Defining the stratification and scope of the survey
## Stratification: The Steering committee should decide on the survey scope and level of stratification. For a survey of national scope, some level of stratification will be useful to obtain representative data for different geographical, administrative or programme-related areas. For example, it may be known that populations in different parts of the country have different vulnerabilities to certain micronutrient deficiencies, or different levels of access to a specific intervention. In that case, it would be useful to have data for each situation so that associated factors and appropriate follow-up can be determined. In countries with decentralized governments, it may be particularly important to consider stratification based on policy and administrative areas where decisions are made.
If the survey aims to assess the impact of an intervention that is implemented in only a single region of the country, then all survey planning decisions and expected outcomes need to relate to that one region. In that case, there is usually no, or only a small degree of stratification, perhaps limited to urban and rural strata.
The overall sample size, fieldwork, and logistical costs are greatly influenced by the number of strata, and it is usually necessary to find a balance between the desired number of strata and the available budget. More information about factors to consider with regard to stratification can be found in Module 4: Survey design.
Micronutrients to assess:
The micronutrient deficiencies with the largest known public health burdens concern iron, vitamin A, iodine, zinc, vitamin D, vitamin B12 and folate.
Here are some key issues to consider when deciding which micronutrients and associated indicators to assess in a survey:
Which populations groups are of interest? Usually women and children are the most highly affected populations.
Which nutrients and associated indicators should be assessed, and what is the feasibility of their assessment? Questions 1 and 2 are discussed in more detail on the section below on “Population Groups to Assess”.
What relevant interventions are currently in place or planned? Consider which micronutrient status data are needed at what level of precision to understand impact and accountability, and to adjust advocacy or policies. For example:
- Can intervention coverage be used instead of a biomarker, where an intervention has previously shown impact?
- What level of confidence in (the precision of) an estimate is needed to show a “real” change after implementation of an intervention?
Are there other available sources of data? What data on micronutrient status or effective implementation of proven interventions are already available, how recent are they, what is the representativeness of the data, and what is the quality of the data? If recent, representative, high-quality data on the status of a specific micronutrient among the population group of interest are available from a different source, there may be no need to include this indicator in the planned micronutrient survey.
What additional data would be useful to understand and improve implementation? Even if data are available from other sources, there might still be a need to assess a particular micronutrient if, for example:
- The status of a different population group is required;
- Existing data do not include programme-related indicators which, together with micronutrient status data would shed light on the relationship between status and implementation of a specific intervention.
- Likewise, if recent indicators of intervention coverage are available, it may not be necessary to include the same indicators in the survey unless, as above, it is important to examine the relationship between status and intervention.
Are there other surveys planned that include information about micronutrient status? For example, Demographic and Health Surveys and Malaria Indicator Surveys often collect data on anaemia and may be able to include coverage of micronutrient supplements or fortified foods. If a micronutrient survey intends to assess only one or two micronutrients, serious consideration should be given to incorporating this into another planned survey. Resources should be invested in a separate micronutrient survey only if there are strong arguments about why the required data cannot be obtained through other channels.
You can find more detail on all these factors in Modules 2-16.
Population Groups to Assess
Priority should be given to those groups that are most vulnerable to one or more micronutrient deficiencies or for whom deficiencies may lead to measurable negative health consequences. Consideration should also be given to a specific population group that is the main intended beneficiary of an existing or planned micronutrient intervention. Children 6–59 months of age and women of reproductive age are the groups most frequently assessed in micronutrient surveys. This is due to their vulnerability to deficiencies and to the common design of interventions to address the needs of these groups.
A survey will be more efficient when it covers population groups vulnerable to multiple micronutrient deficiencies. When selecting the population groups to assess, ensure that:
There are recommended cutoff values to define different categories of status (such as deficient, insufficient and normal) for the micronutrients of interest. For example, for ferritin, there is no internationally recognized cutoff value to define deficiency among pregnant women, thus pregnant women would not be a suitable group in which to assess this indicator.
A sufficient sample size can be accessed without having to visit a large number of households. For example, the population percentage of women of reproductive age who are pregnant may be around 6%. The number of households you would have to visit to obtain a representative sample of pregnant women for each survey stratum may be too high to be feasible. At the same time, if the scope of the survey includes data on pregnant and lactating women, an adequate sample size must be ensured. Sometimes specimens are collected from all consenting pregnant and lactating women in all survey households, with the expectation that the final number of samples may be sufficient to give a reasonable estimate of status at the national level.
