Related Experiment Video
Updated: Jul 28, 2026
Inverse Probability of Treatment Weighting (Propensity Score) using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
Improving the predictive ability of the ICD-based Injury Severity Score
1Department Preventive and Social Medicine, Injury Prevention Research Unit, University of Otago, Dunedin, New Zealand. gabrielle.davie@ipru.otago.ac
Objective:
To assess whether the use of integrated hospitalization and mortality data sources and/or the inclusion of comorbidity improve the predictive ability of the International Classification of Disease (ICD)-based Injury Severity Score (ICISS).
Design:
Models using either the ICISS based solely on hospital discharge data or one of nine modified ICISSs as the predictor variable were assessed on their ability to predict survival using logistic regression modeling.
Setting:
New Zealand. Patients or
Subjects:
Inpatients, with an S00-T89 ICD-10-AM principal diagnosis, and fatalities, with any S00-T89 ICD-10-AM diagnosis, occurring in 2000-2003.
Interventions:
None.
Main Outcome Measures:
Models were compared in terms of their discrimination (concordance), calibration, and goodness-of-fit.
Results:
186 835 cases including 9968 deaths met the inclusion criterion. The modified ICISS that included both mortality data and Charlson comorbid conditions at the ICD-10-AM level had the best concordance and high calibration. Calibration curves indicated that scores using hospital discharge data only to calculate survival risk ratios underestimated mortality, whereas scores using hospital discharge and mortality data overestimated mortality.
Conclusions:
Valid measurement of injury severity is important for both meaningful research and surveillance and to assist in classifying information to meet specific injury policy, prevention, and control needs. This study suggests that the predictive ability of ICISS would be improved if both mortality and comorbidity data were included in its calculation.