Arch Phys Med Rehabil. 2014 (Mar); 95 (3 Suppl): S278–285 ~ FULL TEXT
J. David Cassidy, PhD, DrMedSc, Eleanor Boyle, PhD, Linda J. Carroll, PhD
Faculty of Health,
Institute of Sports Science and Clinical Biomechanics,
University of Southern Denmark,
Division of Health Care and Outcomes Research,
Toronto Western Research Institute,
University Health Network,
University of Toronto,
Toronto, Ontario, Canada
OBJECTIVE: To determine the incidence, course, and prognosis of adult mild traumatic brain injury (MTBI) caused by motor vehicle collisions.
DESIGN: Prospective, population-based, inception cohort study.
SETTING: The province of Saskatchewan, Canada, with a population of about 1,000,000 inhabitants.
PARTICIPANTS: All adults (N=1716) incurring an MTBI in a motor vehicle collision between November 1997 and December 1999 in Saskatchewan.
INTERVENTIONS: Not applicable.
MAIN OUTCOME MEASURES: Age- and sex-stratified incidence rates, time to self-reported recovery, and prognostic factors over a 1-year follow-up.
RESULTS: Of 7170 adults injured in a motor vehicle collision over the 2-year inception period, 1716 (24%) met our cohort definition of MTBI. There were more women affected (53%), and MTBI was most common in the 18- to 23-year-old group. Most were not hospitalized (73%), but 28% reported loss of consciousness and 23% reported posttraumatic amnesia. The annual incidence of MTBI per 100,000 adults was 106.1 (95% confidence interval [CI], 98.9-113.6) in the first year and 118.3 (95% CI, 110.8-126.3) in the second year of the study. The 1-year follow-up rate was 84%. The median time to recovery was 100 days (95% CI, 97-103), and about 23% reported not having recovered by 1 year. Factors associated with delayed recovery included being older than 50 years, having less than a high school education, having poor expectations for recovery, having depressive symptoms, having arm numbness, having hearing problems, having headaches, having low back pain, and having thoracic back pain. Loss of consciousness and posttraumatic amnesia were not associated with recovery.
CONCLUSIONS: MTBI affects almost a quarter of persons reporting an injury after a traffic collision. The median time to recovery is 100 days, but 23% have still not recovered by 1 year. A mix of biopsychosocial factors is associated with recovery, including a strong effect of poor expectations for recovery.
KEYWORDS: Brain concussions; Cohort studies; Epidemiology; Incidence; Prognosis; Recovery of function; Rehabilitation
From the FULL TEXT Article:
After falls, traffic collisions are the most common cause of mild traumatic brain injury (MTBI).  It is estimated that 70% to 90% of all treated brain injuries are mild, and the incidence of hospital-treated MTBI is between 100 and 400 per 100,000 population in developed nations.  However, many persons with MTBI are not treated at hospitals, and therefore, its true incidence is likely above 600 per 100,000 population.  Variability in case definition and diagnosis likely accounts for much of the variation across studies, but cultural factors might also be at play. [3, 4] Most studies capture cases presenting to hospitals, and there are very few population-based estimates of the problem, and even fewer focused on traffic-related MTBI. 
The situation is similar with respect to prognosis because there are few studies on prognostic factors for MTBI after traffic collisions. The World Health Organization Collaborating Centre Task Force on Mild Traumatic Brain Injury systematically reviewed this literature and found only 2 scientifically acceptable studies of traffic-related MTBI.  Friedland and Dawson  compared 64 persons with MTBI admitted to a tertiary care center to 64 persons admitted with other injuries and found that functional recovery and return to work were similar between the groups. Overall, posttraumatic stress was associated with slower recovery in both groups. Cassidy et al  reported 657 persons with MTBI with some loss of consciousness (LOC) who made an insurance claim or were treated by a health professional after a traffic collision in Saskatchewan, Canada, in the period 1994 to 1995. During this study, the insurance system changed from “tort” to “no-fault,” thereby increasing medical benefits, but discontinuing payments for pain and suffering, and therefore, limiting court actions. Their results showed a decrease in the 6-month incidence of MTBI claims from 36 to 27 per 100,000 adults and an improvement in the median time to claim closure from 408 to 233 days. They also showed that claim closure occurred faster when the claimants' health improved. Prognostic factors associated with slower claim closure included being off work because of the collision, being not at fault for the collision, reporting nausea after the collision, reporting memory problems after the collision, and a greater percent of bodily pain.
