ABSTRACT
Objectives
To investigate functional outcome after TBI and identify variables that predict outcome in a multiordinal regression model.
Background
The results of global outcome studies after Traumatic Brain Injury(TBI) differ widely due to differences in outcome measure, attrition to follow-up and selection bias. Outcome information would inform patients/families, guide service development and target high-risk individuals
Subjects/Setting
prospective cohort of 1322 admissions with TBI, assessed by face to face interviews at 1 yr.
Measures
Extended Glasgow Outcome Scale (GOSE) by structured questionnaire.
Results
At 1 year, outcome was determined in 1207(91.3%). Mean age was 46.9(SD17.3); Almost half(49.2%) had mild injury. At one year, 42.9% achieved Good Recovery but GOSE declined in 11.4% of the cohort compared to 10 weeks including 60(4.9%) deaths. In an ordinal logistic regression, increasing TBI severity, etiology (assault), more prominent CT abnormality, past psychiatric history and alcohol intoxication were independent predictors of worse GOSE. A pseudo-R2 of 0.38 suggested that many unmeasured factors also contribute to TBI outcome. Future work needs to identify other variables that may influence outcome.
Conclusions
In a large TBI cohort, there is still considerable functional disability at 1 year. It may be possible to target high-risk groups for rehabilitation
Introduction
Traumatic Brain Injury (TBI) remains a significant source of mortality and disability across the globe, especially in the young (Citation1–3). It has long-lasting consequences for victims and their families(Citation4), for society and for healthcare resources. An accurate prediction of outcome would allow better information for families and individuals thus allowing forecasting of care needs. A good understanding of prognosis after injury would inform the development of services in future as the incidence of TBI continues to increase (Citation1,Citation5).
Yet our knowledge of TBI outcome remains somewhat unclear; few individuals report complete symptom resolution(Citation6) and published studies differ considerably in their findings as a result of design differences. One such key difference is that outcome itself can be measured in many ways with measures across different parameters of health(Citation7). Some may exhibit selection bias such as exclusion of elderly or only include moderate and severe TBI, whereas most TBI is mild. The loss of subjects after TBI to follow-up is very high, as much as 70% within a few months(Citation8). Hence, there is a need for large, prospective, high-retention follow-up studies as highlighted by many (Citation3,Citation5,Citation9). This should be in a non-biased group that is truly characteristic of the TBI population(Citation1).
In our district, an opportunity arose to create a rehabilitation pathway for follow-up of admitted TBI. This allowed a Rehabilitation Medicine team to offer advice and support for families and to organize a follow-up clinic for all admitted TBI cases. It has been shown that early rehabilitation input improves the management of TBI(Citation10) and that coordination can improve the outcome (Citation11,Citation12). The new service identified all admissions, reviewed them after injury, offered support to family members and arranged any further referrals.
This pathway facilitated the follow-up of a prospective cohort with TBI that included all severities, age and injury types; the Sheffield Brain Injury after Trauma (SHEFBIT) cohort. This cohort reflects a valid picture of TBI as treated by health professionals and can help to guide clinicians and patients. It represents a “real-world” pathway and is germane to all professionals with an interest in TBI.
The primary aim of the study was to measure and document the 1 year functional outcome after TBI. The secondary aims were to try to identify if there is any key injury or demographic variable that can predict outcome in a regression model. Some demographic data from this cohort has been previously reported with no data analysis(Citation13). But this paper presents completely new results of 1 year follow-up data and the regression analyses to investigate the best predictors of functional outcome after TBI.
Methods
Clinical pathway/population
The organization of the TBI clinical pathway has been described elsewhere(Citation12).The newly created Brain Injury Rehabilitation Team monitor the care of all admissions with TBI. The team assist with transfers between trauma and rehabilitation services and arranges follow-up.
