Correlates of length of stay in gynecologic oncology patients undergoing inpatient surgery

Abstract

Escalating economic constraints require identification of the determinants of length of stay (LOS), since optimizing these may facilitate cost-effective care. However, risk factors for increased LOS are poorly understood. In order to identify correlates of long LOS, we retrospectively reviewed the charts of 252 consecutive patients who underwent major inpatient surgery on the gynecologic oncology service at Barnes Hospital during 1990. Of these, 86 had benign disease and 38 were excluded for other reasons, leaving 128 for analysis. The recorded variables of age, race, medical history, measures of nutritional state, primary cancer, prior therapy, blood loss, operating time, and attending surgeon were subjected to regression analysis. In order to avoid reducing cell numbers to insignificant sizes, the linked factors of type of surgery and primary site were not analyzed simultaneously. A separate analysis was performed substituting type of procedure for primary site. The mean LOS was 12 days (range, 4-30). Five factors emerged as significant correlates of longer LOS: decreasing preoperative albumin level (P < 0.001) and hemoglobin level (0.029), increasing age (0.017), operative blood loss (0.039), and prior platinum-based chemotherapy (0.043). Presence of vulvar and fallopian tube primaries was associated with significantly shorter LOS (0.023 and 0.037). Together these factors accounted for 36% of total variance in LOS. Substituting procedure type for primary site accounted for 37% of total variance in LOS and showed that only bowel surgery (0.001) and preoperative intracavitary implant were significantly correlated with longer LOS. Factors not associated with LOS (P > 0.05) were race, attending physician, number of preoperative medications or medical conditions, body mass index, other radiotherapy, and other primary sites and procedures. Future studies should attempt to determine why elderly, nutritionally depleted women undergoing long or complicated procedures are at high risk for long LOS and should assess interventions designed to minimize LOS for this high-risk group.