By Srabashi Basu
One of the Sustainable Development Goal (SDG) targets set by United Nations for all countries in the world is to reduce preventable deaths of children under 5 years of age to at most 25 deaths per 1000 live births, by the year 2030. India has been showing steady improvement in reducing under-5 mortality rate (U5MR) and the Statistical Report 2020 estimates U5MR to be 32 deaths per 1000 live births as compared to 35 reported in 2019.
However, within the geographic boundary of India, wide variability exists. Whereas a few states, such as, Kerala, Sikkim, Tamilnadu, Manipur and Maharashtra, have either achieved the SDG30 goal or are very close to it, there are other states, such as UP, MP, Bihar etc, who are at the extreme end of the spectrum. In Uttar Pradesh, U5MR was 72 per 1000 live births in 2015-16, comparable to that of sub-Saharan Africa, one of the poorest parts of the world. To bring equity among the states in India, it is not enough to reduce the country-average U5MR, but each state must also strive at the grassroot level to deal with this problem with high priority.
We have noted that in India, U5MR is almost totally dominated by infant mortality rate (IMR). IMR is defined as the number of deaths occurring between 0 and 364 days after the birth of a child per 1000 live births. Current figures of UP reports U5MR to be 60 whereas IMR is 50 (per 1000 live births). It is evident that reduction of IMR will have a profound effect on achieving SDG30 goal.
Among the important risk factors of IMR are pre-term birth complications, birth asphyxia or trauma, pneumonia, diarrhoea, mother’s health and nutritional status during pregnancy, sanitation status in household, mother’s education, mother’s exposure to media, mother’s ability to decide birth spacing (unmet need for birth-control, easy access to birth-control) and many other social, economic and health-related factors. Let us consider a few of these putative risk factors and how artificial intelligence (AI) can intervene for an effective monitoring program.
There is a significant difference between IMR in urban and rural areas (50 versus 62.5 deaths per 1000 live births in UP as reported in NFHS-5). In India rural healthcare suffers from lack of funding, resulting to lack of manpower as well as lack of facilities in primary care centres. In case of a complicated delivery, referral to a well-equipped secondary or tertiary facility often becomes too late to be useful. Naturally identification of infants at risk is a critical consideration. Typically, low birth weight, mother’s health and care during pregnancy, pre-existing risk factor for the mother, such as smoking and drinking habits, being subject to domestic violence and other negligence contribute directly to an infant’s survival status.
It is possible for the primary care physicians and the ASHA workers to keep a tab on some of these factors and provide counselling to the mother and her household. It is equally likely that, such counsel will not be adhered to! However, there are other factors, such as the mother’s current age, age when she conceived for the first time, total number of children born to the woman, birth order of the current child, spacings between the children born to her, sex of the first-born and sex of the children born previous to the current child and many other such factors, which are too numerous for a physician to take into consideration to determine an infant’s survival probability.
This is precisely the point where AI can intervene in an effective manner. With the use of sophisticated data science algorithms developed with intelligent input from all relevant variables, AI can help in early identification of high-risk births. In fact, AI can even detect patterns of area-specific high-risk infant mortality, thereby raising flags for further investigation about local social practices. Mother’s level of education and exposure to media have been found to be important factors for overall well-being of infants and children. AI can help in sharpening and personalizing the media to extend its effective reach.
AI is a decision support system. It is not to take the places of the physicians and the care-givers, but AI can surely increase the effectiveness of the support system. In countries like India, where in remote and rural areas medical support is inadequate in many senses, AI can ensure that the neediest infants and children get the highest care.
The author is professor and program director, Great Learning