中国成年人群食物频率问卷信效度Meta分析

A META-ANALYSIS OF THE REPRODUCIBILITY AND VALIDITY OF FOOD FREQUENCY QUESTIONNAIRES AMONG CHINESE ADULTS

目的系统分析中国成年人群中食物频率问卷(food frequency questionnaire,FFQ)的信效度研究现状,为进一步推广应用FFQ提供参考依据。方法基于中国知网、万方数据知识服务平台、维普数据库、PubMed、Embase,系统检索1992年至2021年12月的所有中国成年人群FFQ信效度研究文献,文献筛查、评价和数据提取后进行Meta分析。结果最终纳入30篇文献,样本量范围为22-1623名。FFQ食物清单数目为25-172种,5项研究报告了针对FFQ建立相应的食物成分表以计算营养素摄入量,但大多数研究仍然缺乏FFQ设计、实施及数据处理标准化流程的详细说明。总体看,FFQ评估膳食因素的平均信度较好(合并效应值>0.4)。在效度检验中,FFQ对比24小时膳食回顾(24-hour dietary recall,24HR)的相关系数合并效应值:营养素为0.46(95%置信区间(confidence interval,CI):0.39-0.52),食物组为0.43(95%CI:0.36-0.50);对比膳食记录法(dietary record,DR):营养素为0.36(95%CI:0.30-0.42),食物组为0.45(95%CI:0.37-0.52)。FFQ食物清单数量、回忆时间框、研究设计质量、是否调整能量或校正个体内差异均可影响FFQ的效度结果。此外,单一水平看,营养素中维生素B1,食物组中薯类、禽肉、坚果、食用油和盐的测量有效性相对较差。结论中国成年人群FFQ设计及信效度研究结果存在一定差异,整体具有中等可接受的信效度水平,提示FFQ在大型人群研究中开展膳食评估具有良好的应用价值。同时FFQ应用的规范化标准流程仍然需要完善与统一,亟待更多的研究优化规范及推广使用。

Objective To conduct a meta-analysis of the reproducibility and validity of food frequency questionnaires(FFQ)among Chinese adults.Methods Eligible studies published from 1992 to September 2020 were retrieved from the following database:CNKI,WanFang,VIP,PubMed and Embase.Meta-analysis was used to calculated the pooled coefficients for reproducibility and validity based on literature screening and data extraction.Results A total of 30 studies among Chinese free-living adults were included with the sample size ranging from 22 to 1623.The number of food items ranged from 25 to 172.A total of 5 studies described the development of corresponding FFQ food composition tables for the calculation of nutrient intakes.However,the standard procedures for FFQ development,implementation and data analysis were not described in most studies.Participants were aged 18 years older,and most studies had relatively equal proportion of both genders.In general,the FFQ had a modest to high level of reproducibility(pooled correlation coefficient>0.4).Regarding the validity of FFQ,compared with 24-hour diet recalls(24HR),the pooled correlation coefficients(95%confidence interval,95%CI)were 0.46(0.39-0.52)for nutrients and 0.43(0.36-0.50)for food groups;Compared with dietary record(DR),the pooled correlation coefficients(95%confidence interval,95%CI)were 0.36(0.30-0.42)for nutrients,0.45(0.37-0.52)for food groups.The number of food items,recall interval,study quality,and analysis methods potentially influenced the validity of FFQs.Nevertheless,the validity of several specific dietary factors such as vitamin B1,tubers,poultry,nuts,oils and salt was relatively poor.Conclusion Although large variations exit across the validation studies,the Chinese FFQ has moderate to high reproducibility and validity despite some low performance for several specific dietary factors.In large population studies,FFQ could provide valuable dietary information,further studies are warranted to refine the standard procedure of FFQ implementation and extend its application.