Dear Miriam

I will mostly speak from my experience with SMART nutrition surveys.In determining the sample size,the proportion of children below 5 years (0-59m) in the population is considered.However,given that the anthropometric assessment will almost always exclude children below 6 months (0-5m), the final sample size gives the number of children 6-59m required and the number of households which is expected to provide the given number of children as this is a primarily an anthropometric assessment.I hope this helps.

Blessing

Dear Ms. Chang:

This question comes up more and more frequently in planning nutrition assessment surveys. As you mention, young infants used to be routinely excluded from anthropometric measurement because they were thought to be protected from acute malnutrition by breastfeeding and because they were difficult to weigh and measure. However, it began to be appreciated some years ago that they are not in fact immune from wasting. In general, I agree with the UNHCR recommendations you cite, with the following commentary:

Theoretically, the results of surveys for which the target group is children 6-59 months should not be compared to the results of surveys for which the target group is children 0-59 months because they are assessing different populations. However, children less than 6 months of age represent only a bit more than 10% of all children less than 5 years of age, so they would have to be very different from older children to have a major influence on the overall prevalence of undernutrition in children less than 5 years of age. As a result, some comparison, with the appropriate caveats, is probably OK.

It is more difficult to get accurate measurements on such young children. I have never seen infant balance scales taken into the field and used in household-based surveys; I doubt if they would stand the jostling and mistreatment inherent in such activities. Although the UniScale and other electronic scales have poorer precision, even +/- 100 grams still represents only +/- 3% of the measurement of the weight of a 3.5 kilogram child. Length measurements can also pose greater difficulty, so survey teams must be specifically trained and practice handling infants. Perhaps the biggest problem is that mothers really don't like strangers manhandling their young infants. Nonetheless, I have supervised surveys in which reasonable measurements were taken on children less than 6 months of age.

Regarding guidelines, standardization and uniformity should not be ends in and of themselves. I would not recommend routine inclusion of children less than 6 months of age in nutrition assessment surveys, but I think such a decision should be situation-specific. Survey planners must take into account what questions the survey must answer, the circumstances of the target population, and any available data to make an informed judgement regarding whether inclusion of young infants is worth the difficulty. Of course, most decisions in survey planning should take these factors into account.

If you wish to make conclusions about children less than 6 months of age, you will probably need to compensate somehow for their relative rarity in household-based samples. You could oversample this age group. Alternately, you could aim for an overall estimate in this age group without subgroup analysis by location, sex, or other characteristics in order to analyze the maximum sample size. Regardless, you should calculate a separate sample size for these children before embarking on data collection if you really need a precise estimate.

Dear Mirriam,
The 2006 WHO GS should be used for the nutritional assessment of children 0-59 months, however, infants 0 - 6 months anthropometric measurements are generally not measured in emergencies. Wasting prevalence is generally analyzed on 6-59 months, yet stunting is analyzed on 0-59 months if this group is included.

There is already documented evidence on the challenges around obtaining accurate anthropometric (Weight and Height) measurements for 0-6m. There is a general assumption that infants less than 6 months are breastfeeding and are less likely to be nutritionally at risk. Low birth weight and the likelihood of being classified with higher degree of wasting as well as cultural/social beliefs surrounding handling under 6 months during field surveys pose similar challenges.

Some materials on the subject can accessed from;

http://www.ennonline.net/module6

http://www.unscn.org/layout/modules/htp/pdf/mod6_measuring_malnutrition-technical_notes.pdf

My previous field experience shows that the national/large scale nutrition surveys provide rigorous training to the enumerators spanning 2- 4 months .Such training is never feasible in emergency contexts targeting relatively smaller administrative regions. Therefore, other implications involved in sampling 0-59m would be relatively more financial resources and time allocation. It is not possible to draw comparisons on survey results obtained from two different target groups.

It would be more appropriate to develop a comprehensive guideline for 0-59 months while being clear on the challenges involved in getting accurate measurements for 0-6 months infants. As already mentioned above, the choice of your target should be context specific and objective driven. A number of guidelines and materials are already in existence to help you come up with one.

You can access some guidance notes from SMART Survey Planning Tools