Individuals admitted with TBI to a University hospital from 2011 to 15 were entered into the SHEFBIT cohort (Sheffield Brain Injury after Trauma). All participants had GCS recorded in the ED and used for TBI severity. All individuals received a head CT scan and were admitted for a minimum of one day. Those with prior TBI, age <17, resident out of region or dementia, were excluded. The diagnosis of TBI was confirmed by the American Congress of Rehabilitation Medicine criteria and position statement(Citation14).
All individuals were followed up at 8–10 weeks in a Brain Injury clinic and again after 1 year. Those failing to attend were called for re-appointment. All patients were interviewed by the same PRM physician. Demographic and injury factors at time of injury such as gender, employment, GCS and CT scans were recorded. Psychiatric history was defined as any treatment for a diagnosed psychiatric condition at time of injury. Alcohol intoxication at the time of injury was made by taking history or from ED admission file.
Mechanism of injury was recorded as falls, assault, road traffic collisions (RTC), sporting activity and “other” mechanisms (workplace injuries or falls from a height)(Citation15). CT scans were classed under the “overall appearance” method which labels injury as normal, mild focal, medium focal (two adjacent lobes) and diffuse injury(Citation16). Significant medical comorbidity was measured by the Cumulative Illness Rating Scale (CIRS) and a cutoff above 10(Citation17). The National Statistics Socio-Economic Classification (NS-SEC)(Citation18) was used for socioeconomic class (SEC). Pre-injury employment was defined as those working (including full-time students), unemployed or retired.
The study was approved by both the University Hospital Trust and the University of Sheffield Ethics Committees (STH16208). All individuals gave consent for the study at their initial follow-up.
Glasgow outcome scale
The study outcome measure was the Glasgow Outcome Scale-Extended (GOSE) in a structured clinical interview(Citation19). The extended version improves the discrimination between levels as well as excellent correlation with other outcomes and cognitive scores(Citation20). Assessment by a single investigator (RS) with the structured clinical interview was used to minimize misclassification. In some cases, relatives helped with completion. As the level of vegetative state is rare, this level was merged with severe lower outcome.
Statistics
The results for demographic data are presented as frequencies with percentages. Numerical data is presented as mean and standard deviation when approximating to normal distribution or median and range otherwise. Comparison between follow-up and lost individuals was made with t-test or ᵡ2 test/Fisher Exact test for continuous or categorical variables as appropriate.
To assess the independent predictors of the main outcome, an ordinal logistic regression model was estimated with GOSE as the dependent outcome. All demographic and injury varables in the study were entered as continuous or categorical predictors. Post-hoc tests were applied to probe the individual variables that were significantly different but had more than 2 categories (NS-SEC, etiology, CT scan, pre-injury employment). Statistical analysis was performed using SPSS version 23.
Results
Population and injury features
Over the study period, 1934 admissions with TBI were recruited. After cases of prior TBI, dementia or non-local residence were omitted (209 individuals) as well as 319 where the diagnosis of TBI by Emergency Department could not be confirmed after history, 1406 individuals remained; follow-up appointments were arranged for 8–10 weeks and 1322 attended the first appointment. Appointments were repeated after 1 year. After one year there were 60 (4.9%) deaths and 115 (8.7%) cases were lost to follow-up. The lost group were older by 3.7 years but the only significant difference between groups was that CT scans were more likely to have mild or moderate abnormality but not severe in the lost group (). This final study population corresponds to a total of 1207 cases (including deaths) with a GOSE at one year. This represents a one year follow-up of 91.3% ().
Table 1. Cohort demographics and comparison with individuals lost at 1 year
Figure 1. SHEFBIT Study patients and follow-up. Numbers and criteria are explained in the text
The cohort demographics are described in . The mean age of the cohort was 46.7 years (SD 20.0); median age was 45.6 (range 17.7–94.1 yrs). The majority of cases were male (826, 68.4%) and white ethnicity (1117, 92.5%). Women were older with higher GCS and more frequent falls but fewer RTCs (road traffic collisions). By comparison, the 115 cases who were lost to follow-up had milder CT scan abnormality and more social isolation but otherwise there were no significant differences.
