Risk Stratification and Evaluation of American College of Surgeons National Surgical Quality Improvement Program Risk Calculator for Patients Undergoing Panniculectomy
Visakha Suresh, BSE1, Heather Levites, MD2, Sarah Peskoe, PhD1, Rachel Hein, MD2, Yash Avashia, MD2, Detlev Erdmann, MD, PhD, MHSc2.
1Duke University School of Medicine, Durham, NC, USA, 2Duke University Medical Center, Durham, NC, USA.
Purpose: Panniculectomy procedures have been reported to drastically improve quality of life, increase mobility, and improve hygiene in patients affected by a significant pannus formation. While the indications for performing the procedure are well-defined, medical comorbidities, such as morbid obesity, increase the risk of post-operative adverse events for a generally elective procedure.. Thus, the primary aim of this study is to retrospectively review all panniculectomy procedures performed at our institution, and to determine what pre-operative risk factors, such as BMI, may be used to differentiate the rate of post-operative complications amongst patient cohorts and to validate the use of the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) risk calculator in this patient population.
Methods: This study is a single center retrospective analysis. All patients that undergoing a procedure with the CPT code “15830” at our institution from January 2005 to December 2016 were included in this study. Baseline characteristics, preoperative risk factors, and post-operative complications were recorded via chart review. Validation of the ACS-NSQIP risk calculator was performed ssing statistical measures originally utilized in the development of this model.
Results: 264 patients who underwent panniculectomy procedures were identified. The odds ratio of any post-operative complication was 8.26 (95% CI: 2.51-27.1) for patients with class 1 obesity, 7.76 (95% CI: 2.13-28.3) for patients with class 2 obesity, and 16.6 (95% CI: 5.13-53.9) for patients with class 3 obesity, after adjusting for age, gender, smoking status, and diagnosis of diabetes. In the evaluation of the ACS-NSQIP Surgical Risk Calculator, we performed concordance measures. We calculated the c-statistic, a measure of discrimination ranging from 0.5 (chance) to 1.0 (perfect) and reflects the extent to which cases are properly classified as having or not having an event. For this patient population, the c-statistic for the ACS-NSQIP model was only 0.61, showing that although the model is associated with the risk of complication, it does not have a strong predictive value for this particular procedure. We also performed calibration measures to assess how close the actual predicted risks are to the observed rates of complications. We calculated the Brier score, which calculates the average squared difference between patients' predicted probability and the actual outcome (0 for a non-event and 1 for an event). The Brier score for any post-operative complications in this patient population was only 0.55. Thus, based on our analyses, the ACS-NSQIP Risk model is a statistically significantly poor fit for estimating post-operative complications in patients undergoing panniculectomy procedures.
Discussion: This study is one of the first to characterize post-operative complication rate based on extremum of BMI, i.e. obesity Classes 1, 2, and 3, for patients undergoing panniculectomy procedures. Additionally, our results show that the utilization of the ACS-NSQIP Risk Calculator in this particular patient population may underestimate the complication risk as a whole, which may necessitate the development of a separate risk assessment model for this procedure in the future.
Back to 2018 Program