What the newly released infant mortality rates tell us

Late last month, Kanika and I published an article in scroll.in, discussing the newly released infant mortality rates (IMR). We also used the occasion to discuss the importance of the IMR as a health statistic. I am posting the article in full below, partly because scroll posted only screenshots of our charts, and not the original ones we had created using google api. All the data that we used in this article is available in this spreadsheet.

This month, the Office of the Registrar General of India released data on Infant Mortality Rates (IMR) for 18 states and 5 Union Territories for the year 2014. The numbers show that even as most states made some progress in reducing the number of infant deaths, rich states like Delhi and Gujarat continue to have unreasonably high infant mortality rates. Improvement in rural areas was less impressive. Worryingly, rural Rajasthan and Bihar showed slight increases in infant mortality, while rural Madhya Pradesh and Gujarat made no progress.

These numbers come after a long two year drought of mortality indicators in India. Data for 11 states and all-India estimates are still awaited.

The Infant Mortality Rate measures the number of infant deaths before age 1 year for every 1,000 live births. Along with life expectancy, the Infant Mortality Rate is the most important health statistic in any country. Improvements in overall and adult health are impossible without declines in infant mortality. Despite the relative importance of the IMR, and the release of new health indicators in a long time, this new data has received scant attention and commentary.

These numbers are based on the Sample Registration System (SRS) surveys. Ideally, all births and deaths should be registered, and infant mortality or life expectancies should be calculated based on civil registers of deaths and births. But in countries such as India, where birth and death registration is incomplete, sample registration surveys are used to estimate fertility and mortality.

The release of this data was delayed partly because the SRS re-calibrated its sample based on the results of the 2011 census. Usually, IMR figures are released more periodically: for instance, IMR figures for the whole of India and all states for the year 2013 were available by September 2014.

What the data reveals

Figure 1 shows a comparison between the 2014 and 2013 Infant Mortality Rates for states for which the 2014 data has been released. The states have been arranged in the descending order of their 2014 IMR.

The IMR has declined in all states except Bihar and Manipur, a positive development on the whole. It is highest in Madhya Pradesh, a state which has had the highest IMR in India for a long time, and lowest in the states of Sikkim, Nagaland and Manipur.

Delhi, which is richer than all the states here, lags behind Sikkim, Nagaland and Manipur. One possible reason, among others, for this is the high toll of death among infants caused by the toxicity of Delhi’s air.

States of the Hindi-heartland, as well as Assam and Odisha, have worse IMR than other states. This is along expected lines.

Figure 2 presents the same comparison between IMR for 2013 and 2014, but only for rural areas. While the overall state patterns are similar, rural IMR is generally higher than total IMR, and progress in reducing rural IMR is less sharp.

Out of the 18 states for which 2014 figures are available, IMR in rural areas declined in 12. Madhya Pradesh, Meghalaya, and Gujarat made no progress in reducing rural IMR, while it increased in rural areas of Rajasthan, Bihar, and Manipur. Except Assam, where rural IMR showed a decline from 56 to 52, none of the states saw a steep decline in infant mortality.

Curiously, rural Jharkhand has lower infant mortality than rural Gujarat, while rural Bihar, which is also very poor, has the same level of infant mortality as rural Gujarat. Lower IMR in Bihar (and Jharkhand) and high IMR in Gujarat has been a long puzzle in Indian demography.

Figure 3 shows IMR for these states’ urban areas. Bihar and Uttar Pradesh have the highest infant mortality in urban areas, followed by Odisha, Madhya Pradesh, Uttarakhand, and Meghalaya. Nagaland, Sikkim, Arunachal Pradesh and Manipur have the lowest urban infant mortality rates.

Urban Delhi, which had 20 infant deaths for every 1000 live births, lags behind Nagaland, Sikkim, Arunachal Pradesh and Manipur.

The importance of IMR

Infant mortality is one of the best indicators of the overall health of a population. The first reason for this is that the same things that protect infants from death protect them from disease, and those same things improve the health and productivity of adults too.

A very large literature on “early life health” emphasises that the first 1000 days of life, starting from conception are critical. The brain and body are developing at this stage, and damages due to bad health are often irreversible.

Secondly, the biggest improvements in life expectancy and health come from reduction in infant and child mortality.

Figure 4 shows age-specific mortality rates for all ages in India, for the years 1971, 1991, and 2013, based on the Sample Registration System. The age-specific mortality rate is calculated by the SRS as the number of deaths in a particular age group divided by the mid-year population in that age group. All figures are per 1000 population.

The figure shows a pattern well known to demographers and public health experts: mortality is high in the very young ages, low during adolescence and the middle years of life, and rises quickly with the onset of old-age. It also shows declines in mortality at all ages between 1971 and 2013.

But what is even more dramatic is that the big declines in mortality happened at the youngest ages: in the 0-4 years age interval. Child Mortality declined from 52 deaths per 1,000 in 1971 to 27 deaths in 1991, to 11 deaths in 2013.

Adult mortality declined too, but the magnitude of the decline was less: from 109 deaths per 1000 in the 70+ age group in 1971 to 91 deaths in 1991 and 79 deaths in 2013. The children whose lives were saved had many more years to live than the older adults whose survival also increased.

Consequently, in the life expectancy improvements in this period, which was 50 years in the period 1970-75, 60 years in the period 1991-95 and 67.5 years in the period 2009-2013, reduction in child mortality had a much larger role than declines in any other age.

Neglect of health in public discussions

Infant mortality, and these new rates, deserve much more careful scrutiny. The lack of attention accorded to these speaks volumes about the overall neglect of health as an important topic of public discussion and debate in India.
Jean Dreze and Amartya Sen, in their 2013 book, An Uncertain Glory: India and its Contradictions, analysed more than 5,000 op-eds in the last six months of 2012 in all the leading english language papers. They found that just 1% dealt with health related matters.

This neglect, which seems to continue, is tragic. The reasons why Gujarat or Delhi has such high infant mortality given their incomes reiterate the dissociation between India’s growth story and other development indicators. Similarly, India now has higher infant mortality than both Bangladesh and Nepal, even though these countries are much poorer than India. Slow progress, or stagnation and reversals in infant health in some states are also causes of worry.

India’s GDP numbers, which are released quarterly, create a splash in the front pages of all newspapers. Consequently, governments feel that they are constantly monitored and are pressurised to act. Health indicators are released much less frequently. When they are, they generate hardly any debate and discussion, let alone any action.

Progress on the health front in India is unlikely to accelerate unless this changes.