Optimizing the ACS NSQIP Risk Calculator in Microsurgical Free Flap Breast Reconstruction

Background: Pre-operative risk indexes such as the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) have become increasingly popular for predicting patient-specific postoperative risks in plastic surgery. This study aimed to assess the accuracy of the ACS NSQIP tool in microsurgical breast reconstruction with subsequent analysis of the most significant patient-level risk variables that drive these predictions.

Methods: A retrospective review of 98 microvascular free flap breast reconstructions performed between 2018-2022 was completed. 94 patients were included in the analysis. Univariate analysis between complication and non-complication groups was completed to identify significant differences in predictions. Receiver operating characteristic (ROC) curves on these predictions were subsequently completed to evaluate the accuracy of ACS NSQIP. Multivariate regression analysis was performed on pre-operative patient data utilized in the NSQIP risk calculator to identify any significant relationships between predictor variables and observed adverse postoperative events.

Results: Univariate analysis revealed the ACS NSQIP calculator most accurately predicted any complication (p=0.0432) and surgical site infection (p=0.0216) of the 12 outcomes analyzed by the authors. ACS NSQIP predicted “any” complications with a diagnostic accuracy of 63.1%, sensitivity of 61.3%, and specificity of 71.0%. It also predicted surgical site infection with a diagnostic accuracy of 71.6%, sensitivity of 57.1%, and specificity of 84.9%. Subsequent multivariate linear regression analysis of significant predictions revealed that non-white individuals were 6.68 times more likely to have any complication (95% CI 1.89-23.62; p=.003) and 6.38 times more likely to have a serious complication (95% CI 1.81-22.53; p=.004).

Conclusion: The NSQIP Risk Calculator was found to be significant in predicting “any complications” and “surgical site infections'' amongst our institution’s microsurgical breast reconstruction cohort, with minorities experiencing an approximate seven-fold increase in complications. Further analysis of the patient-level risk factors that drive these predictions may optimize the accuracy of this index and provide a powerful tool in pre-operative counseling by plastic surgeons.