Establishment of a Risk Prediction Model for Postoperative Acute Moderate to Severe Pain Based on Real-World Data
LIU Decheng
WANG Yimin
CHEN Weixing
YANG Dong
ZHANG Hui
Abstract：Objective To construct and validate a risk prediction model for acute moderate to severe postoperative pain based on real-world medical data.Methods A total of 2434 adult patients who underwent elective anesthesia surgery at Guangdong Second Provincial General,Hospital Affiliated Hospital of Jinan University from April 2020 to April 2021 were screened.The 24-hour postoperative visual analog scale(VAS)score was used as the outcome indicator and divided into two cohorts:mild pain and moderate to severe pain.Indicators associated with the occurrence of moderate to severe postoperative pain were screened and analyzed using univariate analysis.The statistically significant indicators(P<0.05)were used to construct the prediction model.The dataset was divided in a 7:3 ratio for modeling and internal validation.An additional 1200 patients with the same conditions from May 2021 to August 2021 were selected for external validation.The predictive performance of the model was evaluated using five indicators:the area under the receiver operating characteristic(ROC)curve(AUC),accuracy,and F1 score.The shapley additive explanations(SHAP)plot was used to interpret the optimal model.Results The random forest(RF)prediction model showed the best overall performance.Based on SHAP analysis,the relative importance ranking of risk factors was as follows:anesthesia duration,surgery duration,crystalloid usage,body mass index,colloid usage,and age.Conclusion The postoperative acute pain prediction model constructed in this study demonstrates good predictive performance.
Keywords：Postoperative painReal-world studyRisk factorsPrediction modelMachine learning
Publication Date：2024-09-16
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 1-5,后插1 )