Construction and validation of a web-based dynamic predictive model for the risk of postoperative nausea and vomiting in patients undergoing day-case hysteroscopic surgery.

Objectives: This study aimed to develop and validate a predictive risk model for postoperative nausea and vomiting (PONV) in patients undergoing day-case hysteroscopic surgery.

Methods: The candidate predictors were identified by systematic literature review. Patients who met the study criteria were divided into training group and validation group. The time-period validation was used for the external validation of the model. The candidate predictors with statistical significance through lasso regression analyses were included in multifactor logistic regression analyses. The calibration and receiver operating characteristic (ROC) curves were utilized to assess the accuracy of model. Decision curve analysis (DCA) was used to assess the clinical benefit of Nomogram. All statistical analyses were constructed by RStudio software (version 4.2.1).

Results: A total of five predictors were included in the PONV risk prediction model: (1) motion sickness (OR, 8.53; 95% CI, 6.21-11.81), (2) anesthesia time (OR, 4.20; 95% CI, 2.09-8.65), (3) fasting time (OR, 1.17; 95% CI, 1.13-1.22), (4) anxiety score (OR, 1.10; 95% CI, 1.08-1.12), and (5) artificial airway (OR, 0.54; 95% CI, 0.39-0.74). The area under the ROC curve for the training cohort and validation cohort was 85.0% (95% CI: 82.6-87.5%) and 80.3% (76.2-84.3%), respectively.

Conclusion: The predictive model demonstrated potential in predicting the risk of PONV in patients undergoing day-case hysteroscopic surgery.