Development and Validation of a Machine Learning Model to Predict Postoperative Morbidity and Mortality in Patients With COVID-19 Who Underwent Major Surgery

Sean P McDermott; Leming Zhou; Ata Murat Kaynar

doi:10.7759/cureus.91247

Development and Validation of a Machine Learning Model to Predict Postoperative Morbidity and Mortality in Patients With COVID-19 Who Underwent Major Surgery

Cureus. 2025 Aug 29;17(8):e91247. doi: 10.7759/cureus.91247. eCollection 2025 Aug.

Authors

Sean P McDermott^{1

2}, Leming Zhou², Ata Murat Kaynar^{1

3}

Affiliations

¹ Department of Anesthesiology and Perioperative Medicine, University of Pittsburgh Medical Center, Pittsburgh, USA.
² Department of Health Information Management, School of Health and Rehabilitation Sciences, University of Pittsburgh, Pittsburgh, USA.
³ Department of Critical Care Medicine, University of Pittsburgh Medical Center, Pittsburgh, USA.

Abstract

Active coronavirus disease (COVID-19) increases the risk of postoperative morbidity and mortality. This study applied machine learning (ML) algorithms to predict postoperative outcomes using preoperative features. Multiple supervised ML models were trained on 153 features from 10,613 patients with COVID-19 who underwent major surgery in 2022 and validated on 5,269 patients from 2021. Among patients with COVID-19, 906 (17.2%) in 2021 and 1,248 (11.8%) in 2022 experienced significantly higher 30-day composite mortality or major morbidity compared with 31,019 (3.2%) of 978,582 patients without COVID-19 in 2021 and 29,874 (3.0%) of 1,001,286 patients without COVID-19 in 2022 (p < 0.001). LightGBM was the best-performing algorithm, achieving an area under the receiver operating characteristic curve (AUC) and F1 score of 0.865 and 0.512 when trained on 2022 data, and 0.898 and 0.617 when validated on 2021 data, respectively. This model accurately identified patients with COVID-19 at increased risk of postoperative morbidity and mortality and may provide a more objective approach to risk stratification.

Keywords: ai and machine learning; covid-19; national surgical quality improvement program (nsqip); postoperative outcome; predictive analytics.