The Nottingham Head Injury Register: a survey of 1,276 adult cases of moderate and severe traumatic brain injury in a British neurosurgery centre

Transcription

1 The Intensive Care Society 2011 Audits and surveys The Nottingham Head Injury Register: a survey of 1,276 adult cases of moderate and severe traumatic brain injury in a British neurosurgery centre G Fuller, H Pattani, P Yeoman The Nottingham Head Injury Register was established as an audit and research tool for traumatic brain injury (TBI). Study aims included examination of the epidemiology of UK TBI and identification of factors improving management and outcome. Moderate and severe TBI patients admitted to a UK neurosurgery centre were recruited over a 10-year period from 1993 to A dataset covering all aspects of TBI, including demography, clinical features, investigations, management, complications and outcome were captured prospectively into a computerised database. The study group comprised 1,276 adult patients aged more than 16 years old. The median age of subjects was 35 years, with 78% of the population consisting of male patients. Road traffic accidents were responsible for 49% of injuries; 37% of the study population had moderate TBI and 63% severe TBI. CT brain scan was performed in 92% of cases showing a mass lesion in 41% and diffuse lesion in 59%. Traumatic subarachnoid haemorrhage was present in 34%. At 12 months 32% of patients had died, 13% were severely disabled, 22% were moderately disabled and 33% had made a good recovery. Findings were comparable with other previous large series of TBI patients from international studies. Registry data is useful to define TBI epidemiology and management, and in validation of prognostic scoring systems. Keywords: craniocerebral trauma; epidemiology; outcome measures; critical care; database Introduction Traumatic brain injury (TBI) is an important international cause of morbidity, mortality and socio-economic costs, and the World Health Organisation projects that TBI could rank as the third highest cause of the global burden of disease by Trauma research funding is low and many aspects of TBI including pathogenesis, prognosis and optimum management have yet to be fully established. 2 Clinical trials in TBI are challenging as a result of ethical considerations in unconscious trial participants, difficulty in translating experimental results into clinical practice, and heterogeneity in mechanism and pathology of injury. Over twenty multi-centre clinical trials have been conducted without demonstrating improved outcomes. 3 In contrast, prospectively collected databases have performed an important role in TBI research. Observational studies are necessary to characterise epidemiology, benchmark good practice and define the natural history of TBI. 4 Notable examples include the Traumatic Coma Data Bank s elucidation of the importance of the adverse impact of secondary brain insults 5 and the Trauma Audit and Research Network s demonstration of the importance of specialist neuro-critical care following head injury. 6 The Nottingham Head Injury Register (NHIR) is a database of moderate and severe TBI patients admitted to the Queen s Medical Centre (QMC), a UK-based neurosurgery centre containing a regional neurosurgical unit, general intensive care unit and emergency department. It was conceived as a resource for use in TBI research to: compare demographics, aetiology, severity and long-term outcomes with other head injury cohorts; survey current neurosurgical practice; and audit local head injury management. Furthermore, as a typical UK TBI population, the NHIR provides a representative sample for validation of prognostic models. This report provides a comprehensive description of the NHIR background, design, methods and results. Materials and methods Study population The QMC neurosurgical unit serves a population of 3 million, receiving referrals from 12 secondary level district general hospitals across the East Midlands region of England, while the QMC emergency department manages 160,000 new attendances per year from the city of Nottingham and surrounding countryside. TBI patients may therefore be referrals received by the neurosurgical unit from a district general hospital or may be JICS Volume 12, Number 1, January

