A New Approach to Outcome Prediction in Trauma: A Comparison With the TRISS Model

Prediction of mortality among severely injured trauma patients A comparison between TRISS and machine learning-based predictive models

BACKGROUND

Given the huge impact of trauma on hospital systems around the world, several attempts have been made to develop predictive models for the outcomes of trauma victims. The most used, and in many studies most accurate predictive model, is the "Trauma Score and Injury Severity Score" (TRISS). Although it has proven to be fairly accurate and is widely used, it has faced criticism for its inability to classify more complex cases. In this study, we aimed to develop machine learning models that better than TRISS could predict mortality among severely injured trauma patients, something that has not been studied using data from a nationwide register before.

METHODS

Patient data was collected from the national trauma register in Sweden, SweTrau. The studied period was from the 1st of January 2015 to 31st of December 2019. After feature selection and multiple imputation of missing data three machine learning (ML) methods (Random Forest, eXtreme Gradient Boosting, and a Generalized Linear Model) were used to create predictive models. The ML models and TRISS were then tested on predictive ability for 30-day mortality.

RESULTS

The ML models were well-calibrated and outperformed TRISS in all the tested measurements. Among the ML models, the eXtreme Gradient Boosting model performed best with an AUC of 0.91 (0.88-0.93).

CONCLUSION

This study showed that all the developed ML-based prediction models were superior to TRISS for the prediction of trauma mortality.

Adaptive Bayesian learning for making risk-aware decisions: A case of trauma survival prediction

Decision tree (DT) models provide a transparent approach to prediction of patient's outcomes within a probabilistic framework. Averaging over DT models under certain conditions can deliver reliable estimates of predictive posterior probability distributions, which is of critical importance in the case of predicting an individual patient's outcome. Reliable estimations of the distribution can be achieved within the Bayesian framework using Markov chain Monte Carlo (MCMC) and its Reversible Jump extension enabling DT models to grow to a reasonable size. Existing MCMC strategies however have limited ability to control DT structures and tend to sample overgrown DT models, making unreasonably small partitions, thus deteriorating the uncertainty calibration. This happens because the MCMC explores a DT model parameter space within a limited knowledge of the distribution of data partitions. We propose a new adaptive strategy which overcomes this limitation, and show that in the case of predicting trauma outcomes the number of data partitions can be significantly reduced, so that the unnecessary uncertainty of estimating the predictive posterior density is avoided. The proposed and existing strategies are compared in terms of entropy which, being calculated for predicted posterior distributions, represents the uncertainty in decisions. In this framework, the proposed method has outperformed the existing sampling strategies, so that the unnecessary uncertainty in decisions is efficiently avoided.

Stress-Induced (Not Diabetic) Hyperglycemia is Associated With Mortality in Geriatric Trauma Patients

INTRODUCTION

Stress-induced hyperglycemia (SIH) is associated with worse outcomes among trauma patients. It is also known that injured geriatric patients have higher mortality when compared to younger patients. We sought to investigate the association of all levels of SIH with mortality among geriatric trauma patients at a level 1 academic trauma center. We hypothesized that SIH in the geriatric trauma population would be associated with increased mortality.

METHODS

A retrospective review of all geriatric patients admitted to our level 1 trauma center over a 3-year period (January 2018-December 2020) was performed using the institutional trauma database. Data collected included demographics, injury severity score (ISS), emergency department (ED) blood glucose level, ED systolic blood pressure (SBP), and mortality. Patients were divided into 4 groups based on emergency room blood glucose level, as follows: normoglycemic (<120 mg/dL), mild hyperglycemia (120-150 mg/dL), moderate hyperglycemia (151-199 mg/dL), and severe hyperglycemia (≥200 mg/dL). Multivariable logistic regression analysis was performed to evaluate the association of SIH and in-hospital mortality adjusting for ISS, age, comorbidities, and ED SBP.

RESULTS

A total of 4432 geriatric trauma patients were admitted during the study period, of which 3358 patients (75.8%) were not diabetic. There were 2206 females (65.7%), 2993 were White (89.2%), with a mean age of 81.5 y. There were 114 deaths (3.4%). Univariate results showed that there was a statistically significant association between mortality and glucose groups (P < 0.01). The number of deaths in the four glucose groups were, as follows: 30 (2.0%), 32 (3.8%), 20 (6.2%), and 10 (12.2%), respectively. Multivariable logistic regression analysis results showed that compared to the normoglycemic group, the risk of death was higher in the mild, moderate, and severe glucose groups, as follows: mild group (OR 1.80, 95% confidence interval [CI] 1.04-3.13, P 0.04), moderate group (OR 2.53, 95% CI 1.34-4.80, P < 0.01), and severe group (OR 5.04, 95% CI 2.18-11.67, P < 0.01).

CONCLUSIONS

Mild, moderate, and severe SIH are statistically significant predictors of death among geriatric trauma patients independently of ISS, age, comorbidities, and SBP.

Can the 5-item Modified Frailty Index Predict Outcomes in Geriatric Trauma? A National Database Study

BACKGROUND

Frailty results in increased vulnerability to adverse outcomes following trauma. We investigated the association between the 5-item modified frailty index (mFI-5) and outcomes in geriatric trauma patients.

METHODS

The 2011-2016 Trauma Quality Improvement Program database was used to study outcomes in patients ≥ 65 years old. The mFI-5 was measured and categorized into no frailty (mFI-5 = 0), moderate frailty (mFI-5 = 0.2), and severe frailty (mFI-5 ≥ 0.4). Multivariable logistic regression analyses were performed to identify independent factors of mortality and complications.

RESULTS

26,963 cases met the inclusion criteria, of whom 25.5% were not frail, 38% were moderately frail, and 36.6% were severely frail. Mean age (± SD) was 76 ± 7 years, 61.5% were male, and 97.8% sustained blunt injuries. Median Injury Severity Score (ISS) was 17 (IQR = 10-26), and the median Glasgow Coma Scale was 15 (IQR = 12-15). Overall mortality was 30.6%. Factors independently associated with mortality were age (OR = 1.07 per year, 95%CI 1.06-1.07), blunt trauma (OR = 1.44, 95%CI 1.19 -1.75), ISS (OR = 1.04 per unit increase in ISS, 95%CI 1.03-1.04), and severe frailty (OR = 1.23, 95%CI 1.15-1.32). Interestingly, male sex and GCS appeared to be protective factors with OR of 0.88 (95%CI 0.83 - 0.93) and 0.89 per point change in GCS (95%CI 0.88-0.9), respectively. Moderate (OR = 1.27, 95%CI 1.19-1.25) and severe frailty (OR = 1.49, 95%CI 1.-1.59) were significantly associated with in-hospital complications.

CONCLUSION

Moderate and severe frailty were significant predictors of complications. Only severe frailty was associated with short-term mortality. The mFI-5 can be used as an objective measure to stratify risks in geriatric trauma.

OUP accepted manuscript

Background

Risk-adjusted mortality (RAM) analysis and comparisons of clinically relevant subsets of trauma patients allow hospitals to assess performance in different processes of care. The aim of the study was to develop a RAM model and compare RAM ratio (RAMR) in subsets of severely injured adult patients treated in university hospitals (UHs) and emergency hospitals (EHs) in Sweden.

Methods

This was a retrospective study of the Swedish trauma registry data (2013 to 2017) comparing RAMR in patients (aged 15 years or older and New Injury Severity Score (NISS) of more than 15) in the total population (TP) and in multisystem blunt (MB), truncal penetrating (PEN), and severe traumatic brain injury (STBI) subsets treated in UHs and EHs. The RAM model included the variables age, NISS, ASA Physical Status Classification System Score, and physiology on arrival.

Results

In total, 6690 patients were included in the study (4485 from UHs and 2205 from EHs). The logistic regression model showed a good fit. RAMR was 4.0, 3.8, 7.4, and 8.5 percentage points lower in UH versus EH for TP (P < 0.001), MB (P < 0.001), PEN (P = 0.096), and STBI (P = 0.005), respectively. The TP and MB subsets were subgrouped in with (+) and without (−) traumatic brain injury (TBI). RAMR was 7.5 and 7.0, respectively, percentage points lower in UHs than in EHs in TP + TBI and MB + TBI (both P < 0.001). In the TP–TBI (P = 0.027) and MB–TBI (P = 0.107) subsets the RAMR was 1.6 and 1.8 percentage points lower, respectively.

Conclusion

The lower RAMR in UHs versus EH were due to differences in TBI-related mortality. No evidence supported that Swedish EHs provide inferior quality of care for trauma patients without TBI or for patients with penetrating injuries.

