Estimating plate-based model food proportions in adults living in Scotland using short dietary assessment questionnaires.

AIM: To determine whether a food frequency questionnaire (FFQ) and a short dietary assessment tool can be used to accurately estimate the Eatwell Guide proportions (a plate-based food model) of diets of adults living in Scotland. METHODS: The cross-sectional study was conducted as a follow up of the 2010 Scottish Health Survey (participants aged 18-65 years old). Proportions of the Eatwell Guide food groups (starchy carbohydrates, fruits and vegetables (F&V), dairy and alternatives, protein foods and oils and spreads) were calculated from the Scottish Health Survey Eating Habits Module (SHeS EHM), Scottish Collaborative Group FFQ (SCG FFQ) and a seven-day estimated food diary (reference method), and compared using the Aitchison method and Wilcoxon Signed-Rank Test. Bland-Altman analyses assessed mean difference and 95% limits of agreement between the methods for each food group. RESULTS: Ninety-six adults were included (mean (SD) age = 51.4 (11.1) years; body mass index = 27.1 (4.9) kg/m2 ; 58% female). The SCG FFQ scored a lower median Aitchison distance (1.47) than the SHeS EHM (1.99) (P < 0.001), showing greater agreement with the reference method (P < 0.001). Bland-Altman plots also showed better agreement for the SCG FFQ than the SHeS EHM. Poorest agreement was for starchy carbohydrates (both methods), protein foods (SHeS EHM) and dairy (SCG FFQ). CONCLUSIONS: The SCG FFQ could be used to estimate Eatwell Guide proportions and monitor compliance to the Eatwell Guide recommendations and could be improved with small changes. The SHeS EHM is less suitable, but additional questions on dairy foods, and oils and spreads would improve its ability to estimate the Eatwell Guide proportions.

Is misreporting of dietary intake by weighed food records or 24-hour recalls food specific?

BACKGROUND/OBJECTIVES: Healthy eating advice is informed, in part, by dietary surveys that rely on self-reported data. Misreporting of food intake may distort relationships between diet and health outcomes. This study directly quantified the food groups that were under-reported or over-reported in common dietary assessment techniques. SUBJECTS/METHODS: Food and drink consumption of 59 adults, with ad libitum access to a range of familiar foods, was objectively and covertly measured by investigators, and validated against independent measures of energy balance, while participants were resident in the Human Nutrition Unit of the Rowett Institute. Participants self-reported their diets using weighed dietary records (WDR) and multiple-pass 24-hr recalls over two periods of 3 days using a cross-over design. Foods and drinks were aggregated into 41 food groups. RESULTS: The mean daily weight of food and drinks reported was significantly lower than actually consumed; 3.3 kg (p = 0.004, 95% confidence interval (CI) = 3.07-3.55 kg) and 3.0 kg (p < 0.001, CI = 2.80-3.15 kg) for the WDR and 24-hr recall respectively, compared with 3.6 kg for the objective measure. Reported intakes were significantly lower than the objective measure for four and eight food groups (WDR and 24 h recall, respectively), and not significantly different for the remaining food groups. CONCLUSIONS: Although under-reporting was greater for some food groups than for others, 'healthy' foods were not over-reported and 'unhealthy' foods were not consistently under-reported. A better understanding of which foods tend to be misreported could lead to improvements in the methods of self-reported dietary intakes.