Validation of the FRESH Austin food frequency questionnaire using multiple 24-h dietary recalls

• Published in: Public health nutrition 2022-06, Vol.25 (6), p.1586-1594
• Main Authors: Jovanovic, Christine ES, Whitefield, Jacob, Hoelscher, Deanna M, Chen, Boajiang, Ranjit, Nalini, van den Berg, Alexandra E
• Format: Article
• Language: English
• Subjects: Data collection; Diet; Diet Records; Diet Surveys; Estimates; Evaluation; Female; Food; Fruits; Health care; Humans; Income; Low income areas; Low income groups; Male; Mental Recall; Nutrition research; Nutritional Epidemiology; Public health; Questionnaires; Recruitment; Reproducibility of Results; Research Paper; Surveys and Questionnaires; Validation studies; Validity; Vegetables
• ISSN: 1368-9800; 1475-2727
• DOI: 10.1017/S1368980021002214

The purpose of the current study was to examine the validity of an FFQ utilised in the Food Retail: Evaluating Strategies for a Healthy Austin (FRESH Austin) study, designed to evaluate changes in the consumption of fruits and vegetables (FV) in diverse low-income communities in Austin, TX. The FRESH Austin FFQ was validated against three 24-h dietary recalls (24hDR). All dietary assessments were administered (in-person or by telephone) by trained investigators. Recruitment was conducted at sites within the geographic areas targeted in the FRESH Austin recruitment. People at a community health clinic, a local health centre and a YMCA within the intervention area were approached by trained and certified data collectors, and invited to participate. Among fifty-six participants, 83 % were female, 46 % were non-White, 24 % had income < $25 K/year and 30 % spoke only/mostly Spanish at home. The FFQ and average of three 24hDR produce similar estimates of average total servings/d across FV (6·68 and 6·40 servings/d, respectively). Correlations produced measures from 0·01 for 'Potatoes' and 0·59 for 'Other Vegetables'. Mean absolute percentage errors values were small for all FV, suggesting the variance of the error estimates was also small. Bland-Altman plots indicate acceptable levels of agreement between the two methods. These outcomes indicate that the FRESH FFQ is a valid instrument for assessing FV consumption. The validation of the FRESH Austin FFQ provides important insights for evaluating community-based efforts to increase FV consumption in diverse populations.