion or a slowing down of feelings.8–14 Other commonly reported acute symptoms include visual problems, fatigue and nausea.8 ,11 ,13 ,14 Some observable signs of concussion include a dazed facial expression and unsteady gait.13
Data from 18 studies that reported results from symptom rating scales in the first 24 h following injury were extracted and included in table 1. The time interval following injury varied considerably, from minutes to the following day. Research groups used many variations of concussion rating scales, which varied considerably in the number of symptoms being assessed.
The Post-Concussion Scale15–17 has evolved over the years, and slightly different versions have been used in published studies.18–24 The internal consistency of the scale ranges from 0.88 to 0.94 in high school and college students, and from 0.92 to 0.93 in concussed athletes.25 Normative data for the scale are available for 1391 young men and 355 young women in high school or university.25
SCAT2 is an outcome measure that includes symptom ratings, balance testing and cognitive screening.7 Two scores are derived from the SCAT2 symptom scale: total number of symptoms and symptom severity score. The symptom severity score is intuitive in that it is the sum of the Likert ratings for the entire scale. This score, however, is not included in the SCAT2 total score. The total number of symptoms score is included in the SCAT2 total score. This score is reverse scored in that the total number of symptoms endorsed as mild or greater are subtracted from 22 (thus, if a person endorses 8 symptoms, his score would be 14). A number of concerns have been expressed about the design and scoring of SCAT2, and recommendations have been offered for improving it for clinical use.26 Normative reference values are available for high school students. Data are not available on large samples of concussed athletes.
The Concussion Symptom Inventory (CSI27) is a 12-item scale that was developed using samples of more than 16 000 uninjured athletes and more than 600 concussed athletes. It can be derived from the Post-Concussion Scale and SCAT2. A comparison of the three concussion rating scales is provided in table 2.
One of the most salient findings from this review is that these rating scales are sensitive to the acute effects of concussion in high school, university and professional athletes. Acutely concussed athletes, on average, obtain high scores on symptom rating scales, which is consistent with a meta-analytic review of the literature. Broglio and Puetz28 extracted data from 14 studies involving 1796 injured athletes. Across the 14 studies, the mean overall effect size an average of 1.2 days after injury was large (ie, –3.31). In 10 studies that included both symptom ratings and cognitive assessment (N=913 injured athletes), the average effect sizes for symptom ratings (–1.21) and cognition (–0.95) were more similar. Symptom checklist scores and the corresponding sensitivity of the scales typically diminish with the passage of time due to natural recovery.
A total of 31 studies using cognitive testing on the day of injury qualified for review (see table 3). Data were available from several competitive levels across a broad array of sports. Most of these studies used brief cognitive screening tests intended for rapid assessment in the sporting environment (eg, sideline and rink side). Fewer studies utilised conventional (‘paper and pencil’) neuropsychological testing within 24 h of injury, and 11 studies involved the use of computerised neuropsychological testing. The most widely studied instrument has been the Standardised Assessment of Concussion (SAC; 12 studies). ANAM (n=5), CogSport/Axon (n=3) and ImPACT (3 studies) were the most commonly studied computerised tests.
The array of studies reviewed illustrates the ability to measure or quantify the nature and magnitude of cognitive impairment within the first 24 h of concussion. These studies document that neuropsychological tests can detect changes across multiple domains of cognitive functioning that are susceptible to the acute effects of concussion, such as cognitive processing speed, working memory, attention and concentration, new learning and memory and executive functioning. Measures of orientation appear less sensitive, but might be a marker of a more severe gradient of injury when impaired.
The study by Broglio et al29 was the most extensive comparison of conventional and computerised neurocognitive testing approaches within 24 h of injury. This study revealed different levels of detecting cognitive abnormalities across ImPACT, Headminder CRI and a conventional test battery. A high percentage of symptomatic athletes exhibited a significant decline on both computerised and conventional neuropsychological testing on the day of injury. Approximately 15–30% of athletes were also impaired on testing after reporting a full symptom recovery. It should be noted, however, that 10–25% of athletes who were still symptomatic showed no significant decline on cognitive testing.
