External Validation of the NELA Prognostic Model in Emergency Abdominal Surgery

BACKGROUND: Emergency abdominal surgery is a high-risk procedure, and risk prediction models can aid decision-making and resource allocation. This study aimed to externally validate the National Emergency Laparotomy Audit (NELA) prognostic model. Second, we compared the 2023 NELA prognostic model (P-NELA) with the 2020 NELA risk prediction model.

METHODS: We included all adult patients undergoing emergency laparotomy or laparoscopy for surgical gastrointestinal conditions such as perforation, obstruction, or ischaemia at Copenhagen University Hospital-Herlev from 1 November 2017 to 31 January 2020. The primary outcome was 30-day mortality, and the secondary outcome was 90-day mortality. We evaluated overall performance, discrimination and calibration with the scaled Brier score, area under the curve (AUC) and Integrated Calibration Index.

RESULTS: We included 679 patients, of whom 13.3% (90 patients) died within 30 days. For the P-NELA prognostic model, the scaled Brier score was 18%, AUC (95% confidence interval) was 0.87 (0.83-0.90) and the Integrated Calibration Index was 0.042 for 30-day mortality. Results for the NELA risk prediction model were similar.

CONCLUSION: The P-NELA prognostic model had good discrimination in this external validation. The reduced number of variables in this model compared to the previous NELA risk prediction model did not affect performance.

EDITORIAL COMMENT: This analysis presents testing of a simplified version of an emergency laparotomy prediction model for postoperative complications on a new and independent cohort. The new model demonstrated good prognostic performance.