23 June 2014

Indian Nutrition Data: Too Little or Too Much?

People like me complain that the Indian Government does not collect enough data on nutrition outcomes.

The 3rd  National Family Health Survey (NFHS-3) was conducted in 2005-6.  The 4th is labelled as 2014-15 and given its large size (about 568,000 households compared to 109,000 households in 2005-6) we can count on it coming out at the earliest in 2016 and quite possibly 2017.

So, we have some time to wait for new NFHS data.  

But last week, without fanfare, some new Government data on nutrition were released  -- the 4th District Level Household Survey (DLHS IV 2012-2013).  Both NFHS and DLHS are run by the Ministry of Health and Family Welfare.  NFHS-4 will be representative at the District Level for all 640 districts.

The DLHS IV is also representative at district level, but does not cover all States and Territories, notably omitting some of the "low performing" ones (e.g. Bihar, Jharkhand, Uttar Pradesh, Uttarakhand, Madhya Pradesh, Chhattisgarh, Orissa, Rajasthan and Assam).  

These 9 States and Territories are covered in the Annual Health Survey (AHS) which ran for 3 years (and I think has now been discontinued), but these only cover things like IMR, Under 5 Death rates and Maternal Mortality rates--all important, but not nutrition indicators. 

In addition there are (1) the National Nutrition Monitoring Bureau (NNMB), which conducts more specialised surveys on nutrition and diet, (2) the Indian Human Development Survey (IHDS-II) run by NCAER and the University of Maryland  (results on nutrition to be released soon) and (3) a rapid UNICEF national nutrition survey which has been referred to in meetings I have attended in Delhi, but on which I cannot find any references to via the internet. 

What to make of all this?  Well, the AHS data show that modest improvements in IMR, U5M and MMR are being made, but mainly in the districts that show the lowest initial rates.   The IHDS data are rumoured to show not much improvement in stunting and underweight.  The UNICEF rapid assessment survey is rumoured to show dramatic declines in stunting and underweight.  

Forgetting about rumours, it is interesting to note that the DHLS IV shows Maharashtra with a stunting rate for under 5's of 30% (2013) which is down from 46 % using the NFHS-3 (2006, also for under 5's).  That is 16 percentage points in 7 years. This is very good, and is  a similar rate of decline as indicated by a recent Government of Maharashtra/UNICEF survey for under 2's which showed a decline of about 16 percentage points (from 39% in NFHS-3 for under 2's in 2006 to 23% in 2012). This is about the same rate as the all-Bangladesh figures.  That replication of declines across data sets --and the speed of declines--in Maharashtra is very encouraging.  

(Note the above para was changed after helpful comments from Saul Morris at CIFF and Victor Aguayo of UNICEF, who both pointed out errors in what I had reported--thanks!) 

For researchers interested in nutrition in India, this is fascinating.  For policymakers this is either a nightmare or a convenience (i.e. pick your favourite statistic).  Nevertheless, the Government of India badly needs to harmonise it's data collection because the main victims of the data confusion are malnourished people.  

So, too little or too much nutrition data in India? I think there is too much nutrition data because there is not enough of the right kind: regular NFHS rounds every 2-3 years, reported using the same age group. 

1 comment:

Saul said...

According to the NFHS 3 reports, the prevalence of stunting in Maharashtra in 2005-6 was actually 46.3%. A reduction of 16 percentage points in seven years is actually spectacular by historical standards. Karnataka did nearly as well.