Given the lack of good-quality published studies on MTBI after traffic collisions, there is an obvious gap in knowledge in this respect. Unlike most falls, traffic injuries are complicated by insurance issues as noted above, and the World Health Organization Collaborating Centre Task Force on Mild Traumatic Brain Injury called for more studies in this setting. The purpose of this study was to document the incidence and course of MTBI after traffic collisions in the adult population. We also explore potential prognostic factors that have an impact on recovery.
Participants and setting
A population-based inception cohort was formed of all traffic injuries that occurred between December 1, 1997, and November 30, 1999, in the province of Saskatchewan in Canada. Entry into the cohort occurred if the person was treated by a registered health professional for a traffic injury or if the person made a bodily injury claim to Saskatchewan Government Insurance (SGI), the only insurer for traffic injuries in Saskatchewan. We included participants 18 years and older who were involved in a motor vehicle collision and who made a claim or were treated for injuries within 42 days of their collision. We excluded individuals who died as a result of their injuries or could not answer the baseline questionnaire because of a language barrier or because of serious unrelated illness. We also excluded those injured at work and covered by Workers' Compensation.
We identified the participants with probable MTBI through a 3-step process based on the self-report baseline questionnaire administered by the SGI (Figure 1). We first identified participants who answered “yes” to the question, “Did you hit your head in the collision?” Second, we identified participants who had answered “yes” or “don't know” to one of these symptom-based questions: “Did you lose consciousness immediately after the accident?” or “Immediately after the accident, did you experience amnesia or loss of memory?” or “Immediately after the accident, did you experience disorientation or confusion?” In addition, the study participants had to have answered “yes” to at least 1 of the following questions for inclusion in the cohort: “Did the accident cause dizziness or unsteadiness?” or “Did the accident cause memory problems or forgetfulness?” or “Did the accident cause concentration or attention problems?” Finally, we excluded study participants who reported more than 30 minutes of LOC after the collision.
Flow diagram of cohort assembly.
The baseline research questionnaire was collected by SGI and was available on all injured persons. It contained questions about demographic characteristics (ie, age, sex, height, weight, marital status, highest education level achieved, household income, number of dependents), number of different preexisting comorbid conditions, collision characteristics (ie, position in the car, direction of impact, seatbelt use, headrests), hospitalization, past head injuries, collision-related symptoms (ie, numbness, dizziness or unsteadiness, memory problems or forgetfulness, concentration or attention problems, irritability, vision problems, hearing problems, sleep problems, unusual fatigue or tiredness, anxiety or worry, painful neck movement, painful jaw, LOC, posttraumatic amnesia [PTA], disorientation or confusion), fractures, depressive symptoms, early treating practitioner type, quality of health before and after the injury, pain location and intensity, expectations for recovery, work status, job satisfaction, and activities of daily living. Pain intensity was measured using the 11-point numerical rating scale, where a score of “0” meant they had no pain at all and a score of “10” meant pain as bad as could be. The numerical rating scale has excellent psychometric properties.  The health transition question and the overall general health question of the 36-item short-form health survey were included in the baseline questionnaire along with a question about their general health before the collision.  The Centre for Epidemiological Studies–Depression Scale was included at baseline. It was designed to measure current levels of depressive symptoms with a score range of 0 to 60, where a higher score indicates greater depressive symptoms. The Centre for Epidemiological Studies–Depression Scale has been found to have good test-retest reliability and validity.  Participants were also asked whether they thought their injury would “get better soon,” “get better slowly,” “never get better,” or “don't know.” The presence of comorbid conditions was measured using an inventory that was previously validated. [11, 12]
At 6 weeks and 3, 6, 9, and 12 months postinjury, the study participants were followed by computer-aided telephone interviews conducted at our research center. The interviewers were blinded to the participants' previous responses. The participants were asked about how well they thought they were recovering from their injuries, with the response categories of “all better or cured,” “feeling quite a bit of improvement,” “feeling some improvement,” “feeling no improvement,” “getting a little worse,” or “getting much worse.” Using this question, self-reported recovery was defined by grouping the responses of “all better (cured)” and “feeling quite a bit of improvement” together and those who responded otherwise were classified as having not recovered. This question has been found to have excellent test-retest reliability and criterion validity. [13, 14]
The Research Ethics Boards of the University of Saskatchewan and the University of Alberta approved the original study, and the University Health Network, University of Toronto, approved the analysis for this article (REB Approval 13-6095-AE).
