Harborview assessment for risk of mortality: An improved measure of injuryseverity on the basis of ICD-9-CM

Background: There have been several attempts to develop a scoring system th at can accurately reflect the severity of a trauma patient's injuries, part icularly with respect to the effect of the injury on survival, Current meth odologies require unreliable physiologic data for the assignment of a survi val probability and fail to account for the potential synergism of differen t injury combinations. The purpose of this study was to develop a scoring s ystem to better estimate probability of mortality on the basis of informati on that is readily available from the hospital discharge sheet and does not rely on physiologic data. Methods: Records from the trauma registry from an urban Level I trauma cent er were analyzed using logistic regression. Included in the regression were Internation Classification of Diseases-9th Rev (ICD-9-CM) codes for anatom ic injury, mechanism, intent, and preexisting medical conditions, as well a s age. Two-way interaction terms for several combinations of injuries were also included in the regression model. The resulting Harborview Assessment for Risk of Mortality (WARM) score was then applied to an independent test data set and compared with Trauma and Injury Severity Score (TRISS) probabi lity of survival and ICD-9-CM Injury Severity Score (ICISS) for ability to predict mortality using the area under the receiver operator characteristic curve, Results:The HARM score was based on analysis of 16,042 records (design set) , When applied to an independent validation set of 15,957 records, the area under the receiver operator characteristic curve (AUC) for WARM was 0.9592 , This represented significantly better discrimination than both TRISS prob ability of survival (AUC = 0.9473, p = 0.005) and ICISS (AUC = 0.9402, p = 0.001), HARM also had a better calibration (Hosmer-Lemeshow statistic [HL] = 19.74) than TRISS (HI, = 55.71) and ICISS (NL = 709.19), Physiologic data were incomplete for 6,124 records (38%) of the validation set; TRISS could not be calculated at all for these records. Conclusion: The HARM score is an effective tool for predicting probability of in-hospital mortality for trauma patients, It outperforms both the TRISS and ICD-9-CM Injury Severity Score (ICISS) methodologies with respect to b oth discrimination and calibration, using information that is readily avail able from hospital discharge coding, and without requiring emergency depart ment physiologic data.