The following sections provide considerations for including different population groups.
Children under six months of age
Infants (0–6 months of age) are often included as the focus for questions that address breastfeeding practices. Members of this age group are not usually included for collection of biological specimens, because refusal rates are likely to be high. In addition, there are no internationally recognized cutoff values in this age group to define categories of micronutrient status for many micronutrients.
Preschool children (6-59 months of age)
Preschool children are particularly vulnerable to the consequences of micronutrient deficiency and are therefore frequently included for assessing multiple micronutrients. Priority indicators for this age group in low- and middle-income countries include the status of iron, vitamin A and anaemia, along with markers of inflammation that may affect interpretation of the results. This group is also vulnerable to deficiencies of zinc and vitamins B12 and D, however, these micronutrients are less frequently assessed.
School-age children and adolescents
The exact definition of school-age varies from country to country but is most commonly defined as from 5 years up to but not including 15 years of age (thus 5.0–14.9 years of age). If this group is included in a household-based survey, the survey should take place when children are more likely to be at home, such as during school holidays, on weekends, or after school hours. Alternatively, school-age children can be assessed in a school-based survey, however, the rate of school attendance should be considered, to account for any potential bias in the results. A school attendance rate of around 75% or higher is usually required to obtain what may be considered as representative data through these means. Micronutrient and related health status indicators most often assessed among this group are iodine, zinc, iron, and anaemia, along with markers of inflammation that may affect interpretation of the results. The category of adolescents is usually defined as from 10 to 19 years of age. The focus for this group tends to be females, who are most vulnerable to anaemia due to menstruation and reproductive health risks including early pregnancy.
Women of reproductive age
Women of reproductive, or child-bearing age, are also vulnerable to vitamin and mineral deficiencies. They are usually divided into three subgroups: non-pregnant/non-lactating, pregnant, and lactating. The survey protocol should clearly indicate whether to collect specimens from women who are of unknown pregnancy or lactating status, and how to record these women in the household record. Micronutrient and related health status assessments for pregnant and lactating women usually include iodine status (urinary iodine concentration [UIC] and anaemia [haemoglobin]). Pregnant women comprise a priority group for assessing iodine status, even if only a nationally representative sample is possible. Because their nutritional status can influence the developing fetus, women of reproductive age are often the focus of nutrition-related interventions and are frequently included in micronutrient surveys. Micronutrient and related health status indicators generally assessed among this group include iron, vitamin A, folate, iodine, vitamin B12, and anaemia, along with markers of inflammation that may affect interpretation of the results. See the “Biomarker Selection and Specimen Handling” provides additional details and considerations for assessing vitamin A and iron status.
Micronutrient and related health status assessments for pregnant and lactating women, where included, usually include iodine status (urinary iodine concentration [UIC]) and anemia (hemoglobin [Hb] adjusted for altitude and smoking), defined using internationally recognized cutoffs exist for these indicators for both groups. Data for WRA need to be analyzed and presented separately for each of the three categories above —non-pregnant, non-lactating women; non-pregnant, lactating women; and pregnant women —stating the defined cutoffs to classify the status for each.
Men can also have an inadequate micronutrient status. However, they tend to be less severely affected than children 6–59 months of age and women of reproductive age. Therefore, they are less frequently included for assessment of micronutrient and related health status biomarkers. In some cases, men may be selected from a subsample of households to provide sufficient numbers to understand the overall micronutrient status among this group.
In some settings, men are included in the assessment of hemoglobin concentration to help understand the role that iron deficiency plays in anemia among women and children. If the prevalence of anemia is high among men, as well as among women and children, then causes of anemia other than iron deficiency, such as malaria or other parasitic infections, are likely to be prevalent.
Even if micronutrient status biomarkers are not collected, men are sometimes included in the interview process to assess their level of knowledge, attitudes, and practices (KAP) concerning how to prevent micronutrient deficiencies among household members.
You can find additional information on the selection of micronutrients by population group, choice of biomarker, type of sample required, and related biomarker cutoffs for different population groups in Module 2: Indicators of programme coverage, specimen selection, management and analysis, and Module 3: Biomarker selection and specimen handling.