2 admitted directly from the emergency department. Consecutive TBI patients admitted to QMC between 01/01/1993 and 31/12/2002 were entered prospectively into the NHIR. Patients of any age, including paediatric patients, were enrolled. However paediatric patients, defined as less than 16 years of age, are not considered in this paper. Inclusion criteria For inclusion into the NHIR, patients must have presented to QMC, either directly or after transfer from a referral hospital, following a TBI that resulted in moderate or severe head injury. Moderate head injury was defined as minimum Glasgow Coma Score (GCS) of between 9 and 12, severe head injury was defined as a GCS of 8 or less. 7 The defined minimum GCS followed resuscitation (oxygen, intravenous fluids and vasopressors), occurred within 48 hours of injury, and lasted a minimum of 30 minutes. Minor head injuries (GCS >12), non-traumatic causes of brain injury, patients who died prior to admission, during transport or in the emergency department were excluded. Admission GCS was defined as the GCS recorded in the first emergency department where the patient was treated, after administration of oxygen and intravenous fluids, but prior to receiving sedatives and intubation. Admission GCS of greater than 12 was therefore compatible with inclusion in the database if the GCS subsequently deteriorated within 48 hours. Patient management Following admission, patients were resuscitated according to Advanced Trauma Life Support guidelines, 8 and managed according to a standard head injury protocol consistent with recommendations from the Society of British Neurological Surgeons, 9 European Brain Injury Consortium, 10 and Brain Trauma Foundation guidelines. 11,12 Management priorities included rapid evacuation of mass lesions and prevention of secondary brain injury. Target ranges for physiological variables included: arterial oxygen saturations >93%, mean arterial blood pressure >90 mm Hg, haemoglobin >10 g/dl, temperature C, serum glucose 4-10 mmol/l, pco kpa. Cerebral perfusion pressure was maintained >60 mm Hg, using a noradrenaline infusion if necessary, after correction of hypovolaemia. Elevated intracranial pressure >20 mm Hg was managed medically with deep sedation, surface cooling to <35 C and neuromuscular blocking agents as appropriate, after exclusion of a surgical mass lesion by CT scan. Surgical management included evacuation of mass lesions, ventriculostomy and lobectomy. Data collection The NHIR was based on the Traumatic Coma Data Bank. 5 A dataset of 636 items was determined a priori, encompassing: demographic details, mechanisms of injury, pre-hospital management, clinical variables, radiological findings, laboratory results, therapeutic management, surgical interventions, outcome and rehabilitation. Research officers were trained to collect study variables from patient case notes and nursing records and enter them into a computerised database using Epi-Info version 5.0 (Center for Disease Control and Prevention, Atlanta USA). Each field was protected by entry limiters to minimise inputting errors. Data were then extracted into Microsoft Excel (Microsoft Corporation, Redmond, USA) for analysis. StatsDirect (StatsDirect, Altrincham, UK) was used for statistical analysis. A manual was developed by the research team prior to data collection for use in classifying and extracting data, standardising a defined common clinical vocabulary and rules for each of the data fields. A pilot study was performed which tested data collection and application of data rules. Accuracy was audited monthly by a separate member of the research team, and yearly reports were generated describing interim results. Head computed tomography scan details were recorded on a standardised data sheet by experienced neuro-intensivists. Information was obtained from both the neuroradiology report and reviewing the scans themselves. Patient follow-up data was collected in collaboration with each patient s general practitioner. Survivors were classified into four groups according to the Glasgow Outcome Score recorded at 12 months using a structured questionnaire: good recovery, moderate disability, severe disability, dead. 13 Due to the difficulty in accurately differentiating patients with persistent vegetative state, 14 such patients were