成人住院患者用药风险预测模型的系统评价

打开文本图片集

中图分类号 R952 文献标志码 A 文章编号 1001-008（2025）10-125-06

DOI 10.6039/j.issn.1001-008.2025.10.18

Systematic review on medication risk prediction models for hospitalized adult patients

YANG Yang1，SHAN Xuefeng2，LI Haidong3，LI Yaozheng4，ZHOU Qiwen4，WANG Hongmei4（1. Dept. of Health Management Center，the First Affiliated Hospital of Chongqing Medical University，Chongqing 400016，China； 2. Dept. of Pharmacy，Bishan Hospital of Chongqing Medical University，Chongqing 402760，China；3. Dept. of Science and Technology Education and Foreign Affairs，the Affiliated Stomatological Hospital of Chongqing Medical University，Chongqing 401147，China；4. Dept. of Pharmacy，the First Affiliated Hospital of Chongqing Medical University，Chongqing 400016，China）

ABSTRACT OBJECTIVE To systematically evaluate medication risk prediction models for hospitalized adult patients and provide references for their development and clinical application. METHODS Databases including PubMed，Embase，Web of Science，CNKI，Wanfang data，VIP and CBM were searched for studies on medication risk prediction models from their inception to May 2024. After screening the literature，extracting data，and evaluating the quality of the literature，descriptive analysis was performed on the results of the included studies. RESULTS A total of 13 studies were included，involving 12 models. Nine studies used Logistic regression algorithm for modeling，and the number of included predictive factors ranged from 3 to 11；the area under the receiver operating characteristic curve ranged from 0.65 to 0.865. The literature quality evaluation results showed that 10 studies had high risk of bias；10 studies had high applicability risk. A total of 31 predictive factors were extracted，including 15 items of basic patient information，3 test indicators，and 5 items of medication information，and 8 others. CONCLUSIONS The existing medication risk prediction models for hospitalized adult inpatients are mainly Logistic regression algorithm，with predictive factors mainly focusing on basic indicators such as demographics. The overall prediction performance of the models needs to be improved， and the overall risk of bias is relatively high.

KEYWORDS hospitalized patients；medication risk；drugrelated problems；prediction model；predictive factor

药物治疗是疾病治疗中最重要且广泛的方式之一，但随着治疗药物的广泛使用，药物相关问题（drug-related problems，DRPs）和药物伤害时有发生。（剩余10213字）