Performance of ICD-based injury severity measures used to predict in-hospital mortality and intensive care admission among traumatic brain-injured patients.

M. Gagné,L. Moore,M. Sirois,M. Simard,C. Beaudoin,Brice-Lionel Kuimi

Journal of Trauma and Acute Care Surgery

Abstract:BACKGROUND The International Classification of Diseases (ICD) is the main classification system used for population-based traumatic brain injury (TBI) surveillance activities but does not contain direct information on injury severity. ICD-based injury severity measures can be empirically derived or mapped to the Abbreviated Injury Scale, but no single approach has been formally recommended for TBI. OBJECTIVE To compare the accuracy of different ICD-based injury severity measures for predicting in-hospital mortality and intensive care unit (ICU) admission in TBI patients. METHODS We conducted a population-based retrospective cohort study. We identified all patients aged 16 years and over with a TBI diagnosis who received acute care between April 1, 2006 and March 31, 2013 from the Quebec Hospital Discharge Database. The accuracy of five ICD-based injury severity measures for predicting mortality and ICU admission were compared using measures of discrimination (area under the receiver operating characteristic [AUC]) and calibration (calibration plot and the Hosmer-Lemeshow goodness-of-fit statistic). RESULTS Of 31,087 traumatic brain-injured patients in the study population, 9.0% died in hospital and 34.4% were admitted to the ICU. Among ICD-based severity measures that were assessed, the multiplied derivative of ICD-based injury severity score (ICISS-Multiplicative) demonstrated the best discriminative ability for predicting in-hospital mortality (AUC: 0.858, 95% confidence interval [CI] 0.852-0.864) and ICU admissions (AUC: 0.813, 95%CI 0.808-0.818). Calibration assessments showed good agreement between observed and predicted in-hospital mortality for ICISS measures. All severity measures presented high agreement between observed and expected probabilities of ICU admission for all deciles of risk. CONCLUSIONS ICD-based injury severity measures can be used to accurately predict in-hospital mortality and ICU admission in TBI patients. ICISS-Multiplicative generally outperformed other ICD-based injury severity measures and should be preferred to control for differences in baseline characteristics between TBI patients in surveillance activities or injury research when only ICD codes are available. LEVEL OF EVIDENCE III. STUDY TYPE Prognostic study.