Validity and reliability of food frequency questionnaire used in age-related eye disease studies in Turkish adults

Abstract

The present study aimed to perform the validity and reliability study of the Food Frequency Questionnaire (FFQ) on the frequency of foods rich in antioxidant nutrients and used in Age-Related Eye Diseases (AREDs). In the first interview of the study, the first application of FFQ was carried out, and blank forms of Dietary Records (DRs) were given. For the validity of the FFQ, a total of 12 d (3 days * 4 weeks) of DR were taken. For the reliability of the FFQ, a test-retest application was made with an interval of 4 weeks. The daily intake means of antioxidant nutrients, omega 3 and total antioxidant capacity data obtained from both the FFQ and DR were calculated, and the concordance between the two methods was evaluated with the Pearson Correlation Coefficient (PCC) and Bland-Altman graphs. The present study was carried out at Ege University İzmir/Turkey, Department of Ophthalmology, Retina Unit. The study was conducted with individuals aged ≥50 years who suffered from Age-Related Macular Degeneration (n 100, 72⋅0 ± 8⋅03 years). The values obtained from the test-retest applications for FFQ reliability were the same. The nutrient intake means obtained from the FFQ were similar or significantly higher than DR (P < 0⋅05). In the Bland-Altman graphical approach, the nutrient data were within the limits of agreement, and the PCCs between the two methods were moderately related. Considered together, this FFQ is a suitable tool for determining the dietary intakes of antioxidant nutrients in the Turkish population.

Keywords: AREDS FFQ; Age-related macular degeneration; Antioxidant; Food frequency; Validity and reliability.

Fig. 1.

Study design: The original form of AREDS FFQ was obtained from NEI and translation and adaptation studies were performed. A pilot study was conducted to determine how long the survey would take on average, and a practice was created on how to ask about food consumption. In the first interview, the general information of the individuals was questioned with a structured questionnaire and anthropometric measurements of the individuals were taken. In addition, the first application of AREDS FFQ was carried out and it was explained how to get DR (3 days * 4 weeks). In the last week, AREDS FFQ was reapplied and the study was completed.

Fig. 2.

Bland–Altman graphs of omega 3, antioxidant nutrients and total antioxidant capacity: This figure is intended to determine the fit between the two methods (AREDS FFQ and DR). (a) Omega 3 (mg/gün): The mean of Bland–Altman difference is 0⋅418, and the limits of agreement are between (−0⋅548) and (1⋅385). (b) Vitamin A (mcg/gün): The mean Bland–Altman difference is 450 427, and the fit limits are between (−1 085 212) and (1 986 066). (c) T-Carotene (mg/gün): The average of the Bland–Altman difference is 1⋅312, and the limits of agreement are between (−2⋅945) and (5⋅570). (d) Lutein + Zeaksantin (mcg/gün): The mean Bland–Altman difference is 864 138, and the fit limits are between (−1 866 07) and (3 594 345). (e) Vitamin E (mg/gün): The mean Bland–Altman difference is 41⋅533, and the fit limits are between (10⋅215) and (72⋅851). (f) Vitamin C (mg/gün): The mean of the Bland–Altman difference is 43 585, and the agreement limits are between (−76 381) and (163 551). (g) Zinc (mg/gün): The mean of the Bland–Altman difference is 1322, and the agreement limits are between (−2⋅002) and (4646). (h) T-Antioxidant (mg/gün): The mean of the Bland–Altman difference is 7⋅838, and the agreement limits are between (−2773) and (18 451). The Bland–Altman graphical approaches showed that 95 % of the data is within the limits of agreement, and the scattering is concentrated around the mean (T-carotene, total carotene; T-antioxidants, total antioxidants).