classified as severely disabled. Non-parametric continuous data were compared using the Mann-Whitney test and categorical data compared using Chisquare tests. A two-sided P value of 0.05 was considered to be statistically significant. Injury classification TBIs were classified into two categories, focal and diffuse, based on first admission computed tomography head findings. Focal injuries were those where a high or mixed density mass lesion of greater than 25 ml was visible using the Cavalieri method, 15 or a high or mixed density lesion of any size requiring surgical evacuation was present. Focal head injuries were further sub-classified according to the dominant lesion present, into extradural haematoma, subdural haematoma, or other focal lesion (including intracerebral haematoma and contusions). Cases of mixed focal lesions where no lesion was dominant, were classified in the priority order subdural haematoma >intracerebral haematoma/contusion >extradural haematoma. Thus a patient would be classified as having a subdural haematoma even if an extradural or intracerebral haematoma of equal volume was present. Diffuse injuries were sub classified by grade (I IV) according to the Marshall Diffuse Injury Score. 16 Design features The project was funded by grants from the Nottingham Health Authority, QMC Trust and Department of Health. The project is registered under the Data Protection Act and received ethical approval from the hospital Ethics Committee. Results Demographics A total of 1,662 consecutive patients were enrolled during the 30 Volume 12, Number 1, January 2011 JICS

3 Characteristic Number of patients / percentage All patients Referred patients Directly admitted patients Intensive care patients Demographics Number of patients 1, Mean age (SD) 40.1 (18.3) 40.5 (18.5) 39.8 (18.2) 39.4 (17.3) Median age ( IQR) 35.5 ( ) 36.6 ( ) 34.8 ( ) 35.9 ( ) Male 989 (78%) 423 (77%) 566 (79%) 663 (77%) Cause of injury Road traffic accident 620 (48%) 247 (45%) 373 (52%) 482 (56%) Falls 444 (35%) 221 (40%) 223 (31%) 273 (32%) Assault 176 (14%) 69 (13%) 107 (15%) 83 (9%) Other 34 (3%) 16 (3%) 18 (3%) 22 (3%) Mechanism Blunt 1,263 (99%) 545 (98%) 718 (99%) 849 (99%) Severity Moderate 470 (37%) 181 (33%) 289 (40%) 183 (21%) Severe 806 (63%) 374 (67%) 432 (60%) 679 (79%) Admission GCS Median (range) 8 (3-15) 8 (3-15) 8 (3-15) 7 (3-15) GCS 15 with subsequent 271 (21%) 124 (22%) 147 (20%) 44 (5%) deterioration Extra-cranial injury 187 (32%) 73 (28%) 114 (34%) 168 (39%) AIS>3 (n=593) Hypoxia 300 (24%) 131 (24%) 169 (23%) 253 (29%) Hypotension 179 (14%) 77 (14%) 102 (14%) 154 (18%) Table 1 Demographic, injury mechanism and clinical features of patients included in the Nottingham Head Injury Register. Key: Referred patients: patients admitted to the QMC neurosurgical unit after transfer from a district general hospital. Directly admitted patients: patients admitted to the neurosurgical unit directly from the QMC emergency department. Intensive care patients: Patients requiring admission to the QMC intensive care unit. Data on extra-cranial injury available for 593 patients. All other data complete. SD: standard deviation, IQR: interquartile range, AIS: Abbreviated Injury Scale, GCS: Glasgow Coma Score. study period, of whom 1,276 (77%) were adults. Demographic and clinical features are summarised in Table 1. The median age of the cohort was 35 (range 16-91) with a skewed distribution towards younger ages. There was a large male majority, with 989 cases (78%). Road traffic accidents were the most important causes of TBI, accounting for 620 cases (49%). The majority of other injuries were secondary to assaults or falls. Blunt trauma was prevalent across all groups, responsible for 1,263 cases (99%). Alcohol or drug intake was a common association with TBI, identified in 734 cases (44%). Severe head injuries predominated, accounting for 806 cases (63%). The median admission GCS was 8, range A significant minority (271 patients, 21%) initially presented as minor head injuries, but subsequently their GCS deteriorated to <12 within 48 hours of presentation. Abnormal pupil reaction (one or both pupils unreactive to light) was documented in 172 cases on admission (13%). Causes of secondary brain injury (hypoxia/hypotension) were present in approximately a quarter of patients (24% and 14% respectively). Approximately a third of patients had polytrauma with 187 cases having extracranial injuries of Abbreviated Injury Scale >3 (32%, n=593). There were 555 patients admitted after secondary referral from a district general hospital (44%). These patients reached their primary emergency department in a similar time to patients arriving directly at the QMC emergency department: median 49 minutes (range ) compared to median 54 minutes (range 5-1,556), p>0.05. However, direct admission to QMC was associated with a shorter time from injury to the patient coming under neurosurgical specialist care. TBI pathology The underlying pathology of the cohort, based on computed tomography head scan classification, is summarised in Table 2. JICS Volume 12, Number 1, January

4 Characteristic All patients Referred patients Directly admitted patients CT data available 1,177 (92%) 538 (97%) 639 (89%) Focal lesion: 480 (41%) 277 (51%) 203 (32%) Extradural haematoma 97 (8%) 62 (12%) 35 (5%) Subdural haematoma 277 (23%) 160 (30%) 117 (18%) Intracerebral haematoma/contusion 106 (9%) 55 (10%) 51 (8%) Diffuse lesion 697 (59%) 263 (49%) 434 (68%) MDIS I 86 (7%) 16 (3%) 70 (11%) MDIS II 432 (37%) 170 (32%) 261 (41%) MDIS III 144 (12%) 63 (12%) 82 (13%) MDIS IV 35 (3%) 14 (3%) 21 (3%) Traumatic subarachnoid haemorrhage 399 (34%) 149 (28%) 250 (39%) Table 2 Nottingham Head Injury Register cohort pathology based on computed tomography head scan classification. Key: Referred patients: patients admitted to the QMC neurosurgical unit after transfer from a district general hospital. Directly admitted patients: patients admitted to the neurosurgical unit directly from the QMC emergency department. Denominator for percentage calculation is number of patients undergoing computed tomography head scanning. NIHR: Nottingham Head Injury Register, CT: computed tomography, MDIS: Marshal Diffuse Injury Score. Computed tomography head imaging was not performed in 92 patients and information on whether a scan was performed was unavailable in seven cases. Overall, more than half the scans were classified as diffuse injury (59%). Subdural haematoma was the most common focal lesion (23%). Traumatic subarachnoid haemorrhage was present in approximately one third of cases (31%). Significant differences in the case mix between the directly admitted and referred groups were present, with notably more focal lesions in the referred group (p<0.001). Traumatic subarachnoid haemorrhage was more frequently reported in the directly admitted group (p<0.001). Intensive care and non-surgical management There were 862 patients (68%) admitted to the QMC intensive care unit, with the remaining 414 patients admitted directly to a non critical-care specialty ward, predominantly neurosurgery (18%) and emergency medicine (7%). The intensive care subset of patients had similar demographic characteristics to non-admitted patients, but was more severely injured with lower admission GCS and a higher proportion of pupillary non-reactivity, hypoxia and polytrauma (p<0.01, Table 1). Median length of intensive care unit stay was four days (range 1-94) and patients admitted to intensive care had a significantly longer median total hospital length of stay (10 days (range 1-227) compared to six days (range 1-91), p<0.001). Of patients admitted to intensive care, 93% were ventilated, with a median length of ventilation of three days (range 1-57); further details of intensive care management are summarised in Table 3. Neurosurgical management Data on neurosurgical management were available for 1,115 patients. Of these, 600 cases underwent operative management, which is summarised in Table 3. During the study period intracranial pressure probe insertion was performed in the operating theatre, and for 21% of patients probe insertion was their only procedure. There were 30% of patients who had evacuation of an intracranial haematoma. Referred patients with time critical mass lesions (extradural/subdural haematoma) took a significantly longer time from injury to evacuation than those admitted directly to the QMC emergency department (median 345 minutes (interquartile range minutes) vs 212 minutes (interquartile range minutes), p<0.05). Outcome Outcome data were available at hospital discharge for 1,274 