ABSTRACT
BACKGROUND: Prolonged hospitalization (PH) is associated with increasing rates of complications and cost. OBJECTIVES: To detect and quantify causes and predictors for PH in current practice. METHODS: PH was defined as higher than the 70th percentile of the mean hospitalization duration for each department (two medical departments and one acute geriatric ward). Demographic and clinical data were collected on admission in order to determine predictive factors for PH. Actual causes for PH were tested on the 4th day of hospitalization for all patients by: (1) questioning the patients' attending physicians using a structured questionnaire; (2) assessing the patients' charts using a validated tool. RESULTS: Data were collected during a 5-month study period for all 1092 consecutively admitted patients hospitalized in the three departments in the study of whom 337 (30%) had a PH. In the multivariate analysis we detected the following independent predictors for PH: (1) unmarried patients; (2) dependent-patients; (3) hospitalization in the geriatric versus medical department; (4) an expected high mortality rate according to a validated prediction score; (5) renal failure on admission; (6) prior admission in the previous 6 months. The disease leading to the current admissions was the reason for continued admission on day 4 in the patients with eventual PH in 85% of the cases, as compared to 93% in the control group with regular-duration of hospitalization (p=0.014). On day 4, non-medical reasons for prolonged duration were detected in 7% of those with eventual PH as compared to 1% in the control group (p=0.018). CONCLUSIONS: We already detected demographic and clinical predictors for PH on admission, some of which may be amenable to intervention.
