A QUANTITATIVE METHOD FOR COST REIMBURSEMENT AND LENGTH OF STAY QUALITY ASSURANCE IN MULTIPLE TRAUMA PATIENTS

Objective: To develop a statistically valid method for trauma reimburs ement and quality assurance (QA) length-of-stay filters. This is neede d because diagnosis related group (DRG)-based trauma payment systems a ssume a random sampling of injury severities from a normally distribut ed population and thus result in economic disincentives to level I tra uma centers. Methods: 142 trauma patients with MVC blunt multisystem i njuries (MSI) (ISS greater than or equal to 16) were studied concurren tly during their hospital course. Setting: Level I regional trauma cen ter. Outcome Measures: Outcome measures were (dependent variables) len gth of stay (LOS) and state-approved hospital charges (COST). Results: Mean acute care COST was $74,310, but the distribution of COST was lo g normal, rather than Gaussian normal as assumed by DRGs. The LOS for MSI was more than twice the average for all trauma (22 vs. 9 days), re flecting skewed severities of level I patients and was related to COST (r(2) = 0.802; p < 0.0001). The ISS alone was a weak determinant of C OST or LOS (r(2) = 0.05; p < 0.0001). The best single determinant of C OST and LOS was survival (r(2) = 0.15; p < 0.0001): as it increased, i t increased LOS. The most costly injuries (all p < 0.0001) involved th e lower extremity (LE) or hip joint (HIP), whereas sepsis and pulmonar y and surgical complications constituted the most costly complications (all p < 0.0001). Regression models that accounted for the log-normal distribution of the dependent variable and based on binary variables for survival, LE and HIP injuries, and the complications of sepsis, AR DS, pulmonary failure, MOFS, plus ISS, explained nearly two thirds of the variability in COST (r(2) = 0.621; p < 0.0001) or LOS (r(2) = 0.68 7; p < 0.0001) and the residuals were normally distributed. Conclusion s: These models provide a valid method of reimbursement for MSI trauma for level I trauma centers, since the data imply that good care assoc iated with survival from specific complications of MSI are the major d eterminants of COST, rather than the specific type of injury or the re sultant ISS. Moreover, using survival and ISS plus the disease-related complications as determinants of LOS, this method can be applied to a ny U.S. region since local factors can be used to adjust hospital COST as a highly correlated function of LOS. This method also permits iden tification of LOS outliers for QA, taking into account the influence o f injury complications.