Analyzing hospital high length of stay outliers in Turkey

Abstract

Purpose: The study aims to examine the length of stay (LOS) outliers by analyzing the hospital administrative database.Material and Methods: The Turkish Ministry of Health DRG grouper database was utilized to obtain hospital administrative data on discharges for 15 training and research hospitals in 2012. For each diagnosis-related group (DRG), the geometric mean plus two standard deviations were calculated to identify the outliers. According to the recent data, demographic, clinical, and DRG-specific characteristics influence LOS outliers. Analyses were conducted using descriptive statistics and logistic regression using generalized estimating equations (GEE).Results: High LOS outliers were found to be 4.4 % of the cases. Even though they made up only 4.4 percent of the patients, outliers accounted for 24.50 percent of all discharge days. The top three specific DRGs with the highest LOS outliers were neonatal cases. According to the multivariate model, alcohol, drug use disorders, burns, and ear, nose, mouth, and throat diseases influenced high LOS outliers the most. Furthermore, DRG weight, DRG type, discharge type, and age groups were other significant factors related to high LOS outliers.Conclusion: A quarter of all inpatient days are made up of LOS outliers. Burns, neonate cases, and alcohol/drug use cases should be carefully evaluated. To improve clinical quality and effectively manage hospital resources, hospital administrators and health policymakers should consider the length of stay outliers.