Validity and reproducibility of a web-based dietary assessment tool: a cross-sectional study in an adult Danish population

Abstract

This repeated cross-sectional study assessed the validity and reproducibility of the myfood24^® dietary assessment tool against dietary intake biomarkers in healthy Danish adults. The study included 71 healthy adults (14/57 m/f), aged 53.2 ± 9.1 years with an average BMI of 26.1 ± 0.3 kg/m². Participants were instructed to complete seven-day weighed food records using myfood24^® at baseline and 4 ± 1 weeks thereafter. Estimated mean dietary intake was compared with objective measures of energy metabolism and selected dietary intake biomarkers in fasting blood (folate) and in 24-hour urine (urea, potassium). Resting energy expenditure was measured by indirect calorimetry. Application of the Goldberg cut-off classified 87% (n = 62) of participants as acceptable reporters. A strong Spearman's rank correlation was observed between total folate intake and serum folate (ρ = 0.62). Acceptable correlations were noted for serum folate (ρ = 0.49) and urinary potassium excretion (ρ = 0.44) with estimated and measured protein intake (ρ = 0.45); energy intake and total energy expenditure (ρ = 0.38); potassium intake and potassium excretion (ρ = 0.42); and estimated fruit and vegetable intake. Reproducibility analysis revealed strong correlations (ρ ≥ 0.50) across most nutrients and food groups, except for fish and vitamin D (ρ = 0.30 and ρ = 0.26, respectively). Notably, reproducibility for folate and total vegetable intake exhibited the highest correlations (ρ = 0.84 and ρ = 0.78, respectively). In conclusion, while some limitations exist, myfood24^® remains a useful tool for ranking individuals by intake, particularly in studies focusing on relative comparisons.

Keywords: 7-day weighed food record; Biomarker; Folate: myfood24®; Protein.

Fig. 1.

Study design overview. The figure outlines the study timeline. During the inclusion period, participants attended an information meeting, followed by screening for eligibility. At Visit 1 (week 0), measurements included height and weight (), fasting blood samples (), 24-hour urine sample hand-in (), and indirect calorimetry (). Participants completed a 7-day WFR 1 week prior to both visits (), and handed these in () on both visits. Anthropometric measurements were repeated at Visit 2 (week 5) after a 4 ± 1 week interval.

Fig. 2.

VALID flow chart. The flowchart illustrates the participant recruitment process. Of 244 subjects initially expressing interest, 163 were included in the pre-screening after 81 were excluded based on the exclusion criteria. Of the 163 individuals included in the pre-screening, 96 proceeded to the screening phase, and 67 were excluded for reasons such as being unreachable, personal reasons, or not meeting specific criteria. A total of 71 participants completed Visit 1. Two additional participants were excluded after Visit 1, leaving 69 participants who completed both visits.

Fig. 3.

(a) Bland-Altman plot indicating the agreement between the EI based on the 7-day WFR versus the measured TEE among healthy adults (n = 71). The middle red line represents the mean difference (–45.5 kcal/d), while the upper and lower dashed lines indicate. the limits of agreement (1227 kcal/d, –1318 kcal/d), corresponding to ±1.96 SD. (b) Bland-Altman plot showing the agreement between the protein intake based on 7-day WFR and the ingested protein as calculated from the 24-hour urinary excretion of N, plotted against the mean of the two measures. The solid line indicates the mean difference between the 2 measures (5 g/d), while the upper and lower dashed lines indicate the limits of agreement (59.3 g/d, –50.2 g/d), corresponding to ±1.96 SD.

Fig. 4.

(a) Bland-Altman plot of agreement between total vegetable intake estimated from WFR1 and WFR2. The mean difference is 40.1 g/d, with limits of agreement ranging from –200 g/d to 280 g/d. (b) Bland-Altman plot of agreement between folate intake estimated from WFR1 and WFR2. The mean difference is 37 μg/d, with limits of agreement ranging from –144 μg/d to 218 μg/d. (c) Bland-Altman plot of agreement between potassium intake estimated from WFR1 and WFR2. The mean difference is 190 mg/d, with limits of agreement ranging from –935 mg/d to 1316 mg/d. (d) Bland-Altman plot of agreement between energy intake estimated from WFR1 and WFR2. The mean difference is 50.4 kcal/d, with limits of agreement ranging from –2476 kcal/d to 2902 kcal/d.