Predictors of length of stay for transurethral prostatectomy in Victoria

Background: Transurethral resection of prostate (TURP) is among the top 10 surgical conditions that account for hospital admission in Victoria Bed uti lization for TURF is an increasing concern in current times. This paper des cribes trends in length of stay (LOS) and identifies predictors of LOS for TURF in Victoria. Methods: Trends in TURF were studied using ICD-9-CM coded Victorian hospita l morbidity data from public hospitals from 1987/88 to 1994/95. Detailed mo rbidity data from the same source for the financial year 1995/96 were used to study predictors of LOS by logistic regression. Results: Length of stay decreased significantly between 1987 and 1995 from 10.6 to 6.1 days. The strongest predictor of increased LOS was admission th rough the emergency room (odds ratio (OR) 14.7; 95% confidence interval (CT ) 11.8-18.3). Other significant predictors were older age, lower socio-econ omic status, presence of comorbid conditions, occurrence of procedural morb idity, and hospital type and location. Conclusions: The trend in decreasing LOS may be explained by increasingly e fficient bed management in hospitals who are faced with an increasing need for cost control. Advances in surgical techniques and peri-operative care h ave also contributed to the decrease in LOS. Other factors that influence L OS can be divided into three categories: intrinsic patient factors, such as co-morbid conditions; procedure-specific factors such as peri-operative mo rbidity; and intrinsic hospital factors relating to capacity and resources. Such determinants of LOS may be of value to policy makers when considering the effective application of newer methods for treatment of benign prostat ic hyperplasia.