Pediatric Trauma: Six Years of Experience in a Swiss Trauma Center

OBJECTIVES

The purpose of this study was to provide an internationally comparable overview of pediatric trauma of the University Hospital of Lausanne to improve the care of children.

METHODS

We analyzed the data from all injured children (<16 years of age) listed in our trauma registry from 2011 to 2016. These children were admitted to the resuscitation room after prehospital triage. Our data were analyzed using descriptive statistics.

RESULTS

We included 327 children. Sixty-three percent were male, and the median age was 8 years. Severe trauma (Injury Severity Score (ISS), >15) occurred in 97 children. The principal mechanisms of injury were falls (45%), traffic accidents (29%), and burns (14%). The most frequently affected areas were the head and external body regions. Intensive care admissions amounted to 27%. Twenty percent of patients underwent immediate surgery (wound care, neurosurgery, and orthopedic surgery). The overall mortality rate was 5.5%, with a median ISS of 9. The mortality of severe trauma was 17.5%, with a median ISS of 22. Half of the children died within 6 hours. The main causes of death were falls from greater than 5 m and traffic accidents as pedestrians.

CONCLUSIONS

The demographics and patterns of injury in the pediatric trauma population are similar to other European pediatric trauma centers, but the mortality and the severity of injuries can vary (United Kingdom, 3.7%, median ISS of 9; Denmark, 7.3%, median ISS of 9; and Germany, 13.4%, median ISS of 25). The elevated early mortality rate suggests that improvements in prehospital care and early resuscitation could decrease mortality.

Development and internal validation of China mortality prediction model in trauma based on ICD-10-CM lexicon: CMPMIT-ICD10

Background:

Models to predict mortality in trauma play an important role in outcome prediction and severity adjustment, which informs trauma quality assessment and research. Hospitals in China typically use the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) to describe injury. However, there is no suitable prediction model for China. This study attempts to develop a new mortality prediction model based on the ICD-10-CM lexicon and a Chinese database.

Methods:

This retrospective study extracted the data of all trauma patients admitted to the Beijing Red Cross Emergency Center, from January 2012 to July 2018 (n = 40,205). We used relevant predictive variables to establish a prediction model following logistic regression analysis. The performance of the model was assessed based on discrimination and calibration. The bootstrapping method was used for internal validation and adjustment of model performance.

Results:

Sex, age, new region-severity codes, comorbidities, traumatic shock, and coma were finally included in the new model as key predictors of mortality. Among them, coma and traumatic shock had the highest scores in the model. The discrimination and calibration of this model were significant, and the internal validation performance was good. The values of the area under the curve and Brier score for the new model were 0.9640 and 0.0177, respectively; after adjustment of the bootstrapping method, they were 0.9630 and 0.0178, respectively.

Conclusions:

The new model (China Mortality Prediction Model in Trauma based on the ICD-10-CM lexicon) showed great discrimination and calibration, and performed well in internal validation; it should be further verified externally.

Trauma-related mortality of patients with severe psychiatric disorders: population-based study from the French national hospital database

BACKGROUND

Most research on mortality in people with severe psychiatric disorders has focused on natural causes of death. Little is known about trauma-related mortality, although bipolar disorder and schizophrenia have been associated with increased risk of self-administered injury and road accidents.

AIMS

To determine if 30-day in-patient mortality from traumatic injury was increased in people with bipolar disorder and schizophrenia compared with those without psychiatric disorders.

METHOD

A French national 2016 database of 144 058 hospital admissions for trauma was explored. Patients with bipolar disorder and schizophrenia were selected and matched with mentally healthy controls in a 1:3 ratio according to age, gender, social deprivation and region of residence. We collected the following data: sociodemographic characteristics, comorbidities, trauma severity characteristics and trauma circumstances. Study outcome was 30-day in-patient mortality.

RESULTS

The study included 1059 people with bipolar disorder, 1575 people with schizophrenia and their respective controls (n = 3177 and n = 4725). The 30-day mortality was 5.7% in bipolar disorder, 5.1% in schizophrenia and 3.3 and 3.8% in the controls, respectively. Only bipolar disorder was associated with increased mortality in univariate analyses. This association remained significant after adjustment for sociodemographic characteristics and comorbidities but not after adjustment for trauma severity. Self-administered injuries were associated with increased mortality independent of the presence of a psychiatric diagnosis.

CONCLUSIONS

Patients with bipolar disorder are at higher risk of 30-day mortality, probably through increased trauma severity. A self-administered injury is predictive of a poor survival prognosis regardless of psychiatric diagnosis.

Prediction of mortality in critically-ill elderly trauma patients: a single centre retrospective observational study and comparison of the performance of trauma scores

Background

Trauma in the elderly (≥ 65 years) population is increasing. This study compares the performance of trauma scoring systems in predicting 30-day mortality among the traumatised elderly patients admitted to the intensive care unit in a major trauma centre.

Methods

We collected retrospective data for all elderly trauma patients admitted to our intensive care units between January 2012 and December 2017. We assessed Injury Severity Score (ISS), Geriatric Trauma Outcome Score (GTOS) and the Trauma Audit and Research Network’s (TARN) Probability of Survival (Ps17) between survivors and non-survivors. Receiver operator characteristic (ROC) curves were used to assess the performance of these scoring systems.

Results

There were 255 elderly trauma patients with overall 30-day survival of 76%. There was a statistically significant difference in ISS, GTOS and Ps17 scores between survivors and non-survivors (p < 0.001). The area under the ROC curve (AUROC) was statistically significant for all 3, with AUROC of 0.66 (95% CI 0.59–0.74) for the ISS, 0.68 (95% CI 0.61–0.76) for the GTOS and 0.79 (95% CI 0.72–0.85) for the Ps17. The optimal cut-off points were ≥ 28, ≥ 142, ≤ 76.73 for ISS, GTOS and Ps17, respectively.

Conclusion

Both ISS and GTOS scoring systems preformed equally in predicting 30-day mortality in traumatised elderly patients admitted to the intensive care unit, however neither were robust enough to utilise in clinical practise. The Ps17 performed more robustly, although was not developed for prognosticating on individual patients. Larger prospective studies are needed to validate these scoring systems in critically-ill elderly traumatised patients, which may help to facilitate early prognostication.

Helicopter and ground emergency medical services transportation to hospital after major trauma in England: a comparative cohort study

Background

The utilization of helicopter emergency medical services (HEMS) in modern trauma systems has been a source of debate for many years. This study set to establish the true impact of HEMS in England on survival for patients with major trauma.

Methods

A comparative cohort design using prospectively recorded data from the UK Trauma Audit and Research Network registry. 279 107 patients were identified between January 2012 and March 2017. The primary outcome measure was risk adjusted in-hospital mortality within propensity score matched cohorts using logistic regression analysis. Subset analyses were performed for subjects with prehospital Glasgow Coma Scale <8, respiratory rate <10 or >29 and systolic blood pressure <90.

Results

The analysis was based on 61 733 adult patients directly admitted to major trauma centers: 54 185 ground emergency medical services (GEMS) and 7548 HEMS. HEMS patients were more likely male, younger, more severely injured, more likely to be victims of road traffic collisions and intubated at scene. Crude mortality was higher for HEMS patients. Logistic regression demonstrated a 15% reduction in the risk adjusted odds of death (OR=0.846; 95% CI 0.684 to 1.046) in favor of HEMS. When analyzed for patients previously noted to benefit most from HEMS, the odds of death were reduced further but remained statistically consistent with no effect. Sensitivity analysis on 5685 patients attended by a doctor on scene but transported by GEMS demonstrated a protective effect on mortality versus the standard GEMS response (OR 0.77; 95% CI 0.62 to 0.95).

Discussion

This prospective, level 3 cohort analysis demonstrates a non-significant survival advantage for patients transported by HEMS versus GEMS. Despite the large size of the cohort, the intrinsic mismatch in patient demographics limits the ability to statistically assess HEMS true benefit. It does, however, demonstrate an improved survival for patients attended by doctors on scene in addition to the GEMS response. Improvements in prehospital data and increased trauma unit reporting are required to accurately assess HEMS clinical and cost-effectiveness.

Factors affecting mortality in major trauma patients in Trinidad and Tobago – A view from the developing world

Introduction

There are few data on major trauma in the developing world. This study investigated the characteristics and outcomes of seriously injured patients in Trinidad and Tobago, using Trauma and Injury Severity Score (TRISS) methodology. We also aimed to assess the predictive accuracy of the TRISS model in patients in Trinidad and Tobago.

