The honourable interim finance minister Piyush Goyal has made an allocation in the recently announced interim budget for a National Center for Artificial Intelligence (NCAI) and a national AI portal. This will have far reaching implications in many fields and the biggest impact will be if AI is used effectively in healthcare. In fact, the buzz words in healthcare recently are, machine learning (ML) and AI. There has already been significant interest in AI in healthcare in India and around the world. In my opinion, we have not even scratched the surface and should use the initiative shown by the finance minister to make the best use of it in healthcare. The use of AI in maternal and child health holds a lot of promise and is required most among all medical specialties.
The current healthcare indices in maternal and child health are far from desirable in India. India ranks 128th in terms of meeting the United Nations’ (UN) health-related Sustainable Development Goals (SDGs) by 2030, with low scores on air pollution, sanitation, hepatitis B and child wasting. India is home to 46.6 million stunted children, a third of world’s total. With 46.6 million children who are stunted, India tops the list of countries followed by Nigeria (13.9 million) and Pakistan (10.7 million). One third of all women of reproductive age in India have anaemia. India tops the list of 10 nations contributing 60 per cent of the world’s premature deliveries. The SDG targets for 2030 will reduce the global maternal mortality ratio to less than 70 per 100,000 live births. By 2030, end preventable deaths of newborns and children under five years of age, with all countries aiming to reduce neonatal mortality to at least as low as 12 per 1,000 live births and under-5 mortality to at least as low as 25 per 1,000 live births. It is very clear from the above data that maternal and child health should get top priority as it will have long-term benefits by ensuring that the future of our country stays healthy by having the best outcomes for mothers and their babies. A healthy mother leads to an entire population that is healthy. AI in maternal and child health will have far reaching benefits for individuals and communities across the world. This will go a long way in attaining one of the most important sustainable development goals.
There are many reasons why maternal and child health lends itself to artificial intelligence. The whole period of pregnancy, child birth and preventive care in mother and child is defined and extends from preconception to five years of age. The period of pregnancy has very clear protocols for the number of visits, the lab and radiology investigations that need to be done, the medications that need to be taken, the complications that need to be looked for and a definite end point which is the birth of the child. Similarly, the care of the child is also dictated by protocols in both the immediate care after birth and preventive care for five years till the immunisations are completed. The data for the above touch points are available in many settings of maternal and child care like government clinics and hospitals, NGOs, insurance companies and private healthcare organisations.
What India needs to do?
First and foremost, we need to identify the use cases and questions to be answered where AI will give us the best bang for our buck. Second, we need to identify what data is needed for the use cases. Third, we need to find out where we can get the retrospective data from and how. This will be required to enable machine learning. Fourth, we need to figure out a methodology of collecting prospective data that will enable both machine learning and artificial intelligence. Fifth, we will need the computing power and technology to utilise all this information in the most practical manner. Sixth, we need to write algorithms that will help us solve real world problems. Seventh, we need to implement solutions that are generated by artificial intelligence to help improve maternal and child health.
After having identified the why and what, we need to have a plan regarding how we can accomplish this. We need to have key opinion leaders to identify the use cases and questions that need to be answered. There should be representation from healthcare providers from a wide cross section of providers in the maternal and child health field. The data that is required should be identified by the subject matter experts. We will then have to identify the sources of such data and find out the best way of gaining access to that data and collating it. This should include both retrospective and prospective data. For the data to be used meaningfully, we need to create a data ingestion tool that helps machine learning. Algorithms will need to be prepared by medical and technology experts working closely together. These algorithms should help us identify high risk pregnancies, women at risk for anaemia, pregnancies that may lead to pre-term deliveries, infants and toddlers at risk for being unimmunised, babies in the neonatal intensive care units that are at risk for sepsis and poor outcomes. With huge volumes of data, ML and AI can actually even suggest methods of improving outcomes. Once we have solutions suggested by artificial intelligence, we need to have concerted and coordinated efforts by all stakeholders to implement them on the ground in a practical and realistic manner to improve maternal and child health.
The allure of AI is great and I believe it will help us solve many problems faster than we ever imagined we could. However, while the cliché is that the “devil is in the details” but in this case “the devil is in the data”! There can be no machine learning or AI without data. Unfortunately, this is often forgotten in many discussions. Medical data is pleomorphic. It is unstructured, fragmented, unreliable, illegible, written, digital, inaccessible and also in various other shapes and forms. It is going to take a humongous effort to compile the data required in the manner required to enable ML and AI. This alone will require a number of stake holders to work together. This will include patients as well as it is their data that is going to be used. The government will have to give the necessary impetus by framing the rules to enable easy availability of data for the greater good. All agencies private and government will have to share data and input the data into the common data ingestion tool. There needs to be collaboration with the government, teaching hospitals, WHO, Gates foundation, UNICEF and health insurance providers for getting as much as data as possible. Agencies like Nasscom should be the liaison agency between the data sources and the technology companies to optimise ML and AI. The steps outlined above will be required for use of ML and AI in all medical specialties not just maternal and child health.
All this will be difficult but not impossible. I am an eternal optimist and a complete believer in technology and will bet on this happening. Given the importance of maternal and child health I hope it happens sooner than later.