A Meta-analysis of the American College of Surgeons Risk Calculator’s Predictive Accuracy in Spine Surgery

Abstract

Background: The American College of Surgeons Surgical Risk Calculator (ACS-SRC) has been widely utilized to provide risk estimates of postoperative complications after a surgical procedure. While several retrospective studies have examined the accuracy of the ACS-SRC for specific procedures within spine, these are limited by sample size and institutional biases. Objective: We sought to conduct a meta-analysis to determine the comparative accuracy of the ACS-SRC across postoperative complications in spine. Materials and Methods: Clinical studies in spine surgery that utilized the ACS-SRC, predicted complication rates compared to actual rates, and analyzed at least one metric reported by ACS-SRC met the inclusion criteria. Data for each complication were pooled by the log-transformed event rates using the DerSimonian and Laird random-effect models. All analyses were performed using the binary random-effect model to produce risk difference (RD) and 95% confidence intervals (CIs). Heterogeneity was assessed using the I 2 statistic. Statistical analyses were conducted using Open Meta [Analyst]. All values were two tailed, and P < 0.05 was set as the threshold for statistical significance. Results: After an initial PubMed search of the ACS-SRC yielded 53 studies, a total of 7 studies focused on spine surgery with 12,104 patients across 12 complications. Included studies assessed the ACS-SRC for a variety of procedures: fusion, laminectomy, and deformity correction. The ACS-SRC significantly underpredicted serious complications (RD: −0.074, 95% CI: −0.139 to −0.008, P = 0.027), any complications (RD: −0.131, 95%: CI −0.203 to −0.060, P < 0.001), cardiac (RD: −0.025, 95% CI: −0.040 to −0.011, P < 0.001), venous thromboembolism (VTE) (RD: −0.024, 95% CI: −0.047 to −0.001, P = 0.043), surgical site infection (SSI) (RD: −0.023, 95% CI: −0.043 to −0.004, P = 0.020), and pneumonia (RD: −0.017, 95% CI: −0.306 to −0.005, P = 0.007). There was no significant difference between the actual and ACS-SRC predicted values for discharge to a nursing facility (RD: 0.028, 95% CI: −0.095–0.151, P = 0.654), readmission (RD: −0.015, 95% CI: −0.037–0.001, P = 0.170), renal failure (RD: −0.010, 95% CI: −0.021-0.001, P = 0.086), urinary tract infection (RD: −0.005, 95% CI: −0.014-0.004, P = 0.300), return to OR (RD: 0.003, 95% CI: −0.014–0.019, P = 0.756), and death (RD: 0.000, 95% CI: −0.005–0.006, P = 0.893). Conclusion: Within spine, the ACS-SRC can be an effective tool in predicting select, major complications. Complications such as death, return to OR, discharge to nursing facility, and readmission were accurately predicted. The ACS-SRC is unable to comprehensively show risk of various complications, however. It significantly underpredicts serious complications, any complications, cardiac, VTE, SSI, and pneumonia. To understand the risk of these, it may be advantageous to use the ACS-SRC alongside a specialty or procedure-specific calculator.