社区老年衰弱风险预测模型系统评价

目的 系统评价社区老年衰弱风险预测模型性能。方法 计算机检索PubMed、EMbase、Web of Science、The Cochrane Library、CBM、VIP、WanFang Data和CNKI数据库,搜索关于社区老年人衰弱风险预测模型的研究。检索时限均为建库至2022年8月20日。由2名研究者独立筛选文献、提取数据,并应用预测模型的偏倚风险评估工具（prediction model risk of bias assessment tool,PROBAST）分析纳入文献的偏倚风险和适用性。结果 共纳入10个研究,8个模型开发研究,2个模型开发及验证研究。10个模型的受试者工作曲线下面积为0.695~0.940。所有纳入模型中包含最多的预测因子是年龄。10个研究的适用性较好,但存在一定偏倚,主要是因为缺失数据不详、样本量不足、缺乏模型性能评估及模型过度拟合。结论 社区老年衰弱风险预测模型尚存在一些不足,未来研究应根据预测模型报告规范,完善研究设计和研究报告,并进行内部及外部验证。

Risk predictive models for frailty among community-dwelling older people: a systematic review

Objective To systematically evaluate the performance of community frailty risk prediction model. Methods PubMed, EMbase, Web of Science, The Cochrane Library, CBM, VIP, WanFang Data and CNKI databases were searched for studies on frailty risk prediction models for the elderly in the community. The search period was from the establishment of the database to August 20, 2022. Two researchers independently screened the literature, extracted the data, and applied the bias risk assessment tool ( PROBAST ) of the prediction model to analyze the bias risk and applicability of the included literature. Results A total of 10 studies, 8 model development studies and 2 model development and validation studies were included. The area under the receiver operating curve of the 10 models was 0.695-0.940. The most predictive factor included in all included models was age. The applicability of the 10 studies was good, but there was a certain bias, mainly due to missing data, insufficient sample size, lack of model performance evaluation and model overfitting. Conclusion There are still some deficiencies in the community frailty risk prediction model. Future research should improve the research design and report according to the prediction model reporting specification, and conduct internal and external validation.