Overall, the studies reviewed illustrated a significant decline in cognitive functioning compared with an athlete's individual preinjury baseline performance, relative to the performance of non-injured control athletes or both. This ability to detect and quantify acute cognitive impairment is evident on both brief screening measures (ie, the SAC) and conventional and computerised neuropsychological test batteries. Relatively few studies report sensitivity/specificity data on the ability of particular cognitive tests to differentiate concussed versus non-concussed athletes at the individual case level.
Postural stability/balance tests
Numerous studies illustrate that balance is typically affected in concussed athletes in the early postinjury period2 ,29–34 (see table 4). The majority of studies examine group data in concussed athletes compared with their own baseline and/or compared with uninjured controls.2 ,29–34 Initial studies involved the use of sophisticated computerised test platforms, which are not practical for sideline assessments. Guskiewicz and colleagues later reported that simple clinical tests (eg, Balance Error Scoring System, BESS) could also detect balance deficits following injury.34 ,35 A study by Broglio et al29 assessed the ability of balance testing to detect a significant change (defined as 1 SD below baseline) in the individual athlete following a concussion. Most, but not all, concussed athletes had significant balance deficits following injury.
Taken together, these studies show that balance is an important component of the sideline assessment.36 ,37 Currently, SCAT2 contains a modified BESS (M-BESS) which utilises the three BESS stances on a hard surface only. Further research is required to examine the sensitivity and reliability of the M-BESS for the detection of balance deficits in concussed athletes assessed on the sidelines.
Electrophysiological measures have been used to examine athletes and non-athletes following concussion.38–40 Only recently have researchers examined the role of these measures during the acute phase of injury (see table 5). McCrea et al41 conducted a prospective non-randomised study of 28 high school and college athletes who sustained a concussion while playing football. Analyses of the qEEG data revealed significant within-group differences from baseline to the first 24 h postinjury and at 8 days postinjury. Differences between the concussed and control groups were found within the first day postinjury and at 8 days postinjury.
Barr et al42 reported significant differences using qEEG between the injured and control groups within the first 24 h and 8 days postinjury; no differences were found between groups at 45 days. Drawing from the same cohort of athletes and using the same measures as the previous two studies, Prichep et al43 found statistically significant differences between athletes with mild and moderate concussion within 24 h, 8 days and 45 days postinjury. Sensitivity to moderate concussion was reported as 55% (95% CI 28% to 79%) and specificity was noted to be 94% (95% CI 84% to 98%) within 24 h of injury.
Other assessment measures
There are references in the literature and mainstream media regarding new paradigms, tests and devices in development for the intended use of assessing sport-related concussion. Only three studies in the literature included the day of injury data (see table 6). No conclusion can be drawn at this point regarding the eventual usefulness of these methods or others yet in the development pipeline.
Results from this critical review of the scientific literature inform our understanding of the acute clinical presentation of sport-related concussion and an array of tests in use for concussion assessment. Concussion clearly produces a constellation of self-reported symptoms and impairments in cognitive functioning, balance and other functional capacities during the acute phase (eg, initial 24 h). This review illustrates that clinical tests are capable of detecting and quantifying those effects on the day of injury.
It is important to appreciate that tests do not diagnose whether a concussion has occurred. Rather, tests provide data on the physiological, cognitive, psychological and behavioral changes associated with the injury that can aid the clinician in the overall diagnostic formulation, gauging the gradient of injury severity, assessing clinical recovery and determining fitness to resume participation in sport.
Although concerns are routinely expressed about athletes under-reporting concussion or the resulting symptoms,44 this review suggests that symptom assessment remains a critical component of concussion assessment. The studies reviewed indicate large effect sizes in self-reported symptoms on the day of injury, typically larger than the deficits seen in performance-based methods. In sum, the literature clearly supports the continued use of symptom scales in the assessment of concussed athletes, ideally in combination with other functional tests.
