Predictors of Length of Stay for Inpatients Having Benign Gynaecological Surgery

Abstract

Objective: To determine factors predictive of length of stay (LOS) for women in case-mix group (CMG) 579 (major uterine and [or] adnexal procedures—no malignancy).Methods: All inpatients having benign gynaecologie surgery in CMG 579 between April 1, 1997, and March 30, 1999, were studied. We obtained information from the hospital costing and registration database and from retrospective chart review. Abstracted data included patient demographics, age, admission weight, American Society of Anesthesiologists (ASA) risk score, pre- and postoperative diagnoses, procedure, surgeon, day of the week of surgery, LOS, and costs. Results were considered statistically significant when P ≤ 0.05.Results: There were 606 gynaecology inpatients in the study, with a mean age of 45.1 years. Women were categorized into 4 procedure groups: abdominal (59.2%), vaginal (26.7%), abdominal repair (pelvic floor relaxation) (9.5%), and other (4.6%) forms of surgery. The LOS was highly correlated with overall case costs. Univariable analyses indicated that procedure (R2 = 0.22, P < 0.001), age (R2 = 0.12, P < 0.001), surgeon (R2 = 0.11, P < 0.001), postoperative diagnosis (R2 = 0.11, P < 0.001), ASA score (R2 = 0.09, P < 0.001 ), day of the week of surgery (R2 = 0.03, P < 0.018), and weight (R2 = 0.01, P < 0.017) were all determinants of LOS (all P < 0.001). Multivariable analyses demonstrated that the procedure, age, surgeon, postoperative diagnosis, and ASA score were the significant factors accounting for the most variance in LOS (R2 = 0.46, P < 0.001).Conclusion: The procedure performed, patient’s age, surgeon, postoperative diagnosis, and ASA score were significantly associated with length of stay. For better accuracy and validity, CMG-based comparisons of LOS should take these factors into account.