Methods

Retrospective data from major trauma patients attending the Emergency Department of a tertiary hospital in Trinidad between 2010 and 2014 were analysed. Patients ≥18 years having an Injury Severity Score >15 were included. The impact of age, gender, comorbidities, mechanisms and patterns of injury on mortality was investigated. Using TRISS methodology, predicted mortality was calculated and compared to actual mortality.

Results

Of 323 patients analysed, 284 were male and 24 were aged ≥65 years. The commonest injury mechanisms in younger people were motor vehicle accidents (34.1%) and stabbings (30.8%) compared to falls (66.7%) and motor vehicle accidents (20.8%) in people aged ≥65 years. The commonest areas injured were the chest in younger patients (81.9%) and the head and neck in patients aged ≥65 years (58.3%). Women’s mortality rates were similar to men (RR 1.8; 95% CI 0.7–4.9). Mortality was higher with age ≥65 years (RR 7.0; 95% CI 3.1–15.9), blunt trauma (RR 7.6; 95% CI 1.8–32.4) and Charlson Comorbidity Index of 1 or more (RR 3.2; 95% CI 1.3–8.0). The TRISS model performed well at lower ISS scores and was excellent at predicting survival (discrimination statistic 0.94).

Conclusion

Multiple factors influence mortality in major trauma patients in Trinidad and Tobago, including age, co-morbidities and injury mechanism. TRISS methodology accurately predicted survival in this population but was better at predicting mortality in patients with lower Injury Severity Score.

Clinical forensic medicine in Eastern Denmark: Organisation and assessments

Internationally, clinical forensic medicine (CFM) is diverse in content and conduct, and forensic medical methods are not always evidence based. The first step towards evidence-based practice is to achieve a thorough knowledge of international diversity, which necessitates that CFM practitioners provide information about their national practice. This paper’s aim is to describe the organisation of CFM in Denmark, exemplified by the set-up in Eastern Denmark, and the types of assessments performed. In Denmark, forensic medicine is a board-certified specialty under the health authorities, with mandatory qualifications. The Danish Accreditation Fund accredits the Departments of Forensic Medicine as inspection bodies, according to an international European standard that is approved by Danish Standards. Mainly at police request, forensic doctors perform examinations of both victims and suspected perpetrators of perilous crimes. The examinations’ purposes are documentation and assessment of the findings and collection of biological evidence. The clinical forensic examinations do not include any treatment or medical follow-up. Thus, the forensic doctors must be neutral, objective and impartial. The clinical forensic examinations provide documentation of findings and conclusions not otherwise available for the police investigation and legal aftermath. Moreover, the accredited, standardised protocols ensure that the Departments of Forensic Medicine meet their obligations as inspection bodies, thus ensuring public confidence in the departments’ services.

Epidemiology of penetrating injury in an urban versus rural level 1 trauma center in the Netherlands

Background:

Penetrating injury can encompass a large spectrum of injuries dependent on the penetrating object, the location of entry, and the trajectory of the object through the human body. Therefore, the management of penetrating injuries can be challenging and often requires rapid assessment and intervention. No universal definition of penetrating injury exists in the literature and little is known about the demographics and outcome of penetrating injury in the Netherlands.

Objective:

A research was carried out to ascertain the size and outcome of penetrating injuries in two level-one trauma centers in the Netherlands.

Methods:

Using the trauma registry of the Radboud University Medical Center in Nijmegen and VU University Medical Center in Amsterdam, all patients with penetrating injury were identified who were admitted to these level 1 trauma centers in the period between January 1, 2009, and January 1, 2014. Penetrating injury was defined as an injury that caused disruption of the body surface and extended into the underlying tissue or into a body cavity. Data concerning age, gender, mechanism of injury, Glasgow Coma Scale, number of injuries, type of injury, and Injury Severity Score were collected and analyzed. Patient results were stratified by Injury Severity Score.

Results:

In total, 354 patients were identified, making up around 2% of all admitted trauma patients 3.1% (VU Medical Center) and 1.6% (Radboud Medical Center). Patients were overwhelmingly male (83.1%) and median age was 36 years (range = 1–88 years). Most injuries were caused by stabbings (51.1%) followed by shootings (26.3%). Admission to the intensive care unit occurred in 41.1% of all patients. Median stay in the intensive care unit was 5.1 days (range = 1–96 days) and median total hospital stay was 8 days (range = 1–95 days). Mortality among these patients was 7.1%, ranging from 0% among patients with Injury Severity Score 1–8 to 100% in patients with Injury Severity Score > 34. High mortality figures were associated with injuries caused by firearms (19.4%), injuries to the head (27.9%), and alleged assaults (10.9%). Differences in demographics between the two centers were not significant.

Conclusion:

Penetrating injury is a relative rare occurrence in the Netherlands compared with other countries. It is associated with high mortality and substantial hospital costs. The incidence of penetrating injuries is higher in metropolitan areas than in rural areas. A universal definition of penetrating trauma should be agreed upon in order to ensure that future studies remain free of bias, and also to ensure that data remain homogeneous.

Refining the Trauma and Injury Severity Score (TRISS) to Measure the Performance of the UK Combat Casualty Care System

INTRODUCTION

The Trauma and Injury Severity Score (TRISS) methodology is used in both the UK and US Military trauma registries. The method relies on dividing casualties according to mechanism, penetrating or blunt, and uses different weighting coefficients accordingly. The UK Military Joint Theatre Trauma Registry uses the original coefficients devised in 1987, whereas the US military registry uses updated civilian coefficients, but it is not clear how either registry analyzes explosive casualties according to the TRISS methodology. This study aims to use the UK Military Joint Theatre Trauma Registry to calculate new TRISS coefficients for contemporary battlefield casualties injured by either gunshot or explosive mechanisms. The secondary aim of this study is to apply the revised TRISS coefficients to examine the survival trends of UK casualties from recent military conflicts.

MATERIALS AND METHODS

The Joint Theatre Trauma Registry was searched for all UK casualties injured or killed in Iraq and Afghanistan by explosive or gunshot mechanisms between January 1, 2003 and December 31, 2014. Details of these casualties including injuries and vital signs were reviewed. Logistic regression analysis was performed to devise new TRISS coefficients; these were then used to examine survival over the 12 yr of the study.

RESULTS

Comparing the predictions from the gunshot TRISS model to the observed outcomes, it demonstrates a sensitivity of 98.1% and a specificity of 96.8% and an overall accuracy of 97.8%. With respect to the explosive TRISS model, there is a sensitivity of 98.6%, a specificity of 97.4%, and an overall accuracy of 98.4%. When this updated and mechanism-specific TRISS methodology was used to measure changes in survival over the study period, survival following these injuries improved until 2012 when performance was maintained for the last 2 yr of the study.

CONCLUSION

This study for the first time refines the TRISS methodology with coefficients appropriate for use within combat casualty care systems. This improved methodology reveals that UK combat casualty care performance appears to have improved until 2012 when this standard was maintained.

Protocol for a feasibility exploratory multicentre study of factors influencing trauma patients' outcomes of traffic crashes in Saudi Arabia

INTRODUCTION

Road traffic injury is a leading cause of death for people of all ages. The burden of road traffic injuries is well established in developed countries. However, there has been limited investigation of the incidence and burden of road traffic injury in low/middle-income countries. With a proportionally high number of road users, there is a need to explore the factors in prehospital and hospital care in Saudi Arabia (SA) that are associated with mortality for adult trauma patients following road traffic crashes (RTCs). This paper outlines the method for the planned research.

METHODS AND ANALYSIS

A feasibility exploratory multicentre study will be conducted at three purposefully selected hospitals with different trauma care resources in differing geographic locations of SA. The study sample will include all adult trauma patients who are involved in RTCs in SA and have been admitted to a study site in a 3-month period from May to July 2019. Data regarding the characteristics of the crashes and prehospital health care factors will be extracted from hospital databases where it is available. Information will be collected from patients or carers and hospital records in the two sites that do not have a registry. Patient status at 30 days post-injury, particularly mortality, will be assessed through hospital records. The relative contribution of a range of factors to predicting mortality will be explored using logistic regression analysis.

ETHICS AND DISSEMINATION

This study was approved by the Institutional Review Board Committee at King Saud Medical City (H-01-R-053), the General Department of Research and Studies at the Ministry of Health in SA (1440-1249939) and (1440-1398648), and the La Trobe University Human Research Ethics Committee (HEC19095). The results will be reported in a thesis and in peer-reviewed journal articles and conference presentations.