Means and frequencies were calculated to describe the baseline characteristics of members of the cohort. We calculated the crude age-, sex-, and age-sex–specific incidence rates of MTBI for each of the 2 years of inception into the study. The denominator was calculated using population data from the Saskatchewan government.  Ninety-five percent confidence intervals using the Fisher exact test were calculated for each incidence rate. Time to first reported recovery was modeled using a Kaplan-Meier curve. Participants who were lost to follow up were censored halfway between their last completed questionnaire and the next scheduled interview, and their previous interview was used to determine the final outcome. We used a backward-stepping Cox's proportional hazard model with a P value to remove set at .10 to determine which baseline characteristics were associated with time to recovery during the 1-year follow-up period.
During the 2-year inception period, 8,634 persons were involved in a motor vehicle collision. We excluded 1,464 participants because they were not a driver or a passenger in a motor vehicle during the time of the collision (ie, pedestrians and bicyclists) or filed an injury claim more than 42 days after the collision. Of the remaining 7,170 participants, 31% stated that they had hit their head during the collision. There were 1,768 participants who experienced at least 1 of the following symptoms associated with a TBI: disorientation or confusion (60%), LOC (28%), PTA (23%), dizziness or unsteadiness (72%), memory problems or forgetfulness (36%), or concentration or attention problems (39%). We excluded 52 participants who stated that they lost consciousness for more than 30 minutes. The final cohort had a sample size of 1,716 (see fig 1).
The cohort had a mean age of 37.7 years, and 53% were women. Forty-nine percent reported that they were married or living common-law, and 44% had more than a high school education. Most of them (71%) were driving the vehicle at the time of the collision. Although 1,365 (80%) members of the cohort went to the hospital after their collision, only 463 (34%) of them were admitted overnight. Many experienced pain when moving their neck (75%), sleeping problems (75%), unusual fatigue or tiredness (60%), and anxiety or worry (57%) after the collision. Neck pain (90%), headache (84%), low back pain (63%) and mid back pain (58%) were commonly reported, with neck pain and headache being the most intense. The average neck pain intensity was 6.1 of 10, and the average headache intensity was 5.7 of 10 on the numerical rating scale. Twenty-one percent suffered a fracture because of the collision. Most (76%) reported excellent or good health before the collision, but this was reduced to 31% after the collision (Table 1).
The age-sex–specific incidence rate for the first year of recruitment was 106.1 per 100,000 population, and this increased to 118.3 per 100,000 population for the second year. Overall, women had a marginally higher incidence rate of MTBI than did men. The lowest age-specific incidence rate was in the age group of 50 to 59 years and the highest was in the age group of 18 to 23 years (Table 2).
Characteristics of participants
Age- and sex-specific incidence rates
(95% confidence intervals) per 100,000 population
in first and second years of inception into the cohort
The overall, 1-year follow-up rate for the cohort was 84%. The median time to recovery was 100 days (95% confidence interval, 97–103), but close to 23% of the cohort still considered themselves not recovered by 1 year (Figure 2). Factors associated with delayed recovery included being 50 years of age or older, having lower educational attainment, having poor expectations for recovery, having a higher score on the depressive symptoms scale (Centre for Epidemiological Studies–Depression Scale), having arm numbness as a result of the collision, having hearing problems after the collision, being unsure about disorientation or confusion after the collision, and having low back, mid back, or headache pain after the collision (Table 3). Our final model is based on 1476 persons. LOC, PTA, comorbid health issues, and sex were not associated with time to recovery.
of time to recovery.
Final step of Cox's proportional
hazard model (n=1476)
The average annual incidence of traffic collision–related MTBI in the Canadian province of Saskatchewan during 1998 to 1999 was 109 per 100,000 adults, and this agrees with a population-based estimate of 105 per 100,000 population from Sweden in 2001.2 Overall, 80% of our cohort presented to the emergency department after their injury, which would give an incidence of 87 per 100,000 adults treated in the emergency department, which is within the estimates published in a systematic review by Cassidy et al.  We found no appreciable difference in the incidence of men and women affected, but we did find that the incidence rates varied over age, with higher rates in younger individuals. This age difference confirms what was seen in an earlier cohort study from Saskatchewan.  Furthermore, MTBI accounted for 24% of all traffic injuries, and many of these people also suffered from back pain, which had an impact on their recovery. This reflects the reality of population-based injury studies in which combinations of injuries are common.  On average, those with MTBI took 100 days to recover and 23% had not recovered by 1 year. Delayed recovery was highly associated with poor expectations for recovery measured at baseline. Other factors having an impact on recovery included lower education, older age, depressive symptoms, arm numbness, hearing problems, and pain in the head, mid back, and low back. Sex, LOC, and PTA did not have an impact on recovery.
