Minimum Dietary Diversity - Women

Outcome

% of women of reproductive age (15–49 years) who consumed foods or beverages from at least 5 out of 10 defined food groups during the previous day or night

MDD-W assesses the proportion of women aged 15–49 years who consumed foods or beverages from at least five out of ten defined food groups during the previous day or night. At population level, a higher prevalence of MDD-W indicates a greater likelihood of better micronutrient adequacy. MDD-W measures dietary diversity, one important dimension of diet quality; it does not assess overall diet quality or establish the dietary adequacy of an individual woman.

Determine the indicator's value through individual interviews with a representative sample of women aged 15–49 years, including pregnant women. Record pregnancy status. The steps below describe the open-recall method:

 

1) Ask the respondent to recall chronologically all foods and beverages she consumed during the previous day and night, including meals, snacks and drinks consumed at home or elsewhere. Record them in the Recording Meals Form or an equivalent open-recall grid.

  

2) Probe neutrally for the ingredients of mixed dishes and beverages (e.g. whether porridge contained milk).

  

3) Ask neutrally whether the respondent ate or drank anything else between or outside meals, without suggesting specific foods.

  

4) After completing the open recall, classify the foods and beverages reported by the respondent into the MDD-W food groups. Do not conduct a second pass asking about specific foods or food groups that were not mentioned during the open recall, as this can induce reporting bias and overestimate consumption.

  

5) For each respondent, count the number of the 10 defined MDD-W food groups consumed during the previous day and night to obtain a food-group diversity score ranging from 0 to 10.

   

6) To calculate the indicator's value, divide the number of women aged 15–49 years who consumed foods or beverages from at least 5 of the 10 defined food groups by the total number of women aged 15–49 years surveyed. Multiply the result by 100.

This data can be disaggregated by age group and pregnancy status, and, if relevant and feasible, also by wealth quintile and area of residence.

1) Pregnancy status: MDD-W can be collected among all women aged 15–49 years, including pregnant women, and pregnancy status should be recorded. The 2021 FAO guide noted that the ≥5-of-10 threshold had originally been validated among non-pregnant women, but later research by Verger et al. (2024), cited by FAO, found the same threshold also performed well among pregnant women. However, the current SDG 2.2.4 women’s series still uses non-pregnant women. IndiKit guidance includes all women aged 15–49 years, but results should be disaggregated by pregnancy status when comparability with global MDD-W or SDG estimates is important.

2) FAO recognises two methods for collecting MDD-W data: open recall and list-based recall. In open recall, the respondent describes everything she ate and drank during the previous day and night, and the enumerator later classifies these items into the 10 MDD-W food groups; neutral probing can be used, but specific unmentioned foods should not be suggested. In the list-based method, the enumerator asks directly about specific, locally relevant foods and beverages and records yes/no responses using an adapted questionnaire. Both methods can over-report MDD-W, with greater over-reporting observed for the list-based method in FAO validation work, so use the same method when comparing results over time. Country-adapted Diet Quality Questionnaires (DQQs) from the Global Diet Quality Project provide a standardised list-based option.

3) Seasonality and survey timing: Dietary diversity can vary by season and day of the week. For comparisons over time, collect data during the same season or time of year where possible and ensure that survey timing is comparable. Extended fasting periods such as Ramadan or Lent should also be considered when scheduling surveys or interpreting changes over time.

4) Unusual consumption: Following the updated FAO (2021) MDD-W guidance, do not exclude an individual respondent merely because the previous day involved unusual intake, such as a wedding, feast, celebration or regular fast day; such variation forms part of population-level dietary intake. However, if an unusual event affected the whole community or a large proportion of it and substantially altered food consumption, avoid collecting data during or immediately after that event where possible.

5) According to FAO, MDD-W has replaced the Women’s Dietary Diversity Score (WDDS). WDDS was an earlier indicator that counted how many of nine food groups a woman consumed during the previous 24 hours and was commonly reported as an average score. MDD-W instead uses 10 food groups and reports the proportion of women reaching the ≥5-of-10 threshold. This threshold is intended for population-level interpretation: women reaching it are more likely to have higher micronutrient adequacy, but it is not an individual dietary requirement and there is no universal MDD-W prevalence cut-off defining an acceptable, poor or severe population situation.

 

6) In addition to learning the proportion of women who consumed foods from 5 or more food groups, MDD-W data can also be used to assess:

     i) mean food-group diversity score (0–10)

     ii) proportion of women who consumed any specific food group, such as animal source foods 

  

7) Train data collectors thoroughly on the selected collection method, the locally adapted food list, mixed dishes, food-group classification, foods usually consumed in quantities below 15 g, neutral probing and interview practice. Include pre-testing or field practice where feasible. The duration should be adapted to the method and enumerators' experience; FAO indicates that dedicated MDD-W training generally requires at least one day.

  

8) Following FAO guidance, foods or beverages usually consumed in very small quantities (less than about 15 g, approximately one tablespoon for many foods) should not count towards MDD-W. Respondents do not need to estimate quantities during the interview. Instead, when adapting the questionnaire, identify foods and ingredients normally consumed only in such small amounts and exclude them from scoring or classify them as condiments and seasonings, as appropriate.

Also consider using the following indicators:

• Minimum Dietary Diversity - Children (complements women’s dietary diversity with the corresponding dietary diversity outcome for children aged 6–23 months)

• Household Dietary Diversity Score (helps distinguish household-level food access and dietary variety from women’s population-level dietary diversity)

• Awareness of Appropriate Diet (provides an enabling knowledge measure where nutrition behaviour-change activities are expected to contribute to improved diets)