This review illustrates that the effects of concussion are not confined to one domain (eg, cognitive functioning). That is, athletes, to varying degrees, experience a complex combination of symptoms and exhibit deficits across multiple domains of functioning. As a result, reliance on a single test or multiple measures in a single assessment domain will very likely be less accurate than a multimodal assessment. A multidimensional approach that integrates assessment of self-reported symptoms and other functional domains (eg, cognitive function and balance) known to be affected by concussion is recommended. The specific tests selected to form the multidimensional assessment approach not only need to be reliable and valid, but also may vary depending on the situational constraints of the sporting environment, athlete population and the experience of the assessor. SCAT2 integrates a multimodal assessment into a single outcome measure for use in the competitive sporting environment. Individual components of SCAT-2 have been the focus of recent studies, but further research is required to illustrate the reliability, sensitivity and clinical utility of the entire SCAT2.
The evidence base on the utility of neuropsychological testing has been greatly advanced over the past two decades. Although many critical questions have been answered, further research is required to specifically address a number of important questions related to the usefulness of cognitive testing. Priorities for future study are listed below.
Comparing the added value of computerised versus brief cognitive screening tests on the day of injury and at subsequent assessment points.
Determining how the day of injury cognitive test performance might predict the course and duration of recovery (including markers indicating the risk of prolonged recovery).
Comparing the reliability, validity, sensitivity and specificity of baseline versus no baseline models of cognitive assessment in measuring the acute effects of concussion.
A small number of recent studies provide preliminary data regarding the use of electrophysiological and other measures in the acute evaluation of concussion43 ,45 and suggest that physiological recovery may persist significantly beyond recovery as assessed by clinical measures (eg, symptoms, balance and cognition).41 Although these methods show promise, further research is required to support their use in the sporting environment.
Based on the complex and heterogeneous clinical presentation of concussion, a multimodal approach is recommended to maximise the sensitivity of clinical evaluation on the day of injury. This approach includes formal symptom assessment and standardised testing of cognitive ability and balance. Several tests and measures are appropriate for use in the assessment of acute concussion on the day of injury in the competitive sporting environment. These tests provide important data on the symptoms and functional impairments that clinicians can incorporate into their diagnostic formulation, but they should not solely be used to diagnose concussion. Future research should focus on the incremental value of these assessment instruments for predicting an athlete's recovery course, influencing injury management strategies and reducing risks associated with sport-related concussion.
What are the new findings?
Major progress has been made towards the development and validation of standardised methods for the assessment of sport-related concussion over the past 20 years.
Several tests and measures are appropriate for use in the assessment of acute concussion on the day of injury in the competitive sporting environment.
A multimodal approach that includes formal symptom assessment and standardised testing of cognitive ability and balance is recommended to maximise the sensitivity of clinical evaluation on the day of injury.
How might it impact on clinical practice in the near future?
Concussion tests provide important data on the symptoms and functional impairments that clinicians can incorporate into their diagnostic formulation, but they should not solely be used to diagnose concussion.
Inclusion of standardised measures will provide clinicians with a more systematic and accurate approach to concussion assessment.
Accurate assessment on the day of injury can drive informed clinical management and return to play decision making that improves athlete safety and outcome after sport-related concussion.
The authors acknowledge and appreciate the assistance of Adam Pfaller, Research Assistant in the Brain Injury Research Program at the Medical College of Wisconsin, with the literature search and data extraction for this systematic review.
Contributors MM, GI, RE, MM and MR made substantial contributions to: conception and design, or analysis and interpretation of data; drafting the article or revising it critically for important intellectual content; and final approval of the version to be published. All authors included on a paper fulfil the criteria of authorship. There is no one else who fulfils the criteria but has not been included as an author.
Competing interests See the supplementary online data for competing interests (http://dx.doi.org/10.1136/bjsports-2013-092145).
Provenance and peer review Commissioned; internally peer reviewed.
▸ References to this paper are available online at http://bjsm.bmjgroup.com