patients (99.8%) and at 12 months post injury for 1,225 patients (96%). There were 335 cases who died in hospital (26%), and of these, 251 cases died within one week of admission to QMC (75%). By 12 months, 390 patients had died (32%), 555 patients (45%) had experienced an unfavourable outcome (Glasgow Outcome Score: severe disability, persistent vegetative state or dead) and 670 patients (55%) had recovered to a favourable outcome (Glasgow Outcome Score: good recovery or moderate disability). Discussion Prospective databases have become established as an important resource in investigating many aspects of TBI. The NHIR provides a picture of contemporary practice in a UK neurosurgical centre and aims to contribute to the accumulating international TBI dataset. Strengths and limitations The extensive defined dataset, recruitment of consecutive TBI patients, long-term independent follow up and relative data completeness are particular strengths of this study. The NHIR performs well on evaluation with established quality 32 Volume 12, Number 1, January 2011 JICS

5 Management Admitted to 862 intensive care Number of patients/ percentage Ventilated 800 (93%) Tracheostomy 299 (27%) ICP monitored 469 (55%) Controlled hypothermia 235 (27%) Noradrenaline infusion 245 (28%) Underwent 600 neurosurgical procedure ICP probe insertion only 235 (21%) ICP probe insertion 234 (21%) + other neurosurgical procedure Other neurosurgical 131 (12%) procedure only Haematoma evacuated 329 (30%) Table 3 Summary of intensive care and neurosurgical management of Nottingham Head Injury Register patients. Directory of Nottingham Head Directory of Clinical Clinical Databases Injury Register Databases database Criteria score (1 to 4) median score (1 to 4) Representative 3 3 of country Completeness of 4 3 recruitment Variables Included 4 3 Completeness of 2 2 variables Collection of 3 4 raw data Explicit definitions 4 2 Explicit rules 4 2 Reliability of coding 3 1 Independence of 3 4 observations Data validation 3 3 Table 4 Performance of Nottingham Head Injury Register against Directory of Clinical Databases criteria. Key: Denominator for percentage calculation is number of patients admitted to intensive care or undergoing neurosurgical management. Data on neurosurgical management available for 1115 patients. All other data complete. ICP: Intracranial pressure. assessment instruments. It scores highly across all domains of the Directory of Clinical Databases rating tool, 17 exceeding or equalling the median values of other registered databases except for collection of continuous data (Table 4). 18 There are incomplete 12-month Glasgow Outcome Score data, which is a limitation potentially introducing bias. Difficulty in obtaining outcomes is a common problem in TBI databases with a mean of 6% missing outcome data across published series (range 0-18%). 5,19-27 However, many series with apparently more complete follow up have used discharge GCS or death as a more convenient, but less informative alternative to 6 or 12 month Glasgow Outcome Score. The polytrauma and neurosurgical intervention variables also had incomplete data collection, potentially undermining the internal validity of the NHIR. Missing data are a major problem in databases and the median Directory of Clinical Databases score of registered databases for data completeness is only 2 (1 is the lowest and 4 is the highest possible rating). 18 Although it is probable that data are missing at random, we cannot exclude differential measurement errors. However, the face validity of the NHIR results are supported by internal coherence and consistency of the findings. Assessment of computed tomography scans may also have introduced bias. Original scans were not always available for referred patients and some assessments were made using the radiology report alone, often reported by a non-specialist neuroradiologist; also, some available reports were not of high quality. Key: Nottingham Head Injury Register ratings compared with the median from all 154 databases in the Directory of Clinical Databases. It is possible that this may explain differences in reported subarachnoid haemorrhage rate compared to other studies. The QMC is comparable to other UK neurosurgery centres in terms of TBI admission rate, centre size and facilities. 