Browsing by Author "Bennett, Adam Jacob"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Item A Simplified Risk Score for Predicting the Incidence of Major Complications after Complex Abdominal-Pelvic Resections(2016-04-01) Bennett, Adam Jacob; Mansour, John; Brancaccio, Anne; Abraham, ReeniBACKGROUND: The POSSUM system is used to predict risk of complications following general surgical procedures. This 18-factor instrument has been challenging to apply to most surgical oncology patient populations. Our aim is to develop a simplified scoring system that is highly correlated with the incidence of major complications. OBJECTIVE: To develop a simplified scoring system that is highly correlated with the incidence of major complications. METHODS: We queried a single-institution IRB-approved prospective database from a surgical oncology population from January 2008 to December 2012. We identified patients undergoing complex abdominal or pelvic resections and factors associated with the development of major (Clavien-Dindo ≥ III) complications. Factors not included in the POSSUM system were incorporated into a new scoring system based on univariate correlation with complication rates (Chi-square). Optimal binning generated an ideal cut-off value associated with major complications. A composite scoring system (SOPI) was compared to standard POSSUM predictions using ROC analysis. RESULTS: We identified 831 patients undergoing pancreatic (23%), hepatic (23%), colorectal (22%), esophagogastric (16%), retroperitoneal (4%), combined (3%), or other type (10%) of resection. Major complications occurred in 17% of patients. Two original POSSUM factors were included in the new SOPI model (cardiac history and EBL). Four factors improved correlation with complication rate: gender (female/male-1/3 points); operation type (retroperitoneal/ pancreatic or rectal/others-4/2/1 points); curative intent (curative/non-curative-1/2 points); and cancer (no cancer/cancer-1/2 points). ROC analysis generated a greater AUC for the simplified 6-factor system than standard 18-factor POSSUM (AUC: 0.676 vs 0.631). Increasing SOPI quartiles higher risk of major complications (5%, 15%, 20%, 29%; p-value < 0.001). CONCLUSION: The efficacy of POSSUM or P-POSSUM for complex abdominal and pelvic resections remains unclear. The SOPI system, which is comprised of only 6 factors, was equivalent to P-POSSUM for predicting major complication rates (Clavien-Dindo score of at least III) for surgical oncology patients. Validation in a large, independent dataset is necessary before the system can be widely applied.