JMIR Publications

ABSTRACT

Background:

Postoperative complications increase the in-hospital stay and mortality for intestinal obstruction.

Objective:

We aim to construct an online risk calculator for early postoperative complications in patients after intestinal obstruction surgery based on machine learning algorithms.

Methods:

371 intestinal obstruction patients undergoing surgery were enrolled as the training cohort. Multiple machine learning methods were used to establish prediction models, with their performance appraised via the area under the receiver operating characteristic curve. The best model was validated through two independent cohorts, a publicly available perioperative dataset INSPIRE and a mixed cohort composed of all patients, involving 50, 66, 48 and 535 patients, respectively. SHapley Additive exPlanations (SHAP) were measured to identify risk factors.

Results:

The incidence of postoperative complications in the training cohort was 47.44%, while the incidence in four external validation cohorts were 34%, 56.06%, 52.08%, and 47.66%, respectively. Postoperative complications were associated with 8 item features: POSSUM physiological score, the amount of colloid infusion, shock index before anesthesia induction, ASA classification, the percentage of neutrophils, shock index at the end of surgery, age, and total protein. Random forest model performed best, with an AUC of 0.788 in the training cohort. The RF model also achieved a comparable AUC of 0.755, a greater AUC of 0.817, a similar AUC of 0.786, and the best AUC of 0.868 in four external validation cohorts. We visualized the RF model and created a web-based online risk calculator.

Conclusions:

We have developed an online risk calculator for early postoperative complications based on an RF model to assist clinicians in identifying high-risk patients undergoing intestinal obstruction surgery. Clinical Trial: Not applicable.