老年人认知衰弱风险预测模型的系统评价

Risk prediction model for cognitive frailty in elderly: a systematic review

目的系统评价老年人认知衰弱风险预测模型,以期为认知衰弱的风险预测提供证据支持。方法系统检索PubMed、Web of Science、Cochrane Library、Embase、CINAHL、中国知网、万方数据库、中国生物医学文献数据库发表的老年人认知衰弱风险预测模型相关文献。检索时限为建库至2023年8月,语种限定为中文和英文。根据纳入与排除标准进行文献筛选,由2名研究者独立进行纳入文献的数据提取和质量评价。提取作者、发表年份、国家、随访时间、研究对象、研究类型、预测因子、模型构建方法及预测效能等数据。结果共纳入16项研究,老年人认知衰弱等发生率为4.8%~65.3%。纳入研究的适用性较好,但总体偏倚风险较高。多个模型重复报告的预测因子有:年龄、营养、活动能力和多病等。结论老年人认知衰弱风险预测模型尚处于发展阶段,部分模型存在显著的方法学缺陷和高偏倚风险。未来的研究中应借助大数据,构建内容全面的老年人认知衰弱风险预测模型并加以验证,形成可行性高的信息化风险预警系统。

Objective To systematically review the risk prediction models of cognitive frailty in the elderly,and to provide evidence support for the risk prediction of cognitive frailty.Methods Systematic search of the literature regarding risk prediction models of cognitive frailty in the elderly in PubMed,Web of Science,the Cochrane Library,Embase,CINAHL,CNKI,Wanfang database,and SinoMed was conducted.The search time was from the inception of the database to August 2023,and the language was limited to Chinese and English.Literature screening was conducted according to the inclusion and exclusion criteria,and data extraction and quality evaluation of the included literature were carried out by two researchers independently.The information of author,year of publication,country,follow-up time,study object,study type,predictors,model construction method,and predictive efficacy was extracted.Results Sixteen studies were included,all of which were studies on model development.The prevalence of cognitive frailty in the elderly ranged from 4.8%to 65.3%.The applicability was good in included studies,but the overall risk of bias was high.Predictors reported repeatedly by multiple models were age,nutrition,activity of daily living and multimorbidity.Conclusion Risk prediction model of cognitive frailty in the elderly is still in the developmental stage,and some models have significant methodological flaws and a high risk of bias.Future research should focus on the establishment of comprehensive risk prediction models and their validation with the help of big data to form an informative risk warning system with high feasibility.