Outcomes of traumatic brain injury in Hong Kong: Validation with TRISS, CRASH and IMPACT models

Abstract

Objective: This study aimed to test prognostic models (TRISS, IMPACT and CRASH models) for 14-day mortality, 6-month mortality and 6-month unfavourable outcome in a cohort of trauma patients with TBI in Hong Kong. Methods: This study analysed 661 patients with significant TBI treated in a regional trauma centre in Hong Kong over a 3-year period. Discriminatory power of the models was assessed as the area under the receiver operating characteristics curve. One-sample t-tests were used to compare actual outcomes in the cohort against predicted outcomes. Results: All three prognostic models were shown to have good discriminatory power and no significant systemically over-estimation or under-estimation. For the 153 mild TBI patients, 14-day mortality was 12% (16/133) and 6-month unfavourable outcome was 18% (24/133). Fourteen-day mortality did not differ from the CRASH prediction (difference: 2%, 95% CI ¼ 5% to 1%, p ¼ 0.297), but the CRASH-predicted the 6-month unfavourable outcome rate was higher than the actual rate in this patient cohort (difference: 11%, 95% CI ¼ 7–15%, p50.001). Conclusions: All three predictive models are applicable to the respective eligible TBI patients in Hong Kong. These predictive models can be utilized to audit TBI management outcome for trauma service development in the future. The impact of healthcare delivery models for mild head injury should be further investigated.