How old is a 9 month old? It seems like an easy question with an obvious answer – um, 9 months? But it is actually a bit of a challenge and requires some precise language, and specific coding when analysing child health data, such as in the baseline surveys from our project “Saving lives of mothers and children in Nepal and Vietnam”.

We first need to collect child age data, which itself a challenge. In Nepal, a child will be considered 1 year old on the day of birth, 1 year old on their first year birthday (365 days after birth), and then 2 years old the day after the first year birthday (366 days after birth), and so on. This is manageable as long as we all understand that this is what the mothers mean when we ask them the child’s age. But we get more accurate data when we ask about the child’s date of birth and then count the number of days from birth to the interview date. This should be straightforward too, but Nepal has a different calendar and today is the 21 of Ashoj in the year 2073. We can manage this too, although standard date calculators (such as the one I use in SAS) which are based on our Gregorian Calendar don’t work, so we need to do a few extra calculations to figure out how many days have passed from the child’s birth to the interview date. But we manage that.

Then we need to convert the number of days to the number of months. We divide days by 30.5 days per month to get age in months – it isn’t exact and gets thrown off by leap years, but it is close enough for our purposes. So now we have child age in months, we are home free right? Not quite.

For some health indicators, such as exclusive breastfeeding rates, we want to include all children less than 6 months of age, or for young child feeding we group together those 6 to 8 months old and those 9 to 24 months old. For some indicators, such as number of feedings per day, we want to know the average for children of a specific age. All these are treated differently.

For children less than 6 months of age, we include children 5.99 months old or younger. Children 6 to 8 months are age 6 months to 8.99 months. However, for meal frequency of children “9 months old” the best case we would include only those children who were interviewed on their 9 month birthday – but we would then have a tiny sample size. So we include children 8.5 months old to 9.49 months old and take the average of that group. Although if our sample is small, we may need to expand the age range to, say, 7.5 to 10.49 months old.

Oh my, I have already written 400 words on this and I am afraid I have bored you to tears, but for us data nerds it is a regular issue we confront. It isn’t just me! A few years ago while analysing nutrient intake data I had to figure out the exact age boundaries for the Dietary Reference Intakes (the nutrition requirements used in Canada, the US and elsewhere). The age groups they use are 0-6 months, 7-12 months, 1-3 years and 4-8 years. But the cutoffs for these ages are not explicitly defined. How should we handle a child 6.5 months old? From reading the text and trying to understand the logic, I believe the implied cutoffs are:

• 0-6 months -- from birth to 6.00 months
• 7-12 months -- from 6.01 months to 12.00 months
• 1-3 years -- from 12.01 months to 47.99 months
• 4-8 years -- from 48.00 months to 96.000 months

It is not internally consistent. Whereas 12 months is the exact upper limit for 7-12 months, the upper limit for 3 years is actually 3.99 years (that is, anything up to but not equal to 48 months of age). It is all very confusing!

On one hand it is ok, because we don’t know the nutrient requirements exactly - we can’t say they change by some amount from one day to the next. There is nothing magical that happens to a child on the ninth month birthday, and whether a child is 8 months and three weeks old or 9 months and 1 week old makes no biological difference. But still, we have to apply these guidelines somehow, so we need to be precise in our definitions. On one hand it is trivial, but on the other hand it is critical that we have a common understanding.