The Curious Case of Anaemia

India is considered a country with a chronically high rate of anaemia, particularly for women and children. This issue came up again when the Ministry of Statistics and Programme Implementation released its latest 'Women and Men in India 2022' report based on data from National Family Health Survey(NFHS) 2019-21.

One may quibble about the pace, but India’s per capita income and medical/sanitation coverage have visibly gone up over the last quarter century. Moreover, several government schemes have been used to specifically target anaemia (National Iron Plus Initiative, Anaemia Mukt Bharat, etc.). Yet, as per NFHS, the prevalence of anaemia has hardly budged. The prevalence in under-five children has barely decreased from 70% in 1998-99 to 67% in 2019-21 while the rate for women aged 15-49 has increased from 52% in 1998-99 to 57% in 2019-21. Why is this happening?

When measuring the prevalence of a disease in a population, there are two key factors to consider. Firstly, the thresholds used as a benchmark against the measured biological marker. Secondly, the testing methodology used to ensure accuracy in results. There are serious problems with NFHS data on both accounts.

The current WHO benchmarks go back to five studies of predominantly white adult populations in North America and Europe in 1968. There have been many studies that show that a single global Hb cutoff to define anaemia for all ethnicities and regions isn't scientific. A recent study published in the Indian Journal of Medical Research in 2019 (Varghese et al.) examined the Hb levels of healthy women from using data from the US and UK. The study's results were consistent with previous literature, finding that the mean Hb levels of healthy Blacks, Hispanics, and Asians were lower than those of Whites.

Another study published in Lancet Global Health in 2021 (Sachdev et al.) based on India's Comprehensive National Nutrition Survey, conducted between 2016 and 2018, calculated age-specific and sex-specific Hb percentiles based on biomarkers of 50,000 children and adolescents. The study found that the healthy Hb level was consistently 1-2 g/dL lower than the existing WHO cutoffs at all ages. The difference was even more significant for children of both sexes aged 1-4 years and girls aged 10 years or older. Note that a mere lowering of the threshold by 1 g/dL would result in a drastic fall in the headline anaemia figure by around 29% for under-five children and 26% for women aged 15-49.

Similarly, we found that method of blood sample collection causes wide variation in results. The protocol has two steps - blood collection, and the testing of the sample. The two primary ways to collect the blood sample are – (i) capillary sampling, which involves pricking the fingertip to obtain a drop of blood; (ii) venous sampling, which involves drawing blood from a vein using a needle.

Multiple studies have pointed out accuracy issues with capillary method tested on a portable HemoCue 201+ analyzer used across NFHS. It has shown that capillary blood drawn gives a falsely lower Hb level than venous blood, eventually resulting in a higher headline anaemia prevalence. A major study published by New York Academy of Sciences (Neufeld et al.) in 2019, using data on women aged 18-49 in Uttar Pradesh, found that Hb concentrations measured through capillary samples using Hemocue 201+ were lower on average by 0.9 g/dL. Thus, anaemia prevalence using capillary sample was 59.2% compared to 35.2% using venous samples.

Another study (Levy et al.) published in 2017 found that capillary samples using HemoCue 201+ underestimated Hb concentrations, with a mean difference of 1.5 g/dL. These are very significant differences, and the problem shows up in other countries too. South Africa, which uses a similar methodology to India’s NFHS, found childhood anaemia at 61.3% in 2016. This is clearly suspect given that the country’s per capita income makes it a middle income country.

The previously mentioned Lancet Global study (2021) corrected for both the benchmark and measurement problem. Its estimates used the gold standard cyanmethaemoglobin method. Compared to 67% anaemia in children under-5 as per NFHS, the measurement against updated benchmarks saw a fall in prevalence to 15.4%. This figure drops to 14.9% for girls aged 15-19, compared to 59% as per NFHS. The supposed lack of progress on anaemia, despite growing prosperity and targeted schemes, turns out to be a measurement error!

We are not arguing that anaemia should not be a public health issue. However, we need an accurate measure of the situation. First, public health policy is about trade-offs. By exaggerating one problem, we are diverting resources from others. Second, poor data means we cannot judge the impact of government measures. Feedback loops are essential for effective policy-making.

Sanjeev Sanyal is Member, and Chirag Dudani is Consultant, Economic Advisory Council to the Prime Minister.