Validation of the Norwegian survival prediction model in trauma (NORMIT) in Swedish trauma populations

BACKGROUND

Trauma survival prediction models can be used for quality assessment in trauma populations. The Norwegian survival prediction model in trauma (NORMIT) has been updated recently and validated internally (NORMIT 2). The aim of this observational study was to compare the accuracy of NORMIT 1 and 2 in two Swedish trauma populations.

METHODS

Adult patients registered in the national trauma registry during 2014-2016 were eligible for inclusion. The study populations comprised the total national trauma (NT) population, and a subpopulation of patients admitted to a single level I trauma centre (TC). The primary outcome was 30-day mortality. Model validation included receiver operating characteristic (ROC) curve analysis and GiViTI calibration belts. The calibration was also assessed in subgroups of severely injured patients (New Injury Severity Score (NISS) over 15).

RESULTS

A total of 26 504 patients were included. Some 18·7 per cent of patients in the NT population and 2·6 per cent in the TC subpopulation were excluded owing to missing data, leaving 21 554 and 3972 respectively for analysis. NORMIT 1 and 2 showed excellent ability to distinguish between survivors and non-survivors in both populations, but poor agreement between predicted and observed outcome in the NT population with overestimation of survival, including in the subgroup with NISS over 15. In the TC subpopulation, NORMIT 1 underestimated survival irrespective of injury severity, but NORMIT 2 showed good calibration both in the total subpopulation and the subgroup with NISS over 15.

CONCLUSION

NORMIT 2 is well suited to predict survival in a Swedish trauma centre population, irrespective of injury severity. Both NORMIT 1 and 2 performed poorly in a more heterogeneous national population of injured patients.

Improvement of the performance of survival prediction in the ageing blunt trauma population: A cohort study

Introduction

The overestimation of survival predictions in the ageing trauma population results in negative benchmark numbers in hospitals that mainly treat elderly patients. The aim of this study was to develop and validate a modified Trauma and Injury Severity Score (TRISS) for accurate survival prediction in the ageing blunt trauma population.

Methods

This retrospective study was conducted with data from two Dutch Trauma regions. Missing values were imputed. New prediction models were created in the development set, including age (continuous or categorical) and Anesthesiologists Physical Status (ASA). The models were externally validated. Subsets were created based on age (≥75 years) and the presence of hip fracture. Model performance was assessed by proportion explained variance (Nagelkerke R²), discrimination (Area Under the curve of the Receiver Operating Characteristic, AUROC) and visually with calibration plots. A final model was created based on both datasets.

Results

No differences were found between the baseline characteristics of the development dataset (n = 15,530) and the validation set (n = 15,504). The inclusion of ASA in the prediction models showed significant improved discriminative abilities in the two subsets (e.g. AUROC of 0.52 [95% CI: 0.46, 0.58] vs. 0.74 [95% CI: 0.69, 0.78] for elderly patients with hip fracture) and an increase in the proportion explained variance (R² = 0.32 to R² = 0.35 in the total cohort). The final model showed high agreement between observed and predicted survival in the calibration plot, also in the subsets.

Conclusions

Including ASA and age (continuous) in survival prediction is a simple adjustment of the TRISS methodology to improve survival predictions in the ageing blunt trauma population. A new model is presented, through which even patients with isolated hip fractures could be included in the evaluation of trauma care.

Trauma and Injury Severity Score modification for predicting survival of trauma in one regional emergency medical center in Korea: Construction of Trauma and Injury Severity Score coefficient model

Background:

The problem that is central to trauma research is the prediction of survival rate after trauma. Trauma and Injury Severity Score is being used for predicting survival rate after trauma. Many countries have conducted a study on the classification, characteristics of variables, and the validity of the Trauma and Injury Severity Score model. However, few investigations have been made on the characteristics of coefficients or variables related to Trauma and Injury Severity Score in Korea. Objectives: There is a need for coefficient analysis of Trauma and Injury Severity Score which was created based on the United States database to be optimized for the situation in Korea.

Methods:

This study examined how the currently used Trauma and Injury Severity Score coefficients were developed and created for trauma patients visiting the emergency department in a hospital in Korea using the analytical method. A total of 34,340 trauma patients who were hospitalized into an emergency center from January 2012 to December 2014 for 3 years were analyzed with trauma registry established on August 2006.

Results:

Trauma and Injury Severity Score coefficients were transformed with the methods that were used to make the existing Trauma and Injury Severity Score coefficients using the trauma patients’ data. Regression coefficients (B) were drawn by building up a logistic regression analysis model that used variables such as Injury Severity Score, Revised Trauma Score, and age depending on survival with Trauma and Injury Severity Score.

Conclusion:

With regard to Trauma and Injury Severity Score established in the United States differing from Korea in injury types, it seems possible to realize significant survival rate by deriving coefficients with data in Korea and reanalyzing them.

Performance of the modified TRISS for evaluating trauma care in subpopulations: A cohort study

INTRODUCTION

Previous research showed that there is no agreement on a practically applicable model to use in the evaluation of trauma care. A modification of the Trauma and Injury Severity Score (modified TRISS) is used to evaluate trauma care in the Netherlands. The aim of this study was to evaluate the prognostic ability of the modified TRISS and to determine where this model needs improvement for better survival predictions.

METHODS

Patients were included if they were registered in the Brabant Trauma Registry from 2010 through 2015. Missing values were imputed according to multiple imputation. Subsets were created based on age, length of stay, type of injury and injury severity. Probability of survival was calculated with the modified TRISS. Discrimination was assessed with the Area Under the Receiver Operating Curve (AUROC). Calibration was studied graphically.

RESULTS

The AUROC was 0.84 (95% CI: 0.83, 0.85) for the total cohort (N = 69 747) but only 0.53 (95% CI: 0.51, 0.56) for elderly patients with hip fracture. Overall, calibration of the modified TRISS was adequate for the total cohort, with an overestimation for elderly patients and an underestimation for patients without brain injury.

CONCLUSIONS

Outcome comparison conducted with TRISS-based predictions should be interpreted with care. If possible, future research should develop a simple prediction model that has accurate survival prediction in the aging overall trauma population (preferable with patients with hip fracture), with readily available predictors.

Performance evaluation and validation of the animal trauma triage score and modified Glasgow Coma Scale with suggested category adjustment in dogs: A VetCOT registry study

OBJECTIVE

To examine the animal trauma triage (ATT) and modified Glasgow Coma Scale (mGCS) scores as predictors of mortality outcome (death or euthanasia) in injured dogs.

DESIGN

Observational cohort study conducted from September 2013 to March 2015 with follow-up until death or hospital discharge.

SETTING

Nine veterinary hospitals including private referral and veterinary teaching hospitals.

ANIMALS

Consecutive sample of 3,599 dogs with complete data entries recruited into the Veterinary Committee on Trauma patient registry.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We compared the predictive power (area under receiver operating characteristic [AUROC]) and calibration of the ATT and mGCS scores to their components. Overall mortality risk was 7.3% (n = 264). Incidence of head trauma was 9.5% (n = 341). The ATT score showed a linear relationship with mortality risk. Discriminatory performance of the ATT score was excellent with AUROC = 0.92 (95% confidence interval [CI] 0.91 to 0.94) and pseudo R² = 0.42. Each ATT score increase of 1 point was associated with an increase in mortality odds of 2.07 (95% CI = 1.94-2.21, P < 0.001). The "eye/muscle/integument" category of the ATT showed poor discrimination (AUROC = 0.55). When this component together with the skeletal and cardiac components were omitted from calculation of the overall score, there was no loss in discriminatory capacity (AUROC = 0.92 vs 0.91, P = 0.09) compared with the full score. The mGCS showed good performance overall, but performance improved when restricted to head trauma patients (AUROC = 0.84, 95% CI = 0.79-0.90, n = 341 vs 0.82, 95% CI = 0.79-0.85, n = 3599). The motor component of the mGCS showed the best predictive performance (AUROC = 0.79 vs 0.66/0.69); however, the full score performed better than the motor component alone (P = 0.002). When assessment was restricted to patients with head injury (n = 341), the ATT score still performed better than the mGCS (AUROC = 0.90 vs 0.84, P = 0.04).

CONCLUSIONS

In external validation on a large, multicenter dataset, the ATT score showed excellent discrimination and calibration; however, a more parsimonious score calculated on only the perfusion, respiratory, and neurological categories showed equivalent performance.

Bayesian averaging over decision tree models: An application for estimating uncertainty in trauma severity scoring

INTRODUCTION

For making reliable decisions, practitioners need to estimate uncertainties that exist in data and decision models. In this paper we analyse uncertainties of predicting survival probability for patients in trauma care. The existing prediction methodology employs logistic regression modelling of Trauma and Injury Severity Score (TRISS), which is based on theoretical assumptions. These assumptions limit the capability of TRISS methodology to provide accurate and reliable predictions.

