Using Trauma and Injury Severity Score (TRISS)-based analysis in the development of regional risk adjustment tools to trend quality in a voluntary trauma system: The experience of the Trauma Foundation of Northeast Ohio

Background: Presently, no trauma system exists in Ohio. Since 1993, all hos pitals in Cuyahoga County (CUY), northeast Ohio (n = 22) provide data to a trauma registry, In return, each received hospital-specific data, compariso n data by trauma care level and a county-wide aggregate summary. This repor t describes the results of this approach in our region. Methods: All cases were entered by paper abstract or electronic download. I nterrater reliability audits and z score analysis was performed by using th e Major Trauma Outcome Study and the CUY 1993 baseline groups. Risk adjustm ent of mortality data was performed using statistical modeling and logistic regression (Trauma and Injury Severity Score, Major Trauma Outcome Study, CUY). Trauma severity measures were defined. Results: In 1995, 3,375 patients were entered. Two hundred ninety-one died (8.6%). Severity measures differed by level of trauma care, indicating diff erences in case mis. Probability of survival was lowest in the Level I cent ers, highest in the acute care hospitals. Outcomes z scores demonstrated su rvival differences for all levels. Conclusions: In a functioning trauma system, the most severely injured pati ents should be cared for at the trauma centers. A low volume at acute care hospitals is desirable. By using Trauma and Injury Severity Score with comm unity-specific constants, NE Ohio is accomplishing these goals, The Level I performance data are an interesting finding compared with the data from th e Level II centers in the region.