24 However, the proportion of direct admissions and the general ICU may be unrepresentative, with many other neurosurgical centres having a dedicated neuro-intensive care unit and a smaller proportion of direct admissions. A significant variation in TBI management has been reported both nationally and internationally, 24,27,28 despite most centres basing management on similar treatment guidelines, which may challenge the external validity of these results. Furthermore, the NHIR transfer times for referred patients with expanding mass lesions may be longer than those achieved in other international systems of trauma care. 29 Comparison with other databases Over the past 40 years, a growing TBI literature, including population-based studies and case series, have examined the epidemiology of moderate and severe TBI regionally, nationally and internationally (summarised in Table 5). 5,19-27 However, few prospective databases have accrued large numbers of patients and included data on CT classifications, neurosurgery and community outcome. The NIHR is a study similar in size and scope to many of these important TBI cohorts. Each database has a heterogeneous design with differences in inclusion criteria, TBI management, sample sizes, time frames and computed tomography head scan classification. The subsequent difficulty in comparing datasets emphasises the importance of developing a standardised definition for JICS Volume 12, Number 1, January

6 Study ICD TCDB IMPCT UK4 NHIR EBIC ICNARC NeuroL JNCTB CRASH AHTR Year Country UK/NL/US US US/Europe UK UK Europe UK Italy Japan International Greece Sample size 2, , ,276 1,005 11, ,100 10,008 1,000 TBI severity S S M/S S M/S M/S S Mi/M/S S Mi/M/S Mi/M/S Inclusion Any Any Any Any >16 >16 >16 Any >5 >16 >14 age (years) Direct admissions (%) Median/ 32/- 25/- -/34 29/- 35/40 -/38 44/- -/33 48/- 37/- -/43 mean age Age range Male (%) Falls (%) Road traffic accidents (%) Focal brain injury (%) Diffuse brain injury (%) Subarachnoid haemorrhage (%) Intracranial surgery (%) Mortality (%) Favourable outcome (%) Missing follow up data (%) Table 5 Characteristics of Nottingham Head Injury Register compared to prospective TBI surveys of greater than 500 patients. Key: 1. Measured at discharge; 2. measured at 14 days; 3. measured at 6 months; 4. measured at 12 months; ' - ' data unavailable. ICD; International coma databank, NTCD; National Traumatic Coma Databank, IMPCT; International Mission for Prognosis and Clinical Trial study, UK4S; United Kingdom 4 centres study, NHIR; Nottingham Head Injury Register, EBIC; European Brain Injury Consortium survey of Head Injuries, ICNARC; Intensive Care National Audit and Research Centre study, NeuroL; Italian Neurolink study, JNTCD; Japanese National Traumatic Coma Databank, CRASH; Corticosteroid Randomisation after Significant Head Injury cohort, AHTR; Athens head trauma registry, UK; United Kingdom, US; United States of America, NL; Netherlands, TBI; Traumatic brain injury, Mi; mild, M: moderate, S; severe. reporting TBI parameters. However, comparison of the NHIR with previous prospective series reveals a number of interesting observations. A strong male predominance in TBI is seen across all studies including the NHIR (range 72-80%). The NIHR median age of 35 is similar to most other series. However differences in age inclusion criteria make direct comparison difficult. The older ages seen in the Japanese National Traumatic Coma Databank may reflect the older demography of the general population in this country, 26 while lower mean ages of other studies could reflect the inclusion of paediatric patients. The proportion injured in road traffic accidents were very similar in the International Coma Databank, 22 UK Four Centres study and the current study (52-65%), 24 whereas in the North American, Greek and Italian series many more patients (71-75%) received their injury as a result of a road traffic accident, possibly reflecting local traffic conditions. The comparison of TBI pathology based on computed tomography head imaging is hindered by differing approaches to classification. In the International Coma Databank, the first 700 patients were studied before computed tomography was available 22 and further difficulties arise due to timing of head scans and inter-observer variation in interpretation. The NIHR (in common with the majority of studies) used the Marshall Diffuse Injury Score and included only admission head scans. The Japanese series used a novel classification system divided into focal, diffuse, or focal and diffuse, 26 while NeuroLink and the Intensive Care National Audit and Research Centre study 34 Volume 12, Number 1, January 2011 JICS