Our study suggests that poor expectations for recovery have an important impact on prognosis. Hou et al  also found that negative head injury perceptions measured by the Brief Illness Perceptions Questionnaire were associated with more postconcussion symptoms at follow-up in patients with MTBI seen in the emergency department of a United Kingdom general hospital. Patients who viewed their MTBI as having a number of symptoms out of their control, lasting a long time, and having a serious impact on their life were likely to develop postconcussion syndrome 3 months later. Also, poor expectations for recovery have been identified as strong prognostic factors in whiplash  and occupational soft-tissue injuries.  These relations need to be tested more rigorously in the MTBI population, and negative beliefs might be treatable early in the course of MTBI through reassurance and education. We also found that baseline depressive symptoms were associated with poor recovery. This too might be modifiable with early intervention. The World Health Organization Collaborating Centre Task Force on Mild Traumatic Brain Injury identified depression as an important determinant of MTBI recovery and recommended intervention studies directed at early postinjury depression.  However, to this date, there have been no such studies reported. It may be that poor recovery perceptions and depressive symptoms are related, and they might respond to a cognitive behavior approach in those at risk of poor recovery. Future studies on this are warranted.
Many persons in our cohort reported bodily pain, particularly in the head, neck, and back. It is likely that MTBI occurs with other soft-tissue injuries during motor vehicle collisions. It therefore stands to reason that pain might also determine recovery in these persons and our results reflect that. Furthermore, there is some uncertainty about the overlap between MTBI and whiplash-associated disorders because pain and depression are also associated with perceived cognitive deficits.  MTBI and whiplash-associated disorders might share a common causal mechanism in exposure to acceleration-deceleration forces. Until there are better diagnostic tools developed, this uncertainty will continue. The clinical reality is that they probably occur together in some cases and can be difficult to differentiate. 
Our study has some strengths and limitations. The cohort included all treated traffic injuries in a defined population of about 1 million residents, limiting any selection bias. However, it is possible that we missed some patients with very mild injury who did not seek health care and did not file a bodily injury claim. Our definition of MTBI included all persons who hit their head during the collision and complained of symptoms compatible with MTBI.  Our follow-up rate was more than 80%, which limits any selection bias due to attrition.  We used reliable and valid measures of self-reported recovery, depressive symptoms, pain intensity, general health, and comorbid illnesses. Other simple questions on symptoms were designed for this study, but they have not been subjected to psychometric testing, which could be a weakness resulting in information bias. Our study is large enough to give good precision around estimates of incidence and prognosis. However, we conceived our study as an exploratory prognostic study, designed to test independent associations between multiple candidate prognostic variables and recovery, and these associations need to be confirmed in other studies.  In addition, many of the persons in our study had other soft-tissue injuries that likely affected our prognostic findings. However, this is likely a strength of our study because it reflects the clinical reality of traffic-related injuries. Finally, although our baseline research data were collected by SGI, we do not think that this affected answers to these questions because at the time, Saskatchewan operated under a pure no-fault insurance program with no financial or other incentives to embellish baseline symptoms. Our follow-up questions were collected by our research center, and SGI did not have access to this information.
There are few population-based studies of the incidence, course, and prognosis of MTBI after traffic collisions. This study provides this information. MTBI was found to affect 24% of all injured persons in the population, and recovery took 100 days on average. At 12 months, 23% of the participants viewed themselves as not recovered. Potential prognostic variables included a mix of baseline biopsychosocial factors, including older age, lower education, poor expectations for recovery, depressive symptoms, arm numbness, hearing problems, possible confusion, and pain in the head, mid back, and low back. MTBI is a fairly common consequence of traffic collisions that can occur with other soft-tissue injuries, and it can result in a prolonged recovery.
Cassidy, J.D., Carroll, L.J., Peloso, P.M. et al.
Incidence, risk factors and prevention of mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury.
J Rehabil Med. 2004; 36: 28–60
Styrke, J., Stålnacke, B.M., Sojka, P., and Björnstig, U.
Traumatic brain injuries in a well-defined population: epidemiological aspects and severity.
J Neurotrauma. 2007; 24: 1425–1436
Carroll, L.J., Cassidy, J.D., Holm, L. et al.
