| BACKGROUND: The fundamental condition for ensuring the quality of treatment is monitoring the results. For an objective assessment, proper data is required. The main purpose of trauma registries is to gather data on the complete course from the accident site to the completion of therapy. Because of the dispersion of trauma patients across multiple hospitals in Slovenia, a national major trauma registry is being considered for implementation. The TRISS methodology is most commonly used to compare treatment results to international standards. It was developed using a multivariate analysis of a trauma patient group (MTOS), which significantly differs from the Slovenian pattern of major trauma incidents. Because of this, the usefulness of the TRISS methodology has come under doubt. After including Slovenian trauma patients in the German Traumaregistry (TR DGU), we have implemented the RISC method for comparing treatment results with other centres. The new method has not yet been evaluated on a patient group which differs from the TR DGU pattern. The objective of the study is to assess the usefulness of the RISC method on the Slovenian pattern of major trauma patients and to compare it to the TRISS method.;METHODS: Since the inclusion in the TR DGU in 2006, we have been prospectively gathering data on a cohort group of 376 major trauma patients at General Hospital Celje and sent it to the TR DGU. We have compared risk factors between trauma patients at General Hospital Celje and the TR DGU and calculated the TRISS and RISC score for every patient. The M-statistic was used to compare the distribution of probability of survival between General Hospital Celje, MTOS and TR DGU. TRISS, RISC and RISC II were evaluated using the statistical methods of discrimination (aROC), precision (difference in survival) and calibration (H-L Statistic).;RESULTS: The average age of patients was 47 years, 83% of them were men, 95% had blunt trauma. The average ISS was 26,4 (90% ≥16). Their in-hospital mortality rate was 17,5%. The standardized mortality rate has shown a 1,9% lower mortality than predicted with statistical models.;A significant discrepancy in risk factors between trauma patients at General Hospital Celje and the TR DGU was discovered, which can be attributed to different inclusion criteria. The Slovenian sample of major trauma patients was confirmed to be different, as was the need to evaluate the effectiveness of the RISC method.;We have compared survival rates between patient groups and found a poor match between the patient group at General Hospital Celje and the MTOS group (M=0,50) and a good match with TR DGU (M=0,88). The distribution at General Hospital Celje has not changed significantly between the periods of 2006--07 and 2011--12 (M=0,90). We have confirmed a significant difference between the General Hospital Celje and the MOTS patients groups, which affects the usefulness of the TRISS method. We have confirmed a good match of results between the General Hospital Celje and the TR DGU patient groups, despite differences in risk factors, which justifies the use of the RISC method on the Slovenian pattern of major trauma patients.;The RISC and RISCII scores have shown the best discrimination (aROC 0,91 and 0,90). The RISC score was the most precise (difference in mortality rate of 1,9%) and also had the best and almost perfect calibration (H-L 0,53).;CONCLUSIONS: The RISC method was shown to have better discrimination, calibration and precision than the TRISS methodology for the Slovenian pattern of major trauma patients. The comparison of data at General Hospital Celje between 2006--07 and 2011--12 has shown a few differences. The improvement is best seen with indicators of treatment quality (shorter prehospital time, shorter ER time). These differences have not affected the usefulness of the RISC method, however. It has stood the test of time, as the structure and treatment of patients have changed. The TRISS method, despite new coefficients, is not suitable for the Slovenian pattern of major trauma patients. The updated RISCII is a good method, although not any better than the established RISC method. |