METHODS

We adopt the methodology of Bayesian model averaging and show how this methodology can be applied to decision trees in order to provide practitioners with new insights into the uncertainty. The proposed method has been validated on a large set of 447,176 cases registered in the US National Trauma Data Bank in terms of discrimination ability evaluated with receiver operating characteristic (ROC) and precision-recall (PRC) curves.

RESULTS

Areas under curves were improved for ROC from 0.951 to 0.956 (p = 3.89 × 10^-18) and for PRC from 0.564 to 0.605 (p = 3.89 × 10^-18). The new model has significantly better calibration in terms of the Hosmer-Lemeshow Hˆ statistic, showing an improvement from 223.14 (the standard method) to 11.59 (p = 2.31 × 10^-18).

CONCLUSION

The proposed Bayesian method is capable of improving the accuracy and reliability of survival prediction. The new method has been made available for evaluation purposes as a web application.

Comparison of two prognostic models in trauma outcome

BACKGROUND

The Trauma Audit and Research Network (TARN) in the UK publicly reports hospital performance in the management of trauma. The TARN risk adjustment model uses a fractional polynomial transformation of the Injury Severity Score (ISS) as the measure of anatomical injury severity. The Trauma Mortality Prediction Model (TMPM) is an alternative to ISS; this study compared the anatomical injury components of the TARN model with the TMPM.

METHODS

Data from the National Trauma Data Bank for 2011-2015 were analysed. Probability of death was estimated for the TARN fractional polynomial transformation of ISS and compared with the TMPM. The coefficients for each model were estimated using 80 per cent of the data set, selected randomly. The remaining 20 per cent of the data were used for model validation. TMPM and TARN were compared using calibration curves, measures of discrimination (area under receiver operating characteristic curves; AUROC), proximity to the true model (Akaike information criterion; AIC) and goodness of model fit (Hosmer-Lemeshow test).

RESULTS

Some 438 058 patient records were analysed. TMPM demonstrated preferable AUROC (0·882 for TMPM versus 0·845 for TARN), AIC (18 204 versus 21 163) and better fit to the data (32·4 versus 153·0) compared with TARN.

CONCLUSION

TMPM had greater discrimination, proximity to the true model and goodness-of-fit than the anatomical injury component of TARN. TMPM should be considered for the injury severity measure for the comparative assessment of trauma centres.

Validating performance of TRISS, TARN and NORMIT survival prediction models in a Norwegian trauma population

Introduction

Anatomic injury, physiological derangement, age, injury mechanism and pre‐injury comorbidity are well‐founded predictors of trauma outcome. Statistical prediction models may have poorer discrimination, calibration and accuracy when applied in new locations. We aimed to compare the TRISS, TARN and NORMIT survival prediction models in a Norwegian trauma population.

Methods

Consecutive patients admitted to Oslo University Hospital Ullevål within 24 h after injury, with Injury Severity Score ≥ 10, proximal penetrating injuries, or received by trauma team, were studied. Original NORMIT coefficients were updated in a derivation dataset (NORMIT 2; n = 5923; 2005–2009). TRISS, TARN and NORMIT prediction models were evaluated in the validation dataset (n = 6348; 2010–2013) using two different AIS editions for injury coding. Exclusion due to missing data was 0.26%. Outcome was 30‐day mortality. Validation included AUROC, scaled Brier statistics, and calibration plots.

Results

The NORMIT models had significantly better discrimination, calibration, and overall fit than the TRISS 09, TARN 09 and TARN 12 models. The updated NORMIT 2 had higher numerical values of AUROC and scaled Brier than the original NORMIT, but with overlapping 95%CI. Overlapping 95%CI for AUROCs and Discrimination slopes indicated that the TARN and TRISS models performed similarly. Calibration plots showed tight and consistent predictions over all Ps strata for NORMIT 2 run on AIS'98 coded data, and only little deterioration when AIS'08 data was substituted.

Conclusions

In a Norwegian trauma population, the updated Norwegian survival prediction model in trauma (NORMIT 2) performed better than well‐established British and US alternatives. External validation of these three models in other Nordic populations is warranted.

Retrospective evaluation of 140 dogs involved in road traffic accidents

This study has retrospectively reviewed the medical records of 140 dogs sustaining road traffic accident (RTA), and has examined the population characteristics, medical history, injury type, physical examination, emergency laboratory tests and radiography findings, the animal trauma triage (ATT) score, the length of hospitalisation, the complications and the outcome. The survival rate was 83.2 per cent. Younger dogs sustained more frequently lung contusions and limb fractures, while larger dogs more frequently suffered limb fractures, and smaller dogs and older ones sustained more frequently pelvic fractures and sacroiliac luxation (P<0.05 for all). Dogs sustaining orthopaedic injuries required longer hospitalisation (P<0.001). The survival rates of non-ambulatory dogs (P<0.001) and those with neurological abnormalities (P<0.001), abnormal body temperature (P=0.001), hyperglycaemia (P=0.026) or hypoproteinaemia (P=0.04) at presentation were lower compared with those in which these were absent. The number of injured body systems was significantly (P<0.001) and positively associated with death. Dogs surviving RTA to presentation to the hospital have a good prognosis for survival to discharge. Older age, and high ATT score, abnormal body temperature, neurological deficits, hyperglycaemia and hypoproteinaemia at presentation, and occurrence of multiorgan trauma are negative prognostic indicators in such dogs.

A comparison of base deficit and vital signs in the early assessment of patients with penetrating trauma in a high burden setting

INTRODUCTION

An assessment of physiological status is a key step in the early assessment of trauma patients with implications for triage, investigation and management. This has traditionally been done using vital signs. Previous work from large European trauma datasets has suggested that base deficit (BD) predicts clinically important outcomes better than vital signs (VS). A BD derived classification of haemorrhagic shock appeared superior to one based on VS derived from ATLS criteria in a population of predominantly blunt trauma patients. The initial aim of this study was to see if this observation would be reproduced in penetrating trauma patients. The power of each individual variable (BD, heart rate (HR), systolic blood pressure (SBP), shock index(SI) (HR/SBP) and Glasgow Coma Score (GCS)) to predict mortality was then also compared.

METHODS

A retrospective analysis of adult trauma patients presenting to the Pietermaritzburg Metropolitan Trauma Service was performed. Patients were classified into four "shock" groups using VS or BD and the outcomes compared. Receiver Operator Characteristic (ROC) curves were then generated to compare the predictive power for mortality of each individual variable.

RESULTS

1863 patients were identified. The overall mortality rate was 2.1%. When classified by BD, HR rose and SBP fell as the "shock class" increased but not to the degree suggested by the ATLS classification. The BD classification of haemorrhagic shock appeared to predict mortality better than that based on the ATLS criteria. Mortality increased from 0.2% (Class 1) to 19.7% (Class 4) based on the 4 level BD classification. Mortality increased from 0.3% (Class 1) to 12.6% (Class 4) when classified based by VS. Area under the receiver operator characteristic (AUROC) curve analysis of the individual variables demonstrated that BD predicted mortality significantly better than HR, GCS, SBP and SI. AUROC curve (95% Confidence Interval (CI)) for BD was 0.90 (0.85-0.95) compared to HR 0.67(0.56-0.77), GCS 0.70(0.62-0.79), SBP 0.75(0.65-0.85) and SI 0.77(0.68-0.86).

CONCLUSION

BD appears superior to vital signs in the immediate physiological assessment of penetrating trauma patients. The use of BD to assess physiological status may help refine their early triage, investigation and management.

The effect of CT scanners in the trauma room - an observational study

BACKGROUND

A CT scanner incorporated in the trauma resuscitation bay may benefit trauma patients by fastening work-up times; however, evidence in the area is still sparse. We assessed if time from admission to first CT scan was lower after incorporation of a CT scanner in the resuscitation bay.

METHODS

We included trauma patients admitted in two 1-year periods, before and after a major rebuilding of the trauma room. Beforehand, one CT scanner was located in an adjacent room. After the rebuilding, two mobile CT scanners were placed in the resuscitation bays, where a moving gantry was combined with a trauma resuscitation table. Subgroup analyses were performed on severely injured and patients with traumatic brain injury.