There was a strong bias toward shorter lengths of stay probably due to higher frequency of mild injuries. The median admission was 2 days (range 1–163) and 86.3% had admission less than 10 days.
For etiology, falls (35.0%) and road traffic collisions (27.8%) were the most frequent cause of injury. Assaults were 17.8%. There was a high prevalence of past psychiatric history in 237 (19.6%), intoxication at injury 303 (25.1%) and medical comorbidity in 341 (28.3%) with 87 (7.2%) cases in the cohort on Warfarin or other oral anticoagulants.
For pre-injury employment, the majority were in employment or study at the time of injury (814 (67.4%)), while 172 (14.3%) were unemployed and 221 (18.3%) had retired. Higher SEC levels were less frequent when compared to the UK population from the last census
There was a high proportion of mild TBI (GCS 13–15) which is much closer to the real-life distribution of TBI than many other studies that focus on STBI (GCS < 9) or a specific etiology. Injury severity as categorized by GCS, was largely mild injury; 600 (49.7%) had mild TBI, 408 (33.8%) moderate TBI and 199 (16.5%) had severe TBI.
A normal CT was noted in 38% of admissions and only 7.1% showed diffuse scan abnormalities.
GOSE
The primary outcome measure was the Extended Glasgow Outcome Scale. The changes in this measure from 10 weeks to 1 year showed considerable improvement over time as shown in . However, by one year, a Good Recovery (combining both upper and lower categories) was only achieved by 42.9% of the cohort with 11.8% still showing Severe disability. While 609 individuals had improved their GOSE over a year (50.5%), the majority only increased by one level on the scale (441, 36.5%). At the same time, 138 (11.4%) deteriorated in score. ().
Table 2. Overall GOSE at 10 weeks and 1 year and transitions between groups
In order to establish the independent predictors of GOSE, an ordinal logistic regression model was estimated. Independent variables of age, gender, ethnicity, SEC, pre-injury employment, medical comorbidity, social isolation, GCS, etiology, alcohol intoxication, psychiatric history, CT scan abnormality were also entered. The results are shown in . The model was highly significant with a Nagelkerke R2 of 0.38 (p < .001)
Table 3. Ordinal regression model of 1 year GOSE. Categories described in text. OR odds ratio, *significant for p < .05
The features that were significant for worse outcome in the model were increasing age, lower GCS, etiology (assault versus other mechanisms), positive psychiatric history, alcohol intoxication at time of injury, and worse CT scan abnormality. The odds ratios are shown in .
Discussion
We have followed up a large prospective cohort of mixed TBI admissions at a large Trauma Center. Half of these were MTBI and almost 40% had normal CT scans. It is the largest prospective single center TBI outcome study with face-to-face interviews that we know of. From an initial 1322 individuals at start 1207 (91.3%) had 1 year outcome documented representing a very high retention rate. It is well known that the loss of cases is a significant problem in TBI research. This was obviated by phoning and encouraging follow-up in those who miss their appointments. We also used a new PRM team to identify and recruit all admissions with TBI. Although any study population is subject to selection bias, we believe that this cohort is characteristic of admissions with TBI and of relevance to all health professionals working in brain injury.
We have found that a considerable level of disability remains at 1 year post-TBI; only 42.9% of individuals achieved a Good Recovery (upper and lower) which had improved from 25.1% after 8 weeks. This is surprisingly low if it is considered that most of the cohort had MTBI and is worse than most previous studies. In fact, it is comparable to levels of recovery that have previously been reported only after STBI (Citation21,Citation22). At the same time there are individuals with STBI who show marked improvements and several attain a Good Recovery. There was also considerable movement across levels in both directions; only 37% retained the same status as at 10 weeks and 11.4% had a deterioration of functional status which may indicate the delayed or ongoing effects of TBI which may reach their maximal effect sometime after the injury (Citation9,Citation13).
This highlights the difficulty in predicting outcome, based only on TBI severity. This poor outcome is disappointing and illustrates the level of disability that persists after TBI and that the condition has significant repercussions for individuals, families and for society.
