Department of Public Health and Clinical Medicine/Sustainable Health, Umeå University, Umeå, S-901 87, Sweden.
Department of Odontology, Umeå University, Umeå, Sweden.
Nutr J. 2024 Jul 24;23(1):83. doi: 10.1186/s12937-024-00984-8.
Dietary habits strongly influence health, with poor diets contributing to numerous deaths annually. Addressing this requires improved dietary habits and consistent monitoring thereof. In northern Sweden, a validated food frequency questionnaire (FFQ) has been used for decades, but trends show that its ability to accurately measure intake has diminished. With changing eating habits and food supply, updating the FFQ was crucial, leading to the development of FFQ2020. This study assessed FFQ2020's relative validity using 24-hour recalls and evaluated its reproducibility.
Participants were recruited from one of the northern-Sweden population-based health screenings and by advertising. Food intake was registered in an electronic food frequency questionnaire (FFQ2020) (test instrument) and reference data were obtained by six repeated electronic 24-hour dietary recalls (24HDR). Intakes of single foods were aggregated into food groups and healthy diet index scores, and daily energy and nutrient intakes were estimated. Results from the two methods were described and tested in univariate analyses and correlation tests, Bland Altman plots, cross-classification validity, and intra-class correlation analyses.
Totally, 628 adults were invited to participate in the study. Of these, 320 joined, and 244 completed at least four 24HDRs. The median intakes in food groups, as well as the mean index scores and estimated nutrient intakes, were largely similar between the FFQ2020 and 24HDR recordings. The correlation coefficients between the two assessments ranged from 0.253 to 0.693 for food groups, 0.520 to 0.614 for diet indices, and 0.340 to 0.629 for energy and nutrients. Intra-class correlation coefficients indicated at least good reproducibility for intakes of food groups, diet index scores, and nutrients. Generally, Bland-Altman plots did not reveal any gross systematic disagreement between the two methods for any of the assessments. However, there were single observations located outside the upper or lower 95% confidence interval (CI) limits for the difference between FFQ2020 and the 24HDR recordings.
In concert, the results suggest that the relative validity and reproducibility of FFQ2020 are acceptable for trend analyses and group comparisons in large-scale studies but also that extended reference periods would improve the precision of less frequently consumed foods.
饮食习惯对健康有很大影响，不良饮食每年导致许多人死亡。要解决这个问题，需要改善饮食习惯并对其进行持续监测。在瑞典北部，一种经过验证的食物频率问卷（FFQ）已经使用了几十年，但趋势表明，它准确测量摄入量的能力已经下降。随着饮食习惯和食物供应的变化，更新 FFQ 至关重要，这导致了 FFQ2020 的发展。本研究使用 24 小时回顾法评估了 FFQ2020 的相对有效性，并评估了其可重复性。
参与者是从瑞典北部一项基于人群的健康筛查和广告招募而来的。食物摄入量是在电子食物频率问卷（FFQ2020）（测试仪器）中记录的，参考数据是通过六次重复的电子 24 小时膳食回顾（24HDR）获得的。单一食物的摄入量被汇总到食物组中，并计算出健康饮食指数评分，以及估计的每日能量和营养素摄入量。两种方法的结果在单变量分析和相关测试、Bland Altman 图、交叉分类有效性和组内相关分析中进行了描述和测试。
总共邀请了 628 名成年人参加这项研究。其中，有 320 人参加，有 244 人至少完成了 4 次 24HDR。FFQ2020 和 24HDR 记录的食物组中位数摄入量、平均指数评分和估计营养素摄入量基本相似。两种评估之间的相关系数在食物组为 0.253 至 0.693，饮食指数为 0.520 至 0.614，能量和营养素为 0.340 至 0.629。组内相关系数表明，食物组、饮食指数评分和营养素的摄入量具有至少良好的可重复性。总体而言，Bland-Altman 图并未显示两种方法在任何评估中存在明显的系统差异。然而，有个别观察值位于 FFQ2020 和 24HDR 记录之间差异的上下 95%置信区间（CI）限之外。
总的来说，结果表明，FFQ2020 的相对有效性和可重复性在大型研究中的趋势分析和组间比较中是可以接受的，但扩展参考期将提高对较少食用食物的精确性。
