ASC-NSQIP Risk Calculator Use in Adult Spinal Deformity Surgery at a Tertiary-Referral University Hospital: Does this Machine-Learning Tool Accurately Predict Risk in ASD Surgery?

Introduction: American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) surgical risk calculator (SRC) is a tool for predicting adverse outcomes and complications following surgical intervention. Its use in the adult spinal deformity literature is limited. The aim of our study is to assess if ACS-NSQIP can accurately predict post-operative outcomes in adult spinal deformity surgery (ASD).

Methods: Consecutive ASD cases were tracked at a tertiary-referral academic hospital for a 6-month period. Cases from one ASD neurosurgeon were compared to national data as well as predicted outcomes by ACS-NSQIP. Patient’s ASA classification, number of risk factors, diabetic status, smoking status, functional status, pulmonary risk factors and case acuity were recorded. Patient complications were documented including superficial surgical site infections (SSI), deep surgical site infections, pulmonary complications, 30-day readmissions, 30-day unplanned return to operating room, venous thromboembolism (VTE), urinary tract infections (UTI), renal failure, cardiac complications, unplanned intubation and mortality.

Results: 526 sites in the United States submitted data to ASC-NSQIP. The individual patient characteristics for the one spinal deformity surgeon were as follows: ASA classification (ASA I 2.3%, ASA II 18.2%, ASA III 77.3%, ASA IV 2.3%), case acuity (95.5% elective, 2.3% urgent, 2.3% emergent), patient risk factors (zero 15.9%, one 44.3%, two 20.5%, three 9.1%, four 5.7%, five or more 4.5%), diabetes mellitus (none 72.7%, non-insulin dependent 12.5%, insulin dependent 3.4%), smoking status (25% smokers), functional status (independent 77.3%, partially dependent 11.4%), pulmonary risk factors (3.8% oxygen support, COPD 6.8%). Patient complication rates for one spinal deformity surgeon compared to ASC-NSQIP expected rates demonstrated: wound occurrences (1.87% vs. 1.97%) pneumonia (5.26% vs. 4.77%), 30-day readmissions (6.9% vs. 5.2%), 30-day unplanned returns to operating room (4.6% vs. 4.21%), VTE (0.92% vs. 1.06%), UTI (1.39% vs 1.31%), renal failure (0.46% vs. 0.47%), cardiac complications (0.92% vs. 0.52%) and mortality (0% vs. 4.23%). The predicted and observed rates for each of the 11 outcomes were within the “As Expected” range according to ASC-NSQIP SRC.

Conclusion : The ACS-NSQIP SRC is a validated tool to accurately predict risk in patients undergoing ASD surgery. This calculator can help spine surgeons mitigate risk, inform patients and optimize patients before surgery.