RESULTS

We included 784 patients before and 742 patients after the reconstruction. Case-mix differed between study periods as there was a higher proportion of severe injuries, traumatic brain injury and penetrating trauma in the after period. We found a minor increase in time to CT in the after period (20 vs. 21 min, P = 0.008). In a multivariate regression analysis adjusted for differences in case-mix and with time to CT as outcome, period was an insignificant explanatory variable [β (before vs. after): 0.96 min 95% CI: 0.9-1.02, P = 0.3]. In both subgroups, we found no significant difference in time to CT.

CONCLUSION

We found no reduction in time to CT scan, when comparing a period with mobile CT scanners incorporated in the resuscitation bay to an earlier period with a CT scanner next to the trauma room.

Are prehospital deaths from trauma and accidental injury preventable? A direct historical comparison to assess what has changed in two decades

BACKGROUND & OBJECTIVES

In 1994, Hussain and Redmond revealed that up to 39% of prehospital deaths from accidental injury might have been preventable had basic first aid care been given. Since then there have been significant advances in trauma systems and care. The exclusion of prehospital deaths from the analysis of trauma registries, giv en the high rate of those, is a major limitation in prehospital research on preventable death. We have repeated the 1994 study to identify any changes over the years and potential developments to improve patient outcomes.

METHODS

We examined the full Coroner's inquest files for prehospital deaths from trauma and accidental injury over a three-year period in Cheshire. Injuries were scored using the Abbreviated-Injury-Scale (AIS-1990) and Injury Severity Score (ISS), and probability of survival estimated using Bull's probits to match the original protocol.

RESULTS

One hundred and thirty-four deaths met our inclusion criteria; 79% were male, average age at death was 53.6 years. Sixty-two were found dead (FD), fifty-eight died at scene (DAS) and fourteen were dead on arrival at hospital (DOA). The predominant mechanism of injury was fall (39%). The median ISS was 29 with 58 deaths (43%) having probability of survival of >50%. Post-mortem evidence of head injury was present in 102 (76%) deaths. A bystander was on scene or present immediately after injury in 45% of cases and prior to the Emergency Medical Services (EMS) in 96%. In 93% of cases a bystander made the call for assistance, in those DAS or DOA, bystander intervention of any kind was 43%.

CONCLUSIONS

The number of potentially preventable prehospital deaths remains high and unchanged. First aid intervention of any kind is infrequent. There is a potentially missed window of opportunity for bystander intervention prior to the arrival of the ambulance service, with simple first-aid manoeuvres to open the airway, preventing hypoxic brain injury and cardiac arrest.

Prehospital deaths from trauma: Are injuries survivable and do bystanders help?

BACKGROUND AND OBJECTIVES

Deaths from trauma occurring in the prehospital phase of care are typically excluded from analysis in trauma registries. A direct historical comparison with Hussain and Redmond's study on preventable prehospital trauma deaths has shown that, two decades on, the number of potentially preventable deaths remains high. Using updated methodology, we aimed to determine the current nature, injury severity and survivability of traumatic prehospital deaths and to ascertain the presence of bystanders and their role following the point of injury including the frequency of first-aid delivery.

METHODS

We examined the Coroners' inquest files for deaths from trauma, occurring in the prehospital phase, over a three-year period in the Cheshire and Manchester (City), subsequently referred to as Manchester, Coronial jurisdictions. Injuries were scored using the Abbreviated-Injury-Scale (AIS-2008), Injury Severity Score (ISS) calculated and probability of survival estimated using the Trauma Audit and Research Network's outcome prediction model.

RESULTS

One hundred and seventy-eight deaths were included in the study (one hundred and thirty-four Cheshire, forty-four Manchester). The World Health Organisation's recommendations consider those with a probability of survival between 25-50% as potentially preventable and those above 50% as preventable. The median ISS was 29 (Cheshire) and 27.5 (Manchester) with sixty-two (46%) and twenty-six (59%) respectively having a probability of survival in the potentially preventable and preventable ranges. Bystander presence during or immediately after the point of injury was 45% (Cheshire) and 39% (Manchester). Bystander intervention of any kind was 25% and 30% respectively. Excluding those found dead and those with a probability of survival less than 25%, bystanders were present immediately after the point of injury or "within minutes" in thirty-three of thirty-five (94%) Cheshire and ten of twelve (83%) Manchester. First aid of any form was attempted in fourteen of thirty-five (40%) and nine of twelve (75%) respectively.

CONCLUSIONS

A high number of prehospital deaths from trauma occur with injuries that are potentially survivable, yet first aid intervention is infrequent. Following injury there is a potential window of opportunity for the provision of bystander assistance, particularly in the context of head injury, for simple first-aid manoeuvres to save lives.

Systematic review and need assessment of pediatric trauma outcome benchmarking tools for low-resource settings

INTRODUCTION

Trauma is a leading cause of mortality and disability in children worldwide. The World Health Organization reports that 95% of all childhood injury deaths occur in Low-Middle-Income Countries (LMIC). Injury scores have been developed to facilitate risk stratification, clinical decision making, and research. Trauma registries in LMIC depend on adapted trauma scores that do not rely on investigations that require unavailable material or human resources. We sought to review and assess the existing trauma scores used in pediatric patients. Our objective is to determine their wideness of use, validity, setting of use, outcome measures, and criticisms. We believe that there is a need for an adapted trauma score developed specifically for pediatric patients in low-resource settings.

MATERIALS AND METHODS

A systematic review of the literature was conducted to identify and compare existing injury scores used in pediatric patients. We constructed a search strategy in collaboration with a senior hospital librarian. Multiple databases were searched, including Embase, Medline, and the Cochrane Central Register of Controlled Trials. Articles were selected based on predefined inclusion criteria by two reviewers and underwent qualitative analysis.

RESULTS

The scores identified are suboptimal for use in pediatric patients in low-resource settings due to various factors, including reliance on precise anatomic diagnosis, physiologic parameters maladapted to pediatric patients, or laboratory data with inconsistent accessibility in LMIC.

CONCLUSION

An important gap exists in our ability to simply and reliably estimate injury severity in pediatric patients and predict their associated probability of outcomes in settings, where resources are limited. An ideal score should be easy to calculate using point-of-care data that are readily available in LMIC, and can be easily adapted to the specific physiologic variations of different age groups.

Mortality prediction models in the general trauma population: A systematic review

BACKGROUND

Trauma is the leading cause of death in individuals younger than 40 years. There are many different models for predicting patient outcome following trauma. To our knowledge, no comprehensive review has been performed on prognostic models for the general trauma population. Therefore, this review aimed to describe (1) existing mortality prediction models for the general trauma population, (2) the methodological quality and (3) which variables are most relevant for the model prediction of mortality in the general trauma population.

METHODS

An online search was conducted in June 2015 using Embase, Medline, Web of Science, Cinahl, Cochrane, Google Scholar and PubMed. Relevant English peer-reviewed articles that developed, validated or updated mortality prediction models in a general trauma population were included.

RESULTS

A total of 90 articles were included. The cohort sizes ranged from 100 to 1,115,389 patients, with overall mortality rates that ranged from 0.6% to 35%. The Trauma and Injury Severity Score (TRISS) was the most commonly used model. A total of 258 models were described in the articles, of which only 103 models (40%) were externally validated. Cases with missing values were often excluded and discrimination of the different prediction models ranged widely (AUROC between 0.59 and 0.98). The predictors were often included as dichotomized or categorical variables, while continuous variables showed better performance.

CONCLUSION

Researchers are still searching for a better mortality prediction model in the general trauma population. Models should 1) be developed and/or validated using an adequate sample size with sufficient events per predictor variable, 2) use multiple imputation models to address missing values, 3) use the continuous variant of the predictor if available and 4) incorporate all different types of readily available predictors (i.e., physiological variables, anatomical variables, injury cause/mechanism, and demographic variables). Furthermore, while mortality rates are decreasing, it is important to develop models that predict physical, cognitive status, or quality of life to measure quality of care.

The use of whole-body computed tomography in major trauma: variations in practice in UK trauma hospitals

INTRODUCTION

Whole-body CT (WBCT) use in patients with trauma in England and Wales is not well documented. WBCT in trauma can reduce time to definitive care, thereby increasing survival. However, its use varies significantly worldwide.

METHODS

We performed a retrospective observational study of Trauma Audit and Research Network (TARN) data from 2012 to 2014. The proportion of adult patients receiving WBCT during initial resuscitation at major trauma centres (MTCs) and trauma units/non-designated hospitals (TUs/NDHs) was compared. A model was developed that included factors associated with WBCT use, and centre effects within the model were explored to determine variation in usage beyond that expected from the model.

