Comparison of single questions and brief questionnaire with longer validated food frequency questionnaire to assess adequate fruit and vegetable intake

• Published in: Nutrition (Burbank, Los Angeles County, Calif.) Los Angeles County, Calif.), 2015-07, Vol.31 (7), p.941-947
• Main Authors: Cook, Amelia, B.App.Sc./B.Sc.(Nutr), Roberts, Kia, B.Sc.(Nutr), O'Leary, Fiona, Ph.D, Allman-Farinelli, Margaret Anne, Ph.D., M.Phil.P.H
• Format: Article
• Language: English
• Subjects: Adult; Agreements; Behavior; Cardiovascular disease; Chronic illnesses; Diet - statistics & numerical data; Diet Surveys - methods; Feeding Behavior; Female; Food; Fruit; Fruits; Gastroenterology and Hepatology; Humans; Legumes; Male; Middle Aged; Nutrition research; Potatoes; Predictive Value of Tests; Questionnaires; Reproducibility of Results; Short food frequency questionnaire; Studies; Surveys and Questionnaires - standards; Test methods; Vegetables
• ISSN: 0899-9007; 1873-1244
• DOI: 10.1016/j.nut.2015.01.006

Abstract Objective The aim of this study was to determine if a single question (SQ) for fruit and a SQ or five-item questionnaire for vegetable consumption (VFQ) could replace a longer food frequency questionnaire (FFQ) to screen for inadequate versus adequate intakes in populations. Methods Participants (109) completed three test screeners: fruit SQ, vegetable SQ, and a five-item VFQ followed by the reference 74-item FFQ (version 2 of the Dietary Questionnaire for Epidemiological Studies [DQESv2]) including 13 fruit and 25 vegetable items. The five-item VFQ asked about intake of salad vegetables, cooked vegetables, white potatoes, legumes, and vegetable juice. The screeners were compared with the reference (DQESv2 FFQ) for sensitivity, specificity, and positive and negative predictive powers (PPV, NPV) to detect intakes of two or more servings of fruit and three or more servings of vegetables. Relative validity was examined using Bland-Altman statistics. Results The fruit SQ showed a PPV of 56% and an NPV of 83%. The PPV for the vegetable SQ was 30% and the NPV was 89%. For the five-item VFQ, the PPV was 39% and the NPV was 85%. Bland-Altman plots and linear regression equations showed that although the screener showed good agreement for fruit (unstandardized b1 coefficient = 0.04) for vegetable intake the difference between methods increased at higher intake levels (unstandardized b1 coefficients = −0.3 for the SQ, b1  = −0.6 for five-item VFQ). Conclusion The fruit SQ and the five-item VFQ are suitable replacements for longer FFQs to detect inadequate intake and assess population mean but not individual intakes.