The majority of previous studies have only examined moderate-severe TBI and the proportion identified with Good Recovery at one year ranges considerably from 1.3%(Citation23) up to 74% (Citation22,Citation24–28). The definition of a “Good recovery” often differs. In this study, we used the structured clinical interview from the original authors of the GOSE(Citation19).
The ordinal regression model was highly significant in predicting outcome, identifying a number of associated factors. However, the pseudo R2 of 0.383 reveals that the model still leaves plenty of the variance in GOSE unexplained. In fact, this is a similar proportion to other large studies which have examined outcome from over 10,000 cases (Citation29,Citation30). It is therefore clear that there must be many factors that influence the variance in outcome which are not being captured or measured. The independent predictors of worse outcome that we have identified were increasing age, lower GCS or more severe injury, past psychiatric history and alcohol intoxication. With respect to the overall CT scan appearance, diffuse scan changes had worse outcome than normal or mildly abnormal scans but not significantly different to moderate abnormalities. With regards to etiology, assault had worse outcome than all other injury mechanisms apart from “Other” which was largely falls from height and workplace injuries. These findings may allow for targeting of certain individuals identified at high risk, for example, past psychiatric or alcohol history(Citation31).
It is also important to recognize the variables that were not associated with outcome; these were gender, ethnicity, socioeconomic class, social support. This is not to say that there are no differences in outcome by these variables, but rather that they do not have independent predictive effects beyond those of the other positive predictors. We also recognize that many features such as cognition or physical function are also likely to influence outcome. Unfortunately in a busy clinic, there is a limit to the number of variables or patient questionnaires that can be reasonably measured.
No single risk factor has been consistently identified in the literature. Studies vary considerably in their design and population selection. Another problem is that many studies dichotomize outcome into Good/Poor. Such categorization group a wide range of possible outcomes into one group with a resultant loss of dimensional quality in an ordinal outcome.
It may be expected that TBI severity would be the likeliest risk factor for outcome and many studies find such an effect (Citation4,Citation22), but negative associations have also been found (Citation6,Citation23,Citation32). Severe TBI is more likely to damage self-awareness and it has been suggested this may affect the ability to gauge one’s limitations or impairment; as a result individuals may over-estimate their recovery and report fewer problems(Citation33) which may account for some of the discrepancy between studies.
The association of increasing age with worse outcome, has also been noted previously (Citation22,Citation34,Citation35) although some find that the effect only occurs beyond around age 40 while others find a more linear relationship. Mortality rates after TBI improved for many years but the relative inertia for any further change in recent years has been attributed to the aging of the population and the increased incidence of TBI in older individuals(Citation1).
Previous attempts to create CT modeling for outcome works best in an acute setting for STBI and focus on the need for neurosurgical intervention (Citation36,Citation37). When applied across the whole TBI population, models are poor at predicting differences in outcome and a new classification system for scans is needed. The use of the “overall appearance” system allows description across the full TBI spectrum. Clearly, there will be cases where even a small lesion in a key location may result in very significant impairment. But this is the first report that we are aware of that shows a clear association of degree of CT abnormality and one year global outcome across all TBI severities.
Even those features that we have found no link with, have been found to be positive in other studies, for example, gender and ethnicity(Citation38). Socioeconomic class or income, has also been associated with outcome in some studies(Citation39) but not others(Citation35) Employment at time of injury was found to be a significant predictor in a study of STBI(Citation22) but was not noted by us.
The fact that so little consistency is apparent in the literature, illustrates the difficulty in predicting long-term outcome. Efforts to produce predictive TBI models have largely focussed on acute prognosis such as mortality with some success in the short term (Citation29,Citation30). However, even these models only account for around 0.35–0.4 of the variance in outcome and the majority of this can be attributed to three factors alone; age, motor score in GCS and pupillary reaction. The pseudo-variance in our study is similar to the IMPACT and CRASH models. It is likely that long-term outcome depends on a complex interplay of many factors including psychological, personality, and social factors (Citation5,Citation22). Many of these are difficult to measure effectively and highlight the problems in devising a model for long-term outcome. Such a model will require considerable sophistication to capture not only small changes in outcomes but also the measurement of the variables themselves.