RESULTS

Of the 115 664 study participants, 16.5% had WBCT. WBCT was performed five times more frequently in MTCs than in TUs/NDHs (31% vs 6.6%). In the multivariate model, increased injury severity, low GCS, shock, comorbidities and triage category increased the chances of having a WBCT, but there was no consistent relation with age. High falls and motor vehicle collisions also increased WBCT usage. Adjusting for casemix, there was a 13-fold intrahospital variation in the use of WBCT between MTCs and a 30-fold variation between TUs/NDHs. The amount of variability between individual hospitals that could not be accounted for by the factors shown to impact on WBCT use was 26% (95% CI 17% to 39%) for MTCs and 17% (95% CI 13% to 21%) for TUs/NDHs.

CONCLUSION

There are significant variations in WBCT use between different hospitals in England and Wales, which require further investigation.

The differential mortality of Glasgow Coma Score in patients with and without head injury

IMPORTANCE

The GCS was created forty years ago as a measure of impaired consciousness following head injury and thus the association of GCS with mortality in patients with traumatic brain injury (TBI) is expected. The association of GCS with mortality in patients without TBI (non-TBI) has been assumed to be similar. However, if this assumption is incorrect mortality prediction models incorporating GCS as a predictor will need to be revised.

OBJECTIVE

To determine if the association of GCS with mortality is influenced by the presence of TBI.

DESIGN/SETTING/PARTICIPANTS

Using the National Trauma Data Bank (2012; N=639,549) we categorized patients as isolated TBI (12.8%), isolated non-TBI (33%), both (4.8%), or neither (49.4%) based on the presence of AIS codes of severity 3 or greater. We compared the ability GCS to discriminate survivors from non-survivors in TBI and in non-TBI patients using logistic models. We also estimated the odds ratios of death for TBI and non-TBI patients at each value of GCS using linear combinations of coefficients.

MAIN OUTCOME MEASURE

Death during hospital admission.

RESULTS

As the sole predictor in a logistic model GCS discriminated survivors from non-survivors at an acceptable level (c-statistic=0.76), but discriminated better in the case of TBI patients (c-statistic=0.81) than non-TBI patients (c-statistic=0.70). In both unadjusted and covariate adjusted models TBI patients were about twice as likely to die as non-TBI patients with the same GCS for GCS values<8; for GCS values>8 TBI and non-TBI patients were at similar risk of dying.

CONCLUSIONS

A depressed GCS predicts death better in TBI patients than non-TBI patients, likely because in non-TBI patients a depressed GCS may simply be the result of entirely reversible intoxication by alcohol or drugs; in TBI patients, by contrast, a depressed GCS is more ominous because it is likely due to a head injury with its attendant threat to survival. Accounting for this observation into trauma mortality datasets and models may improve the accuracy of outcome prediction.

Survival prediction algorithms miss significant opportunities for improvement if used for case selection in trauma quality improvement programs

BACKGROUND

Quality improvement (QI) programs have shown to reduce preventable mortality in trauma care. Detailed review of all trauma deaths is a time and resource consuming process and calculated probability of survival (Ps) has been proposed as audit filter. Review is limited on deaths that were 'expected to survive'. However no Ps-based algorithm has been validated and no study has examined elements of preventability associated with deaths classified as 'expected'. The objective of this study was to examine whether trauma performance review can be streamlined using existing mortality prediction tools without missing important areas for improvement.

METHODS

We conducted a retrospective study of all trauma deaths reviewed by our trauma QI program. Deaths were classified into non-preventable, possibly preventable, probably preventable or preventable. Opportunities for improvement (OPIs) involve failure in the process of care and were classified into clinical and system deviations from standards of care. TRISS and PS were used for calculation of probability of survival. Peer-review charts were reviewed by a single investigator.

RESULTS

Over 8 years, 626 patients were included. One third showed elements of preventability and 4% were preventable. Preventability occurred across the entire range of the calculated Ps band. Limiting review to unexpected deaths would have missed over 50% of all preventability issues and a third of preventable deaths. 37% of patients showed opportunities for improvement (OPIs). Neither TRISS nor PS allowed for reliable identification of OPIs and limiting peer-review to patients with unexpected deaths would have missed close to 60% of all issues in care.

CONCLUSIONS

TRISS and PS fail to identify a significant proportion of avoidable deaths and miss important opportunities for process and system improvement. Based on this, all trauma deaths should be subjected to expert panel review in order to aim at a maximal output of performance improvement programs.

Top 10 TARN research publications

The last 25 years have seen Trauma Audit and Research Network's (TARN) research agenda develop into a significant portfolio of over 100 publications, including a number of international collaborations. Holding the largest trauma registry in Europe, TARN continues to provide researchers with the ability to pursue their interests in both epidemiological and clinical topics relating to traumatic injury. This edition of the Emergency Medicine Journal provides an opportunity to celebrate some of these papers with a ‘Top 10’, which have been voted by members of the TARN Research Committee on the basis of their impact.

The Head Injury Transportation Straight to Neurosurgery (HITS-NS) randomised trial: a feasibility study

BACKGROUND

Reconfiguration of trauma services, with direct transport of traumatic brain injury (TBI) patients to neuroscience centres (NCs), bypassing non-specialist acute hospitals (NSAHs), could potentially improve outcomes. However, delays in stabilisation of airway, breathing and circulation (ABC) and the difficulties in reliably identifying TBI at scene may make this practice deleterious compared with selective secondary transfer from nearest NSAH to NC. National Institute for Health and Care Excellence guidance and systematic reviews suggested equipoise and poor-quality evidence - with regard to 'early neurosurgery' in this cohort - which we sought to address.

METHODS

Pilot cluster randomised controlled trial of bypass to NC conducted in two ambulance services with the ambulance station (n = 74) as unit of cluster [Lancashire/Cumbria in the North West Ambulance Service (NWAS) and the North East Ambulance Service (NEAS)]. Adult patients with signs of isolated TBI [Glasgow Coma Scale (GCS) score of < 13 in NWAS, GCS score of < 14 in NEAS] and stable ABC, injured nearest to a NSAH were transported either to that hospital (control clusters) or bypassed to the nearest NC (intervention clusters).

PRIMARY OUTCOMES

recruitment rate, protocol compliance, selection bias as a result of non-compliance, accuracy of paramedic TBI identification (overtriage of study inclusion criteria) and pathway acceptability to patients, families and staff. 'Open-label' secondary outcomes: 30-day mortality, 6-month Extended Glasgow Outcome Scale (GOSE) and European Quality of Life-5 Dimensions.

RESULTS

Overall, 56 clusters recruited 293 (169 intervention, 124 control) patients in 12 months, demonstrating cluster randomised pre-hospital trials as viable for heath service evaluations. Overall compliance was 62%, but 90% was achieved in the control arm and when face-to-face paramedic training was possible. Non-compliance appeared to be driven by proximity of the nearest hospital and perceptions of injury severity and so occurred more frequently in the intervention arm, in which the perceived time to the NC was greater and severity of injury was lower. Fewer than 25% of recruited patients had TBI on computed tomography scan (n = 70), with 7% (n = 20) requiring neurosurgery (craniotomy, craniectomy or intracranial pressure monitoring) but a further 18 requiring admission to an intensive care unit. An intention-to-treat analysis revealed the two trial arms to be equivalent in terms of age, GCS and severity of injury. No significant 30-day mortality differences were found (8.8% vs. 9.1/%; p > 0.05) in the 273 (159/113) patients with data available. There were no apparent differences in staff and patient preferences for either pathway, with satisfaction high with both. Very low responses to invitations to consent for follow-up in the large number of mild head injury-enrolled patients meant that only 20% of patients had 6-month outcomes. The trial-based economic evaluation could not focus on early neurosurgery because of these low numbers but instead investigated the comparative cost-effectiveness of bypass compared with selective secondary transfer for eligible patients at the scene of injury.

CONCLUSIONS

Current NHS England practice of bypassing patients with suspected TBI to neuroscience centres gives overtriage ratios of 13 : 1 for neurosurgery and 4 : 1 for TBI. This important finding makes studying the impact of bypass to facilitate early neurosurgery not plausible using this study design. Future research should explore an efficient comparative effectiveness design for evaluating 'early neurosurgery through bypass' and address the challenge of reliable TBI diagnosis at the scene of injury.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN68087745.

FUNDING

This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 20, No. 1. See the NIHR Journals Library website for further project information.