Methodological issues and research recommendations for mild traumatic brain injury: the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury.
J Rehabil Med. 2004; 36: 113–125
Ferrari, R., Constantoyannis, C., and Papadakis, N.
Cross-cultural study of symptom expectation following minor head injury in Canada and Greece.
Clin Neurol Neurosurg. 2001; 103: 254–259
Carroll, L.J., Cassidy, J.D., Peloso, P.M. et al.
Prognosis for mild traumatic brain injury: results of the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury.
J Rehabil Med. 2004; 36: 84–105
Friedland, J.F. and Dawson, D.R.
Function after motor vehicle accidents: a prospective study of mild head injury and posttraumatic stress.
J Nerv Ment Dis. 2001; 189: 426–434
Cassidy, J.D., Carroll, L., Côté, P., Holm, L., and Nygren, A.
Mild traumatic brain injury after traffic collisions: a population-based inception cohort study.
J Rehabil Med. 2004; 36: 15–21
Jensen, M.P., Karoly, P., and Braver, S.
The measurement of clinical pain intensity: a comparison of six methods.
Pain. 1986; 27: 117–126
Ware, J.E. Jr. and Sherbourne, C.D.
The MOS 36-item short-form health survey (SF-36), I: conceptual framework and item selection.
Med Care. 1992; 30: 473–483
Devins, G.M., Orme, C.M., Costello, C.G. et al.
Measuring depressive symptoms in illness populations: psychometric properties of the Center for Epidemiologic Studies Depression (CES-D) scale.
Psychol Health. 1988; 2: 139–156
Jaroszynski G, Cassidy JD, Carroll L, Côté P, Yong-Hing K.
Development and validation of a comorbidity scale.
Presented at the Canadian Orthopaedic Research Association Convention,
Quebec City, PQ, June 1996.
Assessing the performance of a self-report comorbidity scale [dissertation].
Univ of Alberta, Edmonton; 2006
Ngo, T., Stupar, M., Côté, P., Boyle, E., and Shearer, H.
A study of the test-retest reliability of the self-perceived general recovery and self-perceived change in neck pain questions in patients with recent whiplash-associated disorders.
Eur Spine J. 2010; 19: 957–962
Carroll, L.J., Jones, D.C., Ozegovic, D., and Cassidy, J.D.
How well are you recovering? The association between a simple question about recovery and patient reports of pain intensity and pain disability in whiplash-associated disorders.
Disabil Rehabil. 2012; 34: 45–52
Health insurance registration: covered population.
Regina (Canada): Government of Saskatchewan; 1997-1999.
Ferrari, R., Russell, A.S., Carroll, L.J., and Cassidy, J.D.
A re-examination of the whiplash associated disorders (WAD) as a systemic illness.
Ann Rheum Dis. 2005; 64: 1337–1342
Hou, R., Moss-Morris, R., Peveler, R., Mogg, K., Bradley, B.P., and Belli, A.
When a minor head injury results in enduring symptoms: a prospective investigation of risk factors for postconcussional syndrome after mild traumatic brain injury.
J Neurol Neurosurg Psychiatry. 2012; 83: 217–223
Holm, L.W., Carroll, L.J., Cassidy, J.D., Skillgate, E., and Ahlbom, A.
Expectations for Recovery Important in the Prognosis of Whiplash Injuries
PLoS Med. 2008 (May 13); 5 (5): e105
Hogg-Johnson, S. and Cole, D.C.
Early prognostic factors for duration on temporary total benefits in the first year among workers with compensated occupational soft tissue injuries.
Occup Environ Med. 2003; 60: 244–253
Sullivan, M.J., Hall, E., Bartolacci, R., Sullivan, M.E., and Adams, H.
Perceived cognitive deficits, emotional distress and disability following whiplash injury.
Pain Res Manag. 2002; 7: 120–126
Kischka, U., Ettlin, T., Heim, S., and Schmid, G.
Cerebral symptoms following whiplash injury.
Eur Neurol. 1991; 31: 136–140
Kristman, V., Manno, M., and Côté, P.
Loss to follow-up in cohort studies: how much is too much?.
Eur J Epidemiol. 2004; 19: 751–760
Côté, P., Cassidy, J.D., Carroll, L., Frank, J.W., and Bombardier, C.
A systematic review of the prognosis of acute whiplash and a new conceptual framework to synthesize the literature.
Spine. 2001; 26: E445–E458
Return to MILD TRAUMATIC BRAIN INJURY