Validation of the ACS-NSQIP Risk Calculator: A Machine-Learning Risk Tool for Predicting Complications and Mortality
Katherine E Pierce1, Bhaveen H Kapadia2, Sara Naessig1
1Departments of Orthopaedic and Neurologic Surgery, NYU Langone Orthopedic Hospital; New York Spine Institute, New York, New York, USA.
International Journal of Spine Surgery
|January 28, 2022
Summary
The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) Surgical Risk Calculator (SRC) accurately predicts postoperative complications for adult spinal deformity (ASD) surgery. This tool aids surgeons in preoperative patient optimization and risk assessment.
Area of Science:
- Spine Surgery
- Surgical Outcomes Research
- Health Informatics
Background:
- Adult spinal deformity (ASD) surgery carries significant risks of postoperative complications and mortality.
- Accurate risk prediction is crucial for patient counseling and preoperative optimization.
Purpose of the Study:
- To evaluate the predictive performance of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) Surgical Risk Calculator (SRC) for postoperative complications and mortality in adult spinal deformity (ASD) patients undergoing corrective surgery.
- To compare the calculator's predicted risks with observed outcomes in a large ASD patient cohort.
Main Methods:
- A cohort of adult patients (≥18 years) undergoing ASD corrective surgery was identified using specific Current Procedural Terminology (CPT) codes and International Classification of Diseases Ninth Revision (ICD-9) scoliosis diagnoses.
- Perioperative complication risk averages were calculated using the ACS-NSQIP SRC and compared with observed complication rates.
- The predictive performance of the calculator was assessed using the Brier score for binary outcomes and a Bland-Altman plot for length of stay (LOS).
Main Results:
- The study included 9143 ASD patients, with a mean age of 58.9 years and 56% females. Procedures involved decompression (36.9%) and fusion (100%).
- The ACS-NSQIP SRC demonstrated good predictive performance for 30-day postoperative complications, with predicted rates ranging from 2.8% to 18.5% compared to an observed rate of 11.4% (Brier score: 0.000002).
- The mean difference between observed and predicted length of stay was 2.375 days.
Conclusions:
- The ACS-NSQIP SRC is a valuable tool for predicting surgical risk in patients undergoing corrective surgery for adult spinal deformity.
- This calculator can assist deformity surgeons in the preoperative optimization process by providing reliable risk estimates.