7 did not define their classification method. 19,21 Interestingly, the NHIR showed a higher proportion of diffuse brain injuries compared with the UK 4 centres series, 24 but comparable with other European and American studies. This may reflect the relatively high proportion of direct TBI admissions to QMC with diffuse brain injury rather than a selected population of patients with focal lesions transferred for neurosurgical intervention to hospitals in the 4 centres sample. This correlates with the experience of the National Traumatic Coma Databank. 5 The spectrum of severity of TBI reported in each cohort is influenced by individual inclusion criteria. Other major databases including only moderate and severe TBI show similar proportion of moderate TBI as the NHIR. Abnormal pupillary reaction, reflecting severe head injury, was recorded in 13% of our patients, comparable to figures cited in other studies with comparable entry criteria (5-41%). Rates of secondary insults are consistent between major TBI databases, with observed hypoxia ranging from 18% to 27% and hypotension ranging from 18% to 25%. The NHIR reports similar prevalence of hypoxia (24%), but a lower recorded hypoxia rate (14%). Current TBI management is guided by consensus and evidence-based guidelines, aiming to standardise effective treatment and avoid secondary brain injury Treatment protocols have increasingly emphasised the importance of intracranial pressure monitoring in TBI neurocritical care. Correspondingly intracranial pressure monitoring use has increased, from 35% of severe TBI cases seen in the earlier International Coma Databank, 22 to 75% in the more recent Athens Head Trauma Registry. 27 The intermediate intracranial pressure monitoring rate of 55% observed in the NHIR is similar to other studies conducted within the same time period, and is consistent with national neurosurgical unit surveys from this period indicating routine monitoring was performed in only 49% of centres. 30,31 Early neurosurgical intervention, other than burr hole for intracranial pressure monitoring, was performed in a remarkably similar proportion of patients in contemporary studies (32% to 39%), with the NIHR showing a comparable rate of 32%. A higher level of 47% was observed in the International Coma Databank, 22 which may reflect lack of access to intracranial imaging in this earlier study. Outcome was broadly similar to North American, European, Japanese and UK series where mortality ranged from 27-61%. However, comparison of outcome results between series of patients is difficult without a robust mechanism for accounting for case mix. The higher mortality observed in the International Coma Databank may be a result of less advanced neurocritical care available when this study was performed. 22 In the Athens Head Trauma Registry, a high level of mortality of 61% was observed, which was explained by the investigators as secondary to deficiencies in transfer policies and primary health care. 27 Conclusions The Nottingham Head Injury Register provides extensive information on current British TBI epidemiology, with sample demographics, injury characteristics, management and outcome shown to be comparable with other large international series of TBI patients. Furthermore, registry data will be important in evaluating management strategies, identifying prognostic factors and validating outcome prediction models. References 1. Finfer SR, Cohen J. Severe traumatic brain injury. Resuscitation 2001; 48: Maas AI. Traumatic brain injury: simple data collection will improve the outcome. Wien Kiln Wochenshr 2007;119: Naryan RK, Michel ME, Ansell B et al. Clinical trials in head injury. J Neurotrauma 2002;19: Foulkes MA. Neurosurgical databases. J Neurosurg 1991;75:S1-S6. 