External Validation of the NELA Prognostic Model in Emergency Abdominal Surgery

Area of Science:

• Surgical outcomes research
• Clinical prediction modeling
• Health services research

Background:

• Emergency abdominal surgery carries significant risks.
• Accurate risk prediction models are crucial for patient management and resource allocation.
• External validation of existing models is essential to confirm their generalizability.

Purpose of the Study:

• To externally validate the National Emergency Laparotomy Audit (NELA) prognostic model (P-NELA).
• To compare the performance of the P-NELA model with the previous 2020 NELA risk prediction model.
• To assess the prognostic performance of a simplified emergency laparotomy prediction model.

Main Methods:

• A cohort of 679 adult patients undergoing emergency laparotomy or laparoscopy for gastrointestinal conditions was analyzed.
• The primary outcome was 30-day mortality; the secondary outcome was 90-day mortality.
• Model performance was evaluated using scaled Brier score, Area Under the Curve (AUC), and Integrated Calibration Index.

Main Results:

• The P-NELA model demonstrated strong discrimination with an AUC of 0.87 (95% CI: 0.83-0.90) for 30-day mortality.
• The scaled Brier score was 18% and the Integrated Calibration Index was 0.042 for 30-day mortality.
• The performance of the P-NELA model was comparable to the previous NELA risk prediction model.

Conclusions:

• The P-NELA prognostic model exhibits good discrimination in external validation.
• A reduced number of variables in the P-NELA model did not compromise its predictive performance.
• The simplified P-NELA model demonstrates good prognostic performance for emergency laparotomy patients.