ارزیابی روایی نسبی و پایایی دریافت گروه‌های غذایی با استفاده از پرسشنامه بسامد خوراک: مطالعه قند و لیپید تهران

    Background & Aim: In epidemiologic studies, food frequency questionnaire (FFQ) is usually the most suitable dietary assessment tool on a long-term basis. Hence assessing its validity and reproducibility  is conceptually important to determine the accurate relationship between diet and diseases. The current study was done to examine the relative validity and reproducibility of food groups of the semi-quantitative FFQ, developed for the Tehran Lipid and Glucose Study (TLGS).Patients and Method: This study investigated 132 subjects (61 men and 71 women), aged ≥ 20 years who completed a 168-item FFQ twice (FFQ1-FFQ2), with a 14-month interval. Dietary data were collected monthly by means of twelve 24-hour dietary recalls. The Results: Means (SD) for age and body mass index of subjects were 35.5 (±16.8) years and 25.5 (±5.2) kg/m2 respectively. In order to assess validity of the FFQ, age-adjusted and deattenuated Spearman correlation coefficients were performed. Coefficients ranged between 0.10 (solid fat) and 0.77 (refined sugar) in men (mean=0.44), and between 0.12 (snacks) and 0.79 (refined sugar) in women (mean= 0.42). Mean age-and energy-adjusted intraclass correlation coefficients, assessing reproducibility of the FFQ, were 0.51 in men, highest for tea (0.91), and 0.59 in women, highest for refined sugar (0.74). Mean percentages of disagreement were 7.3 in men and 9.3 in women, the highest percentage being observed for snacks and desserts in men and for tea and coffee in women.Conclusion: The FFQ designed for TLGS was found to be reliable and valid for several food groups.first DRwas collected one month after FFQ1 and the last DRwas collected one month before FFQ2. Food items were categorized into 17 food groups based on the previous studies. The residual method was used to calculate energy and age-adjusted food group intakes. Spearman and deattenuated correlation coefficients and Chi-square test based on tertiles of food group intakes were used to determine the relative validity between the second FFQ and 24-hour DRs, and intraclass correlation coefficients were calculated between FFQ1 and FFQ2 to assess the reproducibility of the FFQ.