Background: The National Surgical Quality Improvement Program (NSQIP) is a risk-adjusted database designed to benchmark quality initiatives. NSQIP captures uniform morbidity variables for all operations and calculates expected morbidity probabilities. Given the frequent need for reoperation following breast-conserving surgery (BCS) and mastectomy, we hypothesized that NSQIP may inaccurately reflect surgical morbidity after breast cancer operations.
Methods: Using the 2008 NSQIP database, we identified 24,447 breast surgery patients. We calculated the observed versus expected (O/E) morbidity ratios, compared them to other general surgery procedures, and analyzed the O/E morbidity ratios among benign and malignant breast diagnoses.
Results: The NSQIP database shows that breast surgery has an O/E morbidity ratio of 3.11, which is higher than other general surgery procedures. Additionally, breast operations for malignancy have higher O/E morbidity ratios (3.22) than those performed for benign disease (2.59). Analysis of malignant patients by CPT code revealed that BCS patients had an O/E morbidity ratio of 7.75 and attributed 89 % of morbidity to reoperation, whereas mastectomy patients had an O/E morbidity ratio of only 1.7. Elimination of the reoperation variable from morbidity calculations in breast surgery reduces the O/E morbidity ratio to less than expected in all breast procedures.
Discussion: Breast surgery has a higher O/E morbidity ratio than other general surgery procedures. Reoperations are expected in BCS for positive margins and in mastectomy for completion ALND. Breast surgeons should advocate for benchmarking by surgical site-specific metrics, because current NSQIP criteria may negatively affect the quality assessment of high-volume breast centers.