AI in healthcare: India's trillion-dollar opportunity

Limited-resource settings in places like India mean a shortage of qualified healthcare professionals and inequitable access to healthcare.

India has just 64 doctors available per 100,000 people compared to a global average of approximately 150 per 100,000. Primary health centres and sub-centres at the rural peripheries are often woefully understaffed and lack critical infrastructure to meet patient needs. This often translates to a lack of high-quality diagnostic services, especially in rural India, home to more than 70% of the population.

Leveraging predictive analytics by using Artificial Intelligence (AI) for early detection can be a powerful tool for targeted public health interventions, especially in the context of limited healthcare capacity and delayed disease detection capabilities outside of urban centres. AI-enabled tools offer opportunities to bridge these inequities and reach AI maturity in the healthcare market in India, which is expected to reach $372 billion this year.

The growth potential of AI in healthcare

AI expenditure in India increased by over 109% in 2018, totaling $665 million and is expected to reach $11.78 billion by 2025, adding $1 trillion to India’s economy by 2035.

NITI Aayog, a public policy think-tank linked to the Indian government, has been testing the application of AI in primary care for early detection of diabetes complications, and is currently validating the use of AI as a screening tool in eye care, by comparing its diagnostic accuracy with that of retina specialists. Integrating AI capabilities with portable screening devices, such as 3Nethra, can expand the capacity for eye screenings and early detection, and enable access in remote places across the country.

Similar applications are possible in oncology. Tata Medical Center and the Indian Institute of Technology recently launched India’s first de-identified cancer image bank: the Comprehensive Archive of Imaging. AI-based tools can use high-quality de-identified images to enable machine learning models to detect biomarkers and improve outcomes for cancer research.

In cardiovascular healthcare, a major and somewhat unique challenge for India, Microsoft’s AI Network for Healthcare and Apollo Hospitals are developing a machine learning model to better predict heart attack risk. Using clinical and lab data from over 400,000 patients, the AI solution can identify new risk factors and provide a heart risk score to patients without a detailed health check-up, enabling early disease detection.

Realizing the economic potential of AI in healthcare requires a measured calculation and mitigation of the risks. Inaccurate decisions could potentially endanger individuals’ health.

AI requires massive amounts of data from multiple sources, and fragmented data availability can increase the risk of inaccurate decisions — for example, mistakes in tumour detection or inappropriate drug prescriptions.

Access to data must be contingent on informed consent and accountability. Patients need to be aware how their data might be used to train AI models and be clearly informed of what factors influenced a particular treatment decision or recommendation by a physician. This is especially important in the Indian context, where physicians usually spend very limited time with each patient, often just 1-2 minutes.

A high level of automation in disease detection could jeopardize the ability of physicians to accurately detect AI mistakes and prompt overreliance on AI-based tools, rather than investing in healthcare infrastructure to ensure patients have appropriate access to healthcare staff.

AI should support healthcare decision making, not be used to automate decision-making. Adopting meaningful human control by enabling physicians to provide feedback on AI model proposals and providing a continuous learning loop can mitigate some of these harms.

The growth of the healthcare market must come hand-in-hand with the parallel goal of achieving universal health coverage. AI systems should never be used as a substitute for access to primary healthcare, but should instead accompany increases in healthcare spending to ensure rural and peripheral populations can access high-quality healthcare.

India's way forward

AI maturity in health requires critical investments in the capacity of the workforce, data and infrastructure, governance and regulatory mechanisms, design and processes, partnerships and stakeholders as well as innovative business models.

Integrating AI into healthcare systems also requires an understanding of AI in national curricula for medical and public health students, both academic and practical.

Similarly, the Indian government will need to make appropriate investments in data infrastructure, such as interoperability, unified EMR and data stewardship. This is essential to build trust and long-term integration of AI into India’s healthcare system.

The government must also invest in and build public-private partnerships across the healthcare industry to facilitate coordination between academia, government, industry, NGOs and patient advocacy organizations. They should scale governance and regulatory mechanisms to provide appropriate oversight for privacy, fairness and transparency.

NITI Aayog’s National Strategy for AI prioritizes principles of privacy, ethics, security, fairness, transparency and accountability, as well as alignment with the rights afforded by the Indian Constitution. India is a founding member of the Global Partnership on AI alliance and has thus far adopted a measured approach for integration of AI, in keeping with ethical and responsible standards. These principles must be applied in practice as the technology scales.

The way in which AI systems are integrated, too, will be crucial. Human-in-the-loop and human-centric designs that empower healthcare staff to understand how a decision is made and how to incorporate this knowledge into treatment will minimize risk.