TABLE 1.
Summary of datasets used to develop and evaluate metrics1
| Diet methods and sample | Foods included in data, n | Reference period or no. of 24HRs | Portion size information | FFQ frequency options (if applicable) | Outcomes |
|---|---|---|---|---|---|
| Cross-sectional datasets | |||||
| Millennium Villages Project (10 Sub-Saharan African countries) (25) | |||||
| FFQ from 1624 rural NPNL WRA; separate instrument developed for each village | 92–161, depending on country | Past month | Nonquantitative (no portion size information) | Never, 1/mo, 2–3/mo, 1/wk, 2–3/wk, 4–6/wk, 1/d, ≥2/d | Nutrient intake and adequacy, BMI, MUAC, hemoglobin |
| Anemia etiology in Ethiopia study (26) | |||||
| FFQ from 1604 mostly rural NPNL WRA2 | 454 | Past week | Quantitative: 7 food item–specific portion sizes assessed for each food | Never, 1/wk, 2–4/wk, 5–6/wk, 1/d, 2–3/d, 4–5/d, ≥6/d | Nutrient intake and adequacy, BMI, MUAC, hemoglobin, ferritin, serum folate, serum vitamin B12, blood pressure |
| 24HR from 1593 mostly rural NPNL WRA2 | 113 | 1 24HR, and 2nd in subset of participants | Multiple-pass probe incorporating information on no. of meals at which each food was consumed, no. of servings of each food consumed at each meal, and average portion size of each food | NA | Same as above |
| 2010–2012 China National Nutrition and Health Survey (27) | |||||
| 24HR from 15,173 urban and rural NPNL WRA | 1615 | 3 consecutive d (2 wkd and 1 wkend) | Quantitative: estimated g consumed/last 24 h each d of the 3 d | NA | Nutrient intake and adequacy, BMI, waist circumference, hemoglobin, glucose, HDL and total cholesterol, triglycerides, blood pressure, metabolic syndrome |
| Indian Migration Study and Andhra Pradesh Children and Parents Study (28, 29) | |||||
| FFQ from 3065 mostly rural NP WRA 3 | 184 | Past year | Portion size estimates with quantitative: standard household utensils (e.g., tablespoon, ladle, and bowl), data on no. of portion sizes consumed also collected | Never, yearly, monthly, weekly, daily | Nutrient intake and adequacy, BMI, hemoglobin, HDL and total cholesterol, blood pressure |
| 2012 and 2016 Mexican National Surveys of Health and Nutrition (30, 31) | |||||
| FFQ from 4975 urban and rural NPNL WRA2 | 140 | Past week | Quantitative: 2–3 portion sizes offered for each food, data on no. of portion sizes consumed also collected | Never, 1/wk, 2–4/wk, 5–6/wk, 1/d, 2–3/d, 4–5/d, ≥6/d | Nutrient intake and adequacy, BMI, waist circumference, hemoglobin, ferritin, serum folate, serum vitamin B12, glucose, insulin, LDL cholesterol, HDL cholesterol, total cholesterol, triglycerides, metabolic syndrome |
| 24HR from 2545 urban and rural NPNL WRA2 | 544 | 1 24HR, 2nd in subset of participants | Multiple 5-pass probe incorporating weighed amounts or common household measurement implements | NA | Same as above |
| Cohort datasets | |||||
| Mexican Teachers Cohort (32) | |||||
| FFQ from 8967 urban and rural NPNL WRA | 125 | Past year | Semiquantitative: standard portion size or commonly used unit indicated | Never, ≤1/mo, 2–3/mo, 1/wk, 2–4/wk, 5–6/wk, 1/d, 2–3/d, 4–5/d, ≥6/d | Weight change, waist circumference change |
| US Nurses' Health Study II (33) | |||||
| FFQ from 56,321 urban and rural NP WRA3 | 135 | Past year | Semiquantitative: standard portion size or commonly used unit indicated | Never or <1/mo, 1–3/mo, 1/wk, 2–4/wk, 5–6/wk, 1/d, 2–3/d, 4–5/d, ≥6/d | Weight change, incident type 2 diabetes |
In cross-sectional datasets, sample size corresponds to the number of participants with dietary data (for some outcomes, available sample size was smaller; refer to (25–33) for more details. MUAC, mid–upper arm circumference; NA, not applicable; NP, nonpregnant; NL, nonlactating; WRA, women of reproductive age; 24HR, 24-hour recall.
FFQ and 24HR data from the Anemia Etiology in Ethiopia Study were collected from the same sample. FFQ and 24HR data from the 2012 and 2016 Mexican National Surveys of Health and Nutrition were collected from separate samples.
The Indian Migration Study and Andhra Pradesh Children and Parents Study population consists of NP WRA (lactation was not ascertained). In analysis of the Nurses’ Health Study II, women were classified as NP (lactation was not ascertained), but 2-y time periods during which a pregnancy was reported were excluded from analysis to limit the influence of lactation.