Construction and optimization of automatic monitoring module for drug-induced movement disorders based on hospital information system data
10.3760/cma.j.cn114015-20240618-00416
- VernacularTitle:基于医院信息系统数据的药源性运动障碍自动监测模块构建及优化
- Author:
Liqiang CUI
1
;
Daihong GUO
;
Man ZHU
;
Tianlin WANG
;
Ao GAO
;
Anqi ZHAO
;
An FU
;
Jing XIAO
Author Information
1. 解放军医学院，北京 100853
- Publication Type:Journal Article
- Keywords:
Dyskinesia, drug-induced;
Artificial intelligence;
Automatic monitoring;
Text classification technology;
Data mining;
Real world study;
Sodium valproate
- From:
Adverse Drug Reactions Journal
2025;27(2):84-90
- CountryChina
- Language:Chinese
-
Abstract:
Objective:Based on the adverse drug event active surveillance and assessment system-Ⅱ (ADE-ASAS-Ⅱ) and the information of inpatients in the hospital information system (HIS), the automatic monitoring module of movement disorders was constructed and its application effect in the real-world study of drug-induced movement disorders (DIMDs) was explored.Methods:Literature reviews, case reports, spontaneous reports and medical records were collected, the keyword set was screened based on ADE-ASAS-Ⅱ system and text classification technology, and an automatic monitoring module was constructed. The information of hospitalized patients in Chinese PLA General Hospital (our hospital) was selected from October 10 to 16, 2022. The results of manual evaluation and the system alarm by the automatic monitoring module were compared, and the performance of the automatic monitoring module was evaluated and optimized through repeated machine learning. The medical record information of hospitalized patients who used sodium valproate throughout the year in our hospital in 2022 were collected, and the occurrence of movement disorders related to sodium valproate was analyzed using the automatic monitoring module.Results:A total of 4 918 hospitalized patients (146 with movement disorders) were collected, and the final setting conditions of the automatic monitoring module were determined, including inclusion criteria (43 text keywords, 3 diagnosis) and exclusion criteria (11 text and 20 document titles were omitted). Among the 1 138 hospitalized patients using sodium valproate in 2022, the incidence of DIMDs with tic and tremor as main clinical manifestations detected by automatic monitoring module was 1.67% (19/1 138).Conclusion:The automatic monitoring module of drug-induced movement disorders based on machine learning and manual evaluation can be applied to explore the occurrence characteristics of DIMDs in the real world, and provide information for pharmacovigilance in clinic.