5. Foulkes AM, Eisenberg MH, Jane JA et al. The traumatic coma data bank: design, methods, and baseline characteristics. J Neurosurg 1991; 75:S8-S Patel HC, Bouamra O, Woodford M et al. Trends in head injury outcome from 1989 to 2003 and the effect of neurosurgical care: an observational study. Lancet 2005;366: Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet 1974;2: American College of Surgeons Committee on Trauma. Advanced Trauma Life Support for doctors, fifth edition. Chicago; American College of Surgeons; Bartlett J, Kett-White R, Mendelow AD et al. Guidelines for the initial management of head injuries: recommendations from the Society of British Neurological Surgeons. Br J Neurosurg 1998;12: Maas AI, Dearden M, Teasdale GM et al. EBIC guidelines for management of severe head injury in adults. Acta Neurochir (Wien) 1997;139: Bullock R, Chesnut R, Clifton G. Guidelines for the management of severe head injury. Brain Trauma Foundation. American Association of Neurological Surgeons Joint Section on Neurotrauma and Critical Care. J Neurotrauma 1996;13: Brain Trauma Foundation. Guidelines for the management of severe traumatic brain injury. J Neurotrauma 2007;24:S Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet 1975;1: Andrews K, Murphy L, Munday R, Littlewood C. Misdiagnosis of the vegetative state: retrospective study in a rehabilitation unit. BMJ 1996; 313: Stocchetti N, Croci M, Spagnoli D et al. Mass volume measurement in severe head injury: accuracy and feasibility of two pragmatic methods. J Neurol Neurosurg Psychiatry 2000; 68: Marshall LF, Marshall SB, Klauber MR, Clark MB. A new classification of head injury based on computerized tomography. J Neurosurg 1991;75:S14-S Black N, Payne M. Directory of clinical databases: improving and promoting their use. Qual Saf Health Care 2003;12: NHS Information Centre London: Directory of Clinical Databases; cited June Citerio G, Stocchetti N, Cormio M, Beretta L. Neuro-Link, a computerassisted database for head injury in intensive care. Acta Neurochir (Wien) 2000;142: Roberts I, Yates D, Sandercock P et al. Effect of intravenous corticosteroids on death within 14 days in 10,008 adults with clinically significant head injury (MRC CRASH trial): randomised placebo controlled trials. Lancet 2004;364: Hyam JA, Welch CA, Harrison DA, Menon DK. Case mix, outcomes and comparison of risk prediction models for admission to adult, general and specialist critical care units for head injury: a secondary analysis of the ICNARC case mix programme database. Crit Care 2006;10:S Jennet B, Teasdale G, Galbraith S et al. Severe head injury in three countries. J Neurol Neurosurg Psychiatry 1977;40: JICS Volume 12, Number 1, January

8 23.Marmarou A, Lu J, Butcher I et al. The IMPACT database on traumatic brain injury: design and description. J Neurotrauma 2007;24: Murray LS, Teasdale GM, Murray GD et al. Head Injuries in four British neurosurgical centres. Br J Neurosurgery 1999;13: Murray GD, Teasdale GM, Braakman R et al. The European brain injury consortium survey of head injuries. Acta Neurochir (Wien) 1999;141: Nakamura N, Yamaura A, Shigemori M et al. Final report of the Japan Neurotrauma Data Bank project ; 1,002 cases of traumatic brain injury. Neurol Med Chir (Tokyo) 2006;46: Stranjalis G, Bouras T, Korfias S et al. Outcome in 1,000 head injury hospital admissions: The Athens head injury trauma registry. J Trauma 2008;65: Clifton GL, Choi SC, Miller ER et al. Intercenter variance in clinical trials of head trauma experience of the National Acute Brain Injury Study: Hypothermia. J Neurosurg 2001;95: Wilberger JE, Harris M, Diamond DL. Acute subdural haematoma: morbidity, mortality, and operative timing. J Neurosurg 1991;74: Ghajar J, Hairi R, Narayan RK et al. Survey of critical care management of comatose head-injured patients in the United States. Crit Care Med 1995;23: Jeevaratnam DR, Menon D. Intensive care management of severely head injured patients in the United Kingdom. BMJ 1996;312: Gordon Fuller Academic Clinical Fellow in Emergency Medicine, Trauma Audit and Research Network, University of Manchester gordonfuller@doctors.org.uk Hina Pattani Specialist Registrar in Intensive Care, Intensive Care Unit, Queen s Medical Centre Campus, Nottingham, University Hospitals NHS Trust Paddy Yeoman Consultant Intensivist, Intensive Care Unit, Queen s Medical Centre Campus, Nottingham University Hospitals NHS Trust 36 Volume 12, Number 1, January 2011 JICS