RESUMO
Hospital length of stay (LOS) after liver transplantation has been determined to correlate with liver disease severity, post-transplant survival rates, and transplant-associated costs. A patient's model for end-stage liver disease (MELD) score and an organ's Donor risk index (DRI) have both been found to be significant predictors of LOS, but these two factors alone are insufficient to form an accurate prediction. Previous studies have identified other factors predictive of LOS, which can be incorporated with MELD and DRI to create more specific results. The objective of this study was to create an algorithm, or models, based on the most significant LOS predictors as identified from national data at different stages of the transplant process. Four models were developed predicting LOS using recipient factors, payment factors, donor factors, and postoperative factors. A medical care team member can enter a patient's data into the model and receive a reasonably accurate prediction of LOS for each phase of the liver transplant process, specifying the impact of each factor. These predictions would help predict the factors most likely to prolong LOS, inform resource allocation, and provide patients with more specific predictions of their LOS following transplantation.