Performance of new adjustments to the TRISS equation model in developed and developing countries

BACKGROUND: The Trauma and Injury Severity Score (TRISS) has been criticized for being based on data from the USA and Canada—high-income countries—and therefore, it may not be applicable to low-income and middle-income countries. The present study evaluated the accuracy of three adjustments to the TRISS equation model (NTRISS-like; TRISS SpO(2); NTRISS-like SpO(2)) in a high-income and a middle-income country to compare their performance when derived and applied to different groups. METHODS: This was a retrospective study of trauma patients admitted to two institutions: a university medical center in São Paulo, Brazil (a middle-income country), and a level 1 university trauma center in San Diego, USA (a high-income country). Patients were admitted between January 1, 2006, and December 31, 2010. The subjects were 2416 patients from Brazil and 8172 patients from the USA. All equations had adjusted coefficients for São Paulo and San Diego and for blunt and penetrating trauma. Receiver operating characteristic (ROC) curves were used to evaluate performance of the models. RESULTS: Regardless of the population where the equation was generated, it performed better when applied to patients in the USA (AUC from 0.911 to 0.982) compared to patients in Brazil (AUC from 0.840 to 0.852). When the severity was considered and homogenized, the performance of equations were similar to both application in the USA and Brazil. CONCLUSIONS: Survival probability models showed better performance when applied in data collected in the high-income countries (HIC) regardless the country they were derived. The severity is an important factor to consider when using non-adjusted survival probability models for the local population. Adjusted models for severely traumatized patients better predict survival probability in less severely traumatized populations. Other factors besides physiological and anatomical data may impact final outcomes and should be identified in each environment if they are to be used in the development of the trauma care performance improvement process in middle-income countries.