How AI-driven Frugal Innovations Can Boost India’s Healthcare System

Kunal Walia Sep 16, 2024

According to Dalberg analysis, major global philanthropies have pledged more than 1.6 billion USD to AI-related activities over the last five years. And of this, 56 per cent of funds were earmarked for healthcare

Artificial Intelligence (AI) has witnessed a marked rise in investor interest across sectors in recent years. Data from Stanford University’s Institute for Human-Centered AI indicates that 40 per cent of India’s total investments in AI from 2013 to 2022 were made in 2022. The healthcare industry is no exception: According to Dalberg analysis, major global philanthropies have pledged more than 1.6 billion USD to AI-related activities over the last five years. And of this, 56 per cent of funds were earmarked for healthcare.

However, a rise in total investment by itself is no guarantee that the tech will be equitably distributed. Issues of inclusion, access, quality, and restrictive social norms also remain challenges for the delivery of healthcare. To ensure that AI-driven technologies reach vulnerable and underserved communities, these innovations must be customised to cater to diverse users. However, while AI can solve some of the sector’s most fundamental issues, it carries an equal amount of risk. To allay these concerns, developers and funders need to heighten efforts to put responsible and ethical AI at the centre of innovation.

How do frugal innovations build inclusion into the product?

Frugal innovations refer to products that are built with minimal resources, are simple in design, and cost-efficient. These features make them easy to deploy and scale across regions. Importantly, they put the user at the centre, making it ideal for countries such as India where a significant proportion of the population faces a shortage of healthcare workers and has little to no access to primary healthcare. As our population expands, it is likely to create new pressures on our healthcare system and deepen the need for a more even distribution of healthcare workers.

Digital innovations can fill this gap by ensuring last-mile connectivity to remote areas that have poor healthcare infrastructure. An example of this is the AI-powered breast cancer-screening device developed by health tech startup Niramai. The hand-held device does not require substantive electric power and is easy to transport. Having fast emerged as a viable alternative to mammography, it has helped healthcare providers widen the reach of breast cancer diagnostic services. The Niramai device also shows how frugal innovations take cultural sensitivities and realities into account: Mammography machines are large and require patients to undress before undergoing screening. This can be intimidating and uncomfortable for many women. Hand-held devices, with which women can take control of the process, encourage more women to participate and receive care in time.

Importantly, these technologies need to be cost-effective and affordable. The growing popularity of telemedicine, for instance, illustrates this well. Telemedicine and phone consultations have reduced the need for in-person visits, driving down transport cost and increasing time efficiency, for both patients and doctors. This has been instrumental in allowing doctors to deliver medical advice, give prescriptions, and follow up on ongoing treatment at a faster pace. Ethical and responsible integration of AI into these services can further boost its impact potential. We are seeing some part of this happen already: The Government of India has launched an AI-based telemedicine mobile clinic which can interact with patients in their native language in Jammu and Kashmