BACKGROUND: The aim of this study was to identify independent background and perioperative risk factors for prolonged postoperative hospital stay among patients having a resection for colorectal cancer. METHODS: Data from 1095 consecutive resections performed by specialist colorectal surgeons between 1995 and 2001 were examined by multiple least squares regression. Each putative risk factor was coded 0 if absent and 1 if present and postoperative stay was measured in days, so that the unstandardized partial regression coefficients (B) represent days of additional stay if the factor was present. Confidence intervals are at the 95% level. RESULTS: Of 35 factors examined, the following 11 had a significant independent association with postoperative stay: urgent operation (B = 4.2, CI 2.2-6.2); preoperative stay>5 days (B = 4.2, CI 2.7-5.7); perioperative transfusion (B = 3.1, CI 2.2-4.2); adjacent organ or structure involved (B = 3.0, CI 1.2-4.9); stoma constructed (B = 2.6, CI 1.8-3.5); peripheral vascular disease (B = 2.3, CI 0.4-4.1); age > or =75 years (B = 2.2, CI 1.2-3.1); respiratory disease (B = 1.7, CI 0.5-2.8); American Society of Anesthetists' (ASA) classification>Class 2 (B = 1.5, CI 0.4-2.5); splenic flexure mobilized (B = 1.4, CI 0.5-2.3); private hospital (B = 1.4, CI 0.3-2.5). CONCLUSIONS: Together these factors accounted for only a fifth of the variability in length of stay and few, except possibly ASA, were susceptible of interventions that might reduce stay. Postoperative morbidity, which is largely unpredictable, remains the major cause of prolonged hospital stay.
BACKGROUND: The aim of this study was to identify independent background and perioperative risk factors for prolonged postoperative hospital stay among patients having a resection for colorectal cancer. METHODS: Data from 1095 consecutive resections performed by specialist colorectal surgeons between 1995 and 2001 were examined by multiple least squares regression. Each putative risk factor was coded 0 if absent and 1 if present and postoperative stay was measured in days, so that the unstandardized partial regression coefficients (B) represent days of additional stay if the factor was present. Confidence intervals are at the 95% level. RESULTS: Of 35 factors examined, the following 11 had a significant independent association with postoperative stay: urgent operation (B = 4.2, CI 2.2-6.2); preoperative stay>5 days (B = 4.2, CI 2.7-5.7); perioperative transfusion (B = 3.1, CI 2.2-4.2); adjacent organ or structure involved (B = 3.0, CI 1.2-4.9); stoma constructed (B = 2.6, CI 1.8-3.5); peripheral vascular disease (B = 2.3, CI 0.4-4.1); age > or =75 years (B = 2.2, CI 1.2-3.1); respiratory disease (B = 1.7, CI 0.5-2.8); American Society of Anesthetists' (ASA) classification>Class 2 (B = 1.5, CI 0.4-2.5); splenic flexure mobilized (B = 1.4, CI 0.5-2.3); private hospital (B = 1.4, CI 0.3-2.5). CONCLUSIONS: Together these factors accounted for only a fifth of the variability in length of stay and few, except possibly ASA, were susceptible of interventions that might reduce stay. Postoperative morbidity, which is largely unpredictable, remains the major cause of prolonged hospital stay.