From previous work, we know that GOSE is dynamic and that individuals can move outcome groups in both directions(Citation13). It is therefore important to continue to follow this cohort in order to document further changes. Other studies have followed up for up to 10 years and show relatively little change in outcome over a long period (Citation22,Citation26,Citation40) although seem to find better overall outcomes than we have found. This might be due to better case ascertainment in our study, especially for those who may be expected to have worse outcome. Some studies have shown an initial recovery and then a plateauing after 2 years (Citation25,Citation28). The largest cohort study to date, using the TBIMS database, found that GOSE deteriorated after 5–10 years (40,Citation41). While it is a challenge to continue a rate of follow-up beyond 1 year, we hope to report on similar long-term outcome in due course(Citation42).
It is possible to try and predict the number of expected TBI cases over a 4 year period in the region (population 400,000). One would expect around 3500 TBI cases using reported European incidence(Citation1) It follows that the pathway has identified more than half of all cases in the region and followed up. Again this is a notable feat. A similar TBI service reported a likely detection rate of only 3% of TBIs (Citation43) and a multicentre study of only STBI, identified only 1/3 of likely cases(Citation24) There is clearly considerable unrecognized or unmet need and the recognition of this is an important step in informing purchasers and health providers (Citation4).
There are a number of key strengths of this study that should be highlighted. A key achievement has been the success of a clinical team to identify and capture all TBI admissions, prospectively followed from the date of injury onwards. The cohort has minimal exclusions and hence covers the true spectrum of admitted TBI including elderly patients. Some studies have recruited very select groups such as medicolegal cases or RTC(Citation26). The systematic assessment and data collection and the capacity to chase up non-attenders were key strengths given the known attrition in TBI studies. A single observer to assess GOSE minimizes inter-observer bias.
There are some weaknesses to note. The results from a single trauma hospital may not be germane to all other regions despite our attempts to recruit as realistic a TBI population as possible. The GOSE does not distinguish between disability caused by TBI or other injury from trauma. It also has a limited number of levels and small changes may be unmeasured. The clinician who documents outcome was unblinded and could be biased. We did not measure quality of life nor any domain of cognition that would have added another dimension to the picture of TBI disability.
In summary, we have found that outcome after TBI is worse than many other studies have documented even in a largely mild injury group. In future, it is hoped to continue to follow up the group beyond one year and provide longer term results at 5 or even 10 years. We are also measuring a number of outcomes other than GOSE in order to broaden our understanding of the changes that occur across other domains, for example, psychosocial function or symptom levels. Further follow-up may provide more precise information about the exact trajectory of TBI disability and information and advice for individuals and families(Citation30). Identifying those at higher risk would allow us to target rehabilitation efforts and prompt more efficient use of resources.
Disclosure Statement
The authors report no declarations of interest.
References
- Roozenbeek B, Maas AI, Menon DK.Changing patterns in the epidemiology of traumatic brain injury. Nat Rev Neurol. 2013;9(4):231–36.doi:https://doi.org/10.1038/nrneurol.2013.22.
- Corrigan JD, Selassie AW, Orman JA (Langlois).The epidemiology of traumatic brain injury. Journal of Head Trauma Rehabilitation. 2010;25(2):72–80.doi:https://doi.org/10.1097/HTR.0b013e3181ccc8b4.
- Jacobsson LJ, Westerberg M, Lexell J.Health-related quality-of-life and life satisfaction 6–15 years after traumatic brain injuries in northern Sweden. Brain Injury. 2010;24(9):1075–86.doi:https://doi.org/10.3109/02699052.2010.494590.