Comparison of Trauma and Injury Severity Score model with alternative approach in outcome prediction in trauma using National Trauma Database in Malaysia

Objective

The coefficients for the Trauma and Injury Severity Score are derived from the Major Trauma Outcome Study (MTOS) in North America, so the applicability of the MTOS-Trauma and Injury Severity Score (TRISS) in different populations remains challenging. This study proposed an alternate TRISS model called National Trauma Database-TRISS, where the coefficients were derived from the National Trauma Database (NTrD) in Malaysia.

Methods

This study utilised data derived from the National Trauma Database in Malaysia with 70% of the cases being used to develop the NTrD-TRISS model, while model validation was conducted based on the remaining 30% of cases. Fractional polynomial method was applied to correct the non-linearity in the logit of Injury Severity Score (ISS) and New Injury Severity Score (NISS). The predictive validity of the NTrD-TRISS model was compared with the MTOS-TRISS model. The predictive validity of these models was further examined with ISS and NISS.

Results

A total of 5857 major trauma cases reported to 13 trauma centres located throughout Malaysia for patients aged 16 and above were included in the study. The outcome prediction in trauma from NTrD-TRISS model was found to be more accurate compared to the MTOS-TRISS model.

Conclusions

This study has ascertained the applicability of the MTOS-TRISS model in Malaysia. The outcome prediction in trauma from the NTrD-TRISS was found to outperform the MTOS-TRISS model. Given the complex computational nature of ISS, the MTOS-TRISS and NTrD-TRISS models with NISS were recommended for future practical usage.

Association between trauma and socioeconomic deprivation: a registry-based, Scotland-wide retrospective cohort study of 9,238 patients

Background

Trauma remains a leading cause of morbidity and mortality in the UK and throughout the world. Socioeconomic deprivation has been linked with many types of ill-health and previous studies have shown an association with injury in other parts of the world. The aim of this study was to investigate the association between socioeconomic deprivation and trauma incidence and case-fatality in Scotland.

Methods

The study included nine thousand two hundred and thirty eight patients attending Emergency Departments following trauma across Scotland in 2011-12. A retrospective cohort study was conducted using secondary data extracted from the national trauma registry. Postcode of residence was used to generate deciles using the Scottish Index of Multiple Deprivation. The incidence rate ratio (IRR) was calculated to allow comparison of incidence of trauma across SIMD deciles. For mortality, observed: expected ratios were obtained using observed mortality in the cohort and expected deaths using probability of survival based on Trauma and Injury Severity Score (TRISS) method.

Results

Compared with the most deprived decile, the least deprived had an incidence rate ratio (IRR) for all trauma of 0.43 (95 % CI 0.32–0.58, p < 0.001). The association was stronger for penetrating trauma (IRR 0.07, 95 % CI .01–0.56, p = 0.011). There was a significant interaction between age, gender and SIMD. For case fatality, multivariate logistic regression showed that, severity of trauma (ISS > 15) OR 18.11 (95 % CI 13.91 to 23.58) and type of injury (Penetrating versus blunt injury) OR 2.07 (95 % CI 1.15 to 3.72) remain as independent predictors of case fatality in this dataset.

Discussion

Our data shows a higher incidence of trauma amongst a socioeconomically deprived population, in keeping with other areas of the world. In our dataset, outcome, as measured by in-hospital mortality, does not appear to be associated with socioeconomic deprivation.

Conclusion

In Scotland, populations living in socioeconomically deprived areas have a higher incidence of trauma, especially penetrating trauma, requiring hospital attendance. Case fatality is associated with more severe trauma and penetrating trauma, but not socioeconomic deprivation.

Electronic supplementary material

The online version of this article (doi:10.1186/s13049-016-0275-7) contains supplementary material, which is available to authorized users.

Factors affecting mortality in older trauma patients-A systematic review and meta-analysis

INTRODUCTION

Major trauma in older people is a significant health burden in the developed world. The aging of the population has resulted in larger numbers of older patients suffering serious injury. Older trauma patients are at greater risk of death from major trauma, but the reasons for this are less well understood. The aim of this review was to identify the factors affecting mortality in older patients suffering major injury.

MATERIALS AND METHODS

A systematic review of Medline, Cinhal and the Cochrane database, supplemented by a manual search of relevant papers was undertaken, with meta-analysis. Multi-centre cohort studies of existing trauma registries that reported risk-adjusted mortality (adjusted odds ratios, AOR) in their outcomes and which analysed patients aged 65 and older as a separate cohort were included in the review.

RESULTS

3609 papers were identified from the electronic databases, and 28 from manual searches. Of these, 15 papers fulfilled the inclusion criteria. Demographic variables (age and gender), pre-existing conditions (comorbidities and medication), and injury-related factors (injury severity, pattern and mechanism) were found to affect mortality. The 'oldest old', aged 75 and older, had higher mortality rates than younger patients, aged 65-74 years. Older men had a significantly higher mortality rate than women (cumulative odds ratio 1.51, 95% CI 1.37-1.66). Three papers reported a higher risk of death in patients with pre-existing conditions. Two studies reported increased mortality in patients on warfarin (cumulative odds ratio 1.32, 95% CI 1.05-1.66). Higher mortality was seen in patients with lower Glasgow coma scores and systolic blood pressures. Mortality increased with increased injury severity and number of injuries sustained. Low level falls were associated with higher mortality than motor vehicle collisions (cumulative odds ratio 2.88, 95% CI 1.26-6.60).

CONCLUSIONS

Multiple factors contribute to mortality risk in older trauma patients. The relation between these factors and mortality is complex, and a fuller understanding of the contribution of each factor is needed to develop a better predictive model for trauma outcomes in older people. More research is required to identify patient and process factors affecting mortality in older patients.

Prediction modelling for trauma using comorbidity and 'true' 30-day outcome

BACKGROUND

Prediction models for trauma outcome routinely control for age but there is uncertainty about the need to control for comorbidity and whether the two interact. This paper describes recent revisions to the Trauma Audit and Research Network (TARN) risk adjustment model designed to take account of age and comorbidities. In addition linkage between TARN and the Office of National Statistics (ONS) database allows patient's outcome to be accurately identified up to 30 days after injury. Outcome at discharge within 30 days was previously used.

METHODS

Prospectively collected data between 2010 and 2013 from the TARN database were analysed. The data for modelling consisted of 129 786 hospital trauma admissions. Three models were compared using the area under the receiver operating curve (AuROC) for assessing the ability of the models to predict outcome, the Akaike information criteria to measure the quality between models and test for goodness-of-fit and calibration. Model 1 is the current TARN model, Model 2 is Model 1 augmented by a modified Charlson comorbidity index and Model 3 is Model 2 with ONS data on 30 day outcome.

RESULTS

The values of the AuROC curve for Model 1 were 0.896 (95% CI 0.893 to 0.899), for Model 2 were 0.904 (0.900 to 0.907) and for Model 3 0.897 (0.896 to 0.902). No significant interaction was found between age and comorbidity in Model 2 or in Model 3.

CONCLUSIONS

The new model includes comorbidity and this has improved outcome prediction. There was no interaction between age and comorbidity, suggesting that both independently increase vulnerability to mortality after injury.

New Injury Severity Score is a better predictor of mortality for blunt trauma patients than the Injury Severity Score

BACKGROUND

Trauma-related mortality depends on injury severity. Several trauma scores are used to evaluate injury severity. We compared the Injury Severity Score (ISS) and the New Injury Severity Score (NISS) in terms of predicting mortality among hospitalized blunt trauma patients.

METHODS

The data of Al-Ain Hospital Trauma Registry were prospectively collected over 3 years. Data of blunt trauma patients were then analyzed retrospectively. Univariate analysis was used to compare patients who died with those who survived. Sex, age, mechanism of injury, heart rate, systolic blood pressure (SBP), and Glasgow Coma Score (GSC) on arrival at the hospital, ISS, and NISS were studied. Significant factors were then entered into a direct likelihood ratio logistic regression model.

RESULTS

Of 2,573 patients in the registry, 2,115 (82.2 %) suffered blunt trauma at a mean (SD) age of 32 (15.3) years. Among them, 1,838 (87 %) were male. Main mechanisms of injury were road traffic collision (vehicle occupants) (32.8 %) and falling from a height (22.4 %). Fifty patients (2.4 %) died. Univariate analysis showed that GCS and SBP at hospital arrival, ISS, NISS, and mechanism of injury significantly affected mortality. Logistic regression model showed that mortality was significantly increased by low GCS (p < 0.0001), high NISS (p < 0.0001), and low SBP (p = 0.006) at hospital arrival.

CONCLUSIONS

Mortality of blunt trauma in the UAE is significantly affected by high NISS, low GCS, and hypotension. NISS is better than ISS for predicting mortality of blunt trauma patients and may replace it.