Validation of a relative head injury severity scale for pediatric trauma.

BACKGROUND: Brain injury is the most important independent predictor of mortality and morbidity in pediatric trauma. The Glasgow Coma Score (GCS) is the commonly used clinical instrument to assess brain injury. However, the GCS or one of its components is often not applicable in children under a certain age or cannot be computed reliably because of the patient's condition or the circumstances surrounding resuscitation efforts. This limits its usefulness in statistical models of trauma outcomes, which rely on complete data collection and entry into trauma registries. This study provides evidence validating use of a relative head injury severity scale (RHISS) derived from available International Classification of Diseases, 9th Revision (ICD-9) diagnosis codes to stratify degree of head injury. METHODS: The patient population was derived from the National Pediatric Trauma Registry (NPTR;1994-2001). Survival Risk Ratios (SRRs) were computed for each head injury ICD-9 code. ICD-9 diagnosis codes related to head injury were then assigned to a RHISS category based on duration of loss of consciousness, location of skull fracture, or both: 0 = none; 1 = mild; 2 = moderate, or 3 = severe head injury. Analysis of variance compared mean SRRs across RHISS categories. Each patient was then assigned to a RHISS category based on their single worst ICD-9 head injury code. Logistic regression analysis was used to predict mortality based on New Injury Severity Score (NISS), whether the patient had been intubated, RHISS, and the Abbreviated Injury Score (AIS) for head and neck injuries. RESULTS: GCS score was missing for 96% of nonsurvivors in the NPTR. Mean SRRs differed significantly (p < 0.001) among ICD-9 codes assigned to each RHISS category, as follows (Mean +/- SD): RHISS (0) = 0.93 +/- 0.16; RHISS (1) = 0.89 +/- 0.22; RHISS (2) = 0.85 +/- 0.26; RHISS (3) = 0.55 +/- 0.35. Logistic regression identified RHISS as an independent significant predictor (p < 0.01) of mortality. CONCLUSION: RHISS is a valid index of degree of head injury in the pediatric trauma population. Unlike GCS, RHISS is more likely to be available in trauma registries, and can be computed from administrative data. RHISS provides a feasible and valid method for quantifying the degree of brain injury in statistical models of pediatric trauma outcome.

Traumatic brain injury in the elderly: increased mortality and worse functional outcome at discharge despite lower injury severity.

OBJECTIVE: The purpose of this study was to compare data obtained from a statewide data set for elderly patients (age > 64 years) that presented with traumatic brain injury with data from nonelderly patients (age > 15 and < 65 years) with similar injuries. METHODS: The New York State Trauma Registry from January 1994 through December 1995, from trauma centers and community hospitals excluding New York City (45,982 patients), was examined. Head-injured patients were identified by International Classification of Diseases, Ninth Revision diagnosis codes. A relative head injury severity scale (RHISS) was constructed on the basis of groups of these codes (range, 0 = none to 3 = severe). Comparisons were made with nonelderly patients for mortality, Glasgow Coma Scale (GCS) score at admission and discharge, Injury Severity Score, New Injury Severity Score, and RHISS. Outcome was assessed by a Functional Independence Measure score in three major domains: expression, locomotion, and feeding. Data were analyzed by the chi2 test and Mann-Whitney U test, with p < 0.05 considered significant. RESULTS: There were 11,772 patients with International Classification of Diseases, Ninth Revision diagnosis of head injury, of which 3,244 (27%) were elderly. There were more male subjects in the nonelderly population (78% male subjects) compared with the elderly population (50% men). Mortality was 24.0% in the elderly population compared with 12.8% in the nonelderly population (risk ratio, 2.2; 95% confidence interval, 1.99-2.43). The elderly nonsurvivors were statistically older, and mortality rate increased with age. Stratified by GCS score, there was a higher percentage of nonsurvivors in the elderly population, even in the group with only moderately depressed GCS score (GCS score of 13-15; risk ratio, 7.8; 95% confidence interval, 6.1-9.9 for elderly vs. nonelderly). Functional outcome in all three domains was significantly worse in the elderly survivors compared with the nonelderly survivors. CONCLUSION: Elderly traumatic brain injury patients have a worse mortality and functional outcome than nonelderly patients who present with head injury even though their head injury and overall injuries are seemingly less severe.

Development of a model for prediction of survival in pediatric trauma patients: comparison of artificial neural networks and logistic regression.

BACKGROUND/PURPOSE: There is a paucity of outcome prediction models for injured children. Using the National Pediatric Trauma Registry (NPTR), the authors developed an artificial neural network (ANN) to predict pediatric trauma death and compared it with logistic regression (LR). METHODS: Patients in the NPTR from 1996 through 1999 were included. Models were generated using LR and ANN. A data search engine was used to generate the ANN with the best fit for the data. Input variables included anatomic and physiologic characteristics. There was a single output variable: probability of death. Assessment of the models was for both discrimination (ROC area under the curve) and calibration (Lemeshow-Hosmer C-Statistic). RESULTS: There were 35,385 patients. The average age was 8.1 +/- 5.1 years, and there were 1,047 deaths (3.0%). Both modeling systems gave excellent discrimination (ROC A(z): LR = 0.964, ANN = 0.961). However, LR had only fair calibration, whereas the ANN model had excellent calibration (L/H C stat: LR = 36, ANN = 10.5). CONCLUSIONS: The authors were able to develop an ANN model for the prediction of pediatric trauma death, which yielded excellent discrimination and calibration exceeding that of logistic regression. This model can be used by trauma centers to benchmark their performance in treating the pediatric trauma population.

The effects of motorcycle helmet use between hospitals in states with and without a mandatory helmet law.

BACKGROUND: Motorcycle injuries and mortality are different depending on the use of a helmet. Helmet use varies greatly depending on state laws. METHODS: Retrospective study using trauma registry data from two Level 1 Trauma Centers in states with (NY) and without (CT) a mandatory helmet law, from 1996 through 1998. RESULTS: Motorcycle accident victims in both states were similar for sex, age, RTS, TRISS probability of survival, GCS on arrival and ISS. Helmet use was higher in New York than in Connecticut (91% vs 18%, P < .01). Mortality was higher in Connecticut than in New York (15% vs 6%, P < .05). CONCLUSION: The demographics and injury severity of motorcycle accident victims presenting to Level 1 Trama Centers were very similar in the two adjoining states. The most significant difference between the states is that of helmet use. This is closely related to the decreased mortality rate and the higher GCS at discharge seen in the state with the mandatory helmet law.