- Dikmen SS, Machamer JE, Powell JM, Temkin NR.Outcome 3 to 5 years after moderate to severe traumatic brain injury11No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Arch Phys Med Rehabil. 2003;84(10):1449–57.doi:https://doi.org/10.1016/S0003-9993(03)00287-9.
- Olver JH, Ponsford JL, Curran C.Outcome following traumatic brain injury: a comparison between 2 and 5 years after injury. Brain Injury. 1996;10(11):841–48.doi:https://doi.org/10.1080/026990596123945.
- Van der Naalt J, Van Zomeren AH, Sluiter WJ, Minderhound JM. One year outcome in mild to moderate head injury: the predictive value of acute injury characteristics related to work complaints and return to work. J Neurol Neurosurg Psychiatry. 1999;66:207–13.doi:https://doi.org/10.1136/jnnp.66.2.207.
- Stucki G, Ewert T, Cieza A.Value and application of the ICF in rehabilitation medicine. Disabil Rehabil. 2002;24(17):932–38.doi:https://doi.org/10.1080/09638280210148594.
- Corrigan JD, Bogner JA, Mysiw JW, Clinchot D, Fugate L. Systematic bias in outcome studies of persons with traumatic brain injury. Arch Phys Med Rehabil. 1997;78:132–37.doi:https://doi.org/10.1016/S0003-9993(97)90253-7.
- Rosenthal M, Christensen BK, Ross TP. Depression following traumatic brain injury, Arch Phys Med Rehabil. 1998;79:90–103.
- Turner-Stokes L, Pick A, Nair A, Disler PB, Wade DT. Multi-disciplinary rehabilitation for acquired brain injury in adults of working age. Cochrane Database Syst Rev. 2015(12):CD004170.
- Andelic N, Bautz-Holter E, Ronning P, Olafsen K, Sigurdardottir S, Schanke A-K, Sveen U, Tornas S, Sandhaug M, Roe C.Does an early onset and continuous chain of rehabilitation improve the long-term functional outcome of patients with severe traumatic brain injury? J Neurotrauma. 2012;29(1):66–74.doi:https://doi.org/10.1089/neu.2011.1811.
- Singh R, Venkateshwara G, Kirkland J, Batterley J, Bruce S.Clinical pathways in head injury: improving the quality of care with early rehabilitation. Disabil Rehabil. 2012;34(5):439–42.doi:https://doi.org/10.3109/09638288.2011.608146.
- Singh R, Choudhari K, Sinha S, Mason S, Lecky F, Dawson J. Global Outcome after Traumatic Brain Injury in a prospective cohort. Clin Neurol Neurosurg. 2019;186:105526.doi:https://doi.org/10.1016/j.clineuro.2019.105526.
- Menon DK, Schwab K, Wright DW, Maas AI.Position statement: definition of traumatic brain injury. Arch Phys Med Rehabil. 2010;91(11):1637–40.doi:https://doi.org/10.1016/j.apmr.2010.05.017.
- Lecky F, Woodford M, Yates DW. Trends in trauma care in England and Wales 1989-97. UK Trauma Audit and Research Network Lancet. 2000;355:1771–75.
- Wardlaw JM.Which CT features help predict outcome after head injury? J Neurol Neurosurg Psychiatry. 2002;72(2):188–92.doi:https://doi.org/10.1136/jnnp.72.2.188.
- Linn BS, Linn MW, GUREL L.Cumulative Illness Rating Scale. J Am Geriatr Soc. 1968;16(5):622–26.doi:https://doi.org/10.1111/j.1532-5415.1968.tb02103.x.
- Chandola T, Jenkinson C.The new UK National Statistics Socio-Economic Classification (NS-SEC); investigating social class differences in self-reported health status. J Pub Health Med. 2000;22(2):182–90.doi:https://doi.org/10.1093/pubmed/22.2.182.
- Wilson JT, Pettigrew LAURAEL, Teasdale GM.Structured interviews for the Glasgow Outcome Scale and the extended Glasgow Outcome Scale: guidelines for their use. J Neurotrauma. 1998;15(8):573–85.doi:https://doi.org/10.1089/neu.1998.15.573.
