Application of Artificial Intelligence Technology in the Management of An-tibiotic Usage Intensity in the Gastroenterology Department
Objective:To apply artificial intelligence technology to reduce the intensity of antimicrobial use in the gastroenterology department.Methods:The case data of 1752 inpatients in the gastroenterology department of our hospital from February to April 2022 were collected.Based on the infection-related test indicators and the defined daily doses(DDDs)value of antimicrobial use,a classification prediction model was constructed to predict whether patients will use antimicrobials and the intensity of antimicrobial use.For patients with the classification prediction result of"antimicrobials can be used",a regression prediction model was constructed to predict the DDDs value of the antimicrobials that should be used.The XGBoost algorithm was used to construct the classification prediction model,and a total of 5 regression algorithms including LGBMRegressor,XGBRegressor,BaggingRegressor,PLSRegression and LassoLarsCV were used to construct the regression prediction models.Cross-validation and an independent test set were used to eval-uate the performance of the models.Finally,the models were integrated into the clinical decision support system and applied in the gastroenterology ward.Results:Using the artificial intelligence algorithm con-structed in this paper to predict the use of antimicrobials,the mean DDDs of antimicrobials in the gas-troenterology department from June to November 2022 was 32.03,which was significantly lower than that from June to November 2021(62.67)(P＜0.001),while the average length of hospital stay and hospitaliza-tion costs showed no significant changes compared with the same period in 2021(P＞0.05).Conclusion:The artificial intelligence-based antimicrobial intelligent prediction and early warning system is helpful to effectively reduce the intensity of antimicrobial use in the gastroenterology department.
GastroenterologyAntimicrobial use intensityClassification modelRegression model
万方数据
维普
