Tuberculosis is one of the top 10 reasons for death worldwide, and the leading cause of a single infectious agent (above HIV/AIDS) according to the World Health Organisation. A total of 10 million people fell ill with tuberculosis globally in 2018, and of them, 1.5 million people died.  

Caused by a bacterium called mycobacterium, TB can affect any of the body parts, including lungs, kidney, spine, or brain and could be contagious and fatal. WHO says that as many as 50 percent of tuberculosis cases may go unreported and unknown.

In India, the estimated TB incidence is 27 lakhs as per the India TB report 2019 by the Department of Health and Family Welfare, Government of India. This is a 16 per cent increase as compared to 2017, says the report. In the TB mortality rate, currently India stands second in the world. Despite a strong eradication programme, almost half of all tuberculosis cases in India go undetected.

To bring in efficiency and accuracy in early detection and enhance the quality of service delivery in healthcare, the Central government has started utilising AI-based technologies. NITI Aayog’s National Strategy for Artificial Intelligence report says that AI-based solutions can be used to address the problem of “scarce personnel and lab facilities” and “help overcome the barriers to access”. AI-based solution not only -help to diagnose the disease in India but, “an advanced AI-based solution for early diagnosis of tuberculosis, for example, could easily be rolled out to countries in South East Asia or Africa, once developed and refined in India,” says the report.

AI for the aid

To build an AI-based solution in dealing with TB, last year, the Central TB Division of the Health Ministry has signed a Memorandum of Understanding (MoU) with the Mumbai-based Wadhwani Institute for Artificial Intelligence- a research institute that develops AI solutions for social good.

As per the partnership agreement, Wadhwani will be supporting the Revised National TB Control Programme (RNTCP) to be AI-ready to eradicate TB. This includes developing, piloting, and deploying AI-based solutions. These solutions would support the programme in vulnerability and hot-spot mapping, modelling novel methods of screening and diagnostics, and enabling decision support for care-givers.  

As far as India is concerned, the country has put a target of eliminating TB by the year 2025, which is five years ahead of the global target set by WHO.

“We are in the midst of working on a number of exciting opportunities for developing AI for social impact,” says Dr P. Anandan, CEO, Wadhwani ai. He says that to address this through AI, especially, problem identification, testing and deployment, require experience in working in the field with the communities whose problems we are trying to address. “Equally important is experiences in working with local, state, and national level government organisations,” points out Anandan.  

Data is prime 

“Data is fuel for AI. So, we work closely with the government, understand the pain points, get the data we need and build an AI system that would reach the bottom level,” says Rajesh Jain, Senior Director Programs, Wadhwani. Currently, there exists a centralised strategy and execution; dedicated government staff and budgets; support from organisations such as WHO; robust digital infrastructure, and multiple touch-points across a patient’s journey.

Around this, the institute is creating technologies to address multiple challenges across the cascade of TB care, starting with caseload estimation at the district level using a variety of risk and transmission factors to help identify missing cases and the prioritisation of TB patients for health workers through stratification of the risk of drop-off from treatment.

 Other areas 

Apart from TB, NITIAayog also recognises cancer as one of the critical issues where AI-based solutions can be of help. The government is developing an imaging biobank for cancer, that will be a national repository of annotated and curated pathology images. AI-based radionics is another emerging discipline. This will contain comprehensive quantification of tumour phenotypes by applying a large number of quantitative imaging features. Diabetic retinopathy is another major field for building AI-based solutions.