- Wilde EA, Whiteneck GG, Bogner J, Bushnik T, Cifu DX, Dikmen S, French L, Giacino JT, Hart T, Malec JF, et al.Recommendations for the use of common outcome measures in traumatic brain injury research. Arch Phys Med Rehabil. 2010;91(11):1650–60.doi:https://doi.org/10.1016/j.apmr.2010.06.033.
- Corral L, Ventura JL, Herrero JI, Monfort JL, Juncadella M, Gabarrós A, Bartolomé C, Javierre CF, García-Huete L.Improvement in GOS and GOSE scores 6 and 12 months after severe traumatic brain injury. Brain Injury. 2007;21(12):1225–31.doi:https://doi.org/10.1080/02699050701727460.
- Forslund MV, Roe C, Perrin PB, Sigurdardottir S, Lu J, Berntsen S, Andelic N.The trajectories of overall disability in the first 5 years after moderate and severe traumatic brain injury. Brain Injury. 2017;31(3):329–35.doi:https://doi.org/10.1080/02699052.2016.1255778.
- Siponkoski ST, Wilson L, von Steinbüchel N, Sarajuuri J, Koskinen S.Quality of life after traumatic brain injury: Finnish experience of the QOLIBRI in residential rehabilitation. J Rehabil Med. 2013;45(8):835–42.doi:https://doi.org/10.2340/16501977-1189.
- Jourdan C, Bosserelle V, Azerad S, Ghout I, Bayen E, Aegerter P, Weiss JJ, Mateo J, Lescot T, Vigué B, et al. Predictive factors for 1-year outcome of a cohort of patients with severe traumatic brain injury (TBI): results from the Paris-TBI study. Brain Injury. 2013;27(9):1000–07.doi:https://doi.org/10.3109/02699052.2013.794971.
- Ruttan L, Martin K, Liu A, Colella B, Green RE.Long-Term Cognitive Outcome in Moderate to Severe Traumatic Brain Injury: a Meta-Analysis Examining Timed and Untimed Tests at 1 and 4.5 or More Years After Injury. Arch Phys Med Rehabil. 2008;89(12):S69–S76.doi:https://doi.org/10.1016/j.apmr.2008.07.007.
- Draper K, Ponsford J, Schonberger M.Psychosocial and emotional outcomes 10 years following traumatic brain injury. J Head Trauma Rehabil. 2007;22(5):278–87.doi:https://doi.org/10.1097/01.HTR.0000290972.63753.a7.
- Thornhill S.Disability in young people and adults one year after head injury: prospective cohort study. BMJ. 2000;320(7250):1631–35.doi:https://doi.org/10.1136/bmj.320.7250.1631.
- Sandhaug M, Andelic N, Langhammer B, Mygland A.Functional level during the first 2 years after moderate and severe traumatic brain injury. Brain Injury. 2015;29(12):1431–38.doi:https://doi.org/10.3109/02699052.2015.1063692.
- Trial Collaborators MRCCRASH, Perel P, Arango M, Clayton T, Edwards P, Komolafe E, Poccock S, Roberts I, Shakur H, Steyerberg E, et al. Predicting outcome after traumatic brain injury: practical prognostic models based on large cohort of international patients. BMJ. 2008;336:425–29.
- Maas AI, Murray GD, Roozenbeek B, Lingsma HF, Butcher I, McHugh GS, Weir J, Lu J, Steyerberg EW. International Mission on Prognosis Analysis of Clinical Trials in Traumatic Brain Injury (IMPACT) Study Group. Advancing Care for Traumatic Brain Injury: Findings from the IMPACT Studies and Perspectives on Future Research Lancet Neurol. 2013;12:1200–10.
- Backhaus SL, Ibarra SL, Klyce D, Trexler LE, Malec JF.Brain injury coping skills group: a preventative intervention for patients with brain injury and their caregivers. Arch Phys Med Rehabil. 2010;91(6):840–48.doi:https://doi.org/10.1016/j.apmr.2010.03.015.
