Objective: To review two predictive models, based on the American Society of Anaesthesiologists (ASA) and the Physiological and Operative Severity Score for the enumeration of Mortality and morbidity (POSSUM)-used for estimating postoperative mortality in patients, undergoing surgery for colorectal disease, in the UK.
Methods: Data was derived from three multicentre, UK-based studies involving a total of 16,006 patients with malignant or non-malignant bowel pathologies. Data sources were: The Colorectal-POSSUM (CR-POSSUM) Study population, comprising 6883 patients undergoing colorectal surgery in 15 UK hospitals between 1993 and 2001; The Association of Coloproctology of Great Britain and Ireland (ACPGBI) Colorectal Cancer (CRC) Database, encompassing 8077 newly diagnosed CRC patients, undergoing surgical resections in 79 hospitals, between April 2000 and March 2002; The ACPGBI Malignant Bowel Obstruction (MBO) Study, encompassing 1046 patients with MBO in 148 hospitals, treated between April 1998 and March 1999. Multifactorial logistic regression analyses were used to adjust for case-mix, identify risk factors for in-hospital/30-day operative mortality and to accommodate the variability of outcomes between hospitals.
Results: In the ACPGBI CRC study, 7374 patients had surgery, 6622(89.8%) a major bowel resection and 1465(19.9%) emergency surgery. Nine hundred and eighty-nine (94.6%) patients with MBO had surgery and 854(86.3%) underwent bowel resection. In the CR-POSSUM study, of the 6790(98.6%) patients undergoing surgery, 3451(50.8%) had a major colorectal resection, including 2107(31.0%) as an emergency. The operative mortality was 7.5% for the ACPGBI CRC study, 15.7% for patients with MBO and 5.7% for patients in the CR-POSSUM study. When tested, the predictive models showed good discrimination, with an area under the receiver-operator characteristic curve of 77.5% for the ACPGBI CRC, 80.1% for the MBO and 89.8% for the CR-POSSUM.
Conclusions: Prediction of postoperative death can be made by the clinician using simple, numerical, tables derived from the ACPGBI CRC, MBO and CR-POSSUM models. The models can be used in everyday practice for pre-operative counselling of patients and their carers, as a part of the process of informed consent. They may also be used to compare the outcomes between multidisciplinary CRC teams.