- Whitnall L.Disability in young people and adults after head injury: 5-7 year follow up of a prospective cohort study. J Neurol Neurosurg Psychiatry. 2006;77(5):640–45.doi:https://doi.org/10.1136/jnnp.2005.078246.
- Sherer M, Bergloff P, Levin E, High WM, Oden KE, Nick T.Impaired awareness and employment outcome after traumatic brain injury. J Head Trauma Rehabil. 1998;13(5):52–61.doi:https://doi.org/10.1097/00001199-199810000-00007.
- Hukkelhoven CWPM, Steyerberg EW, Rampen AJJ, Farace E, Habbema JD, Marshall LF, Murray GD, Maas AIR.Patient age and outcome following severe traumatic brain injury: an analysis of 5600 patients. J Neurosurg. 2003;99(4):666–73.doi:https://doi.org/10.3171/jns.2003.99.4.0666.
- Mushkudiani NA, Engel DC, Steyerberg EW, Butcher I, Lu J, Marmarou A, Slieker F, McHugh GS, Murray GD, Maas AIR.Prognostic value of demographic characteristics in traumatic brain injury: results from the IMPACT study. J Neurotrauma. 2007;24(2):259–69.doi:https://doi.org/10.1089/neu.2006.0028.
- Maas AIR, Hukkelhoven CWPM, Marshall LF, Steyerberg EW.Prediction of outcome in traumatic brain injury with computed tomographic characteristics: a comparison between the computed tomographic classification and combinations of computed tomographic predictors. Neurosurgery. 2005;57(6):1173–82.doi:https://doi.org/10.1227/01.NEU.0000186013.63046.6B.
- Stenberg M, Koskinen L-OD, Jonasson P, Levi R, Stålnacke B-M.Computed tomography and clinical outcome in patients with severe traumatic brain injury. Brain Injury. 2017;31(3):351–58.doi:https://doi.org/10.1080/02699052.2016.1261303.
- Gary KW, Arango-Lasprilla JC, Ketchum JM, Kreutzer JS, Copolillo A, Novack TA, Jha A.Racial differences in employment outcome after traumatic brain injury at 1, 2, and 5 years postinjury. Arch Phys Med Rehabil. 2009;90(10):1699–707.doi:https://doi.org/10.1016/j.apmr.2009.04.014.
- Godbolt AK, Stenberg M, Lindgren M, Ulfarsson T, Lannsjo M, Stalnacke BM, Borg J. Associations Between Care Pathways and Outcome 1 year After Severe Traumatic Brain Injury. J Head Trauma Rehabil. 2015;30:E41–51.doi:https://doi.org/10.1097/HTR.0000000000000050.
- Corrigan JD, Cuthbert JP, Harrison-Felix C, Whiteneck GG, Bell JM, Miller AC, Coronado VG, Pretz CR. US population estimates of health and social outcomes 5 years after rehabilitation for traumatic brain injury. J Head Trauma Rehabil. 2014;29:E1–9.doi:https://doi.org/10.1097/HTR.0000000000000020.
- Pretz CR, Dams-O’Connor K. Longitudinal description of the Glasgow Outcome Scale-Extended for individuals in the traumatic brain injury model systems national database: a National Institute on Disability and Rehabilitation Research traumatic brain injury model systems study. Arch Phys Med Rehabil. 2013;94:2486–93.doi:https://doi.org/10.1016/j.apmr.2013.06.021.
- Raj R, Bendel S, Reinikainen M, Kivisaari R, Siironen J, Lång M, Skrivars M. Hyperoxemia and long-term outcome after traumatic brain injury. Crit Care. 2013;17:R177.
- Seeley H, Pickard J, Allanson J, Hutchinson P.The epidemiology of a specialist neurorehabilitation clinic: implications for clinical practice and regional service development. Brain Injury. 2014;28(12):1559–67.doi:https://doi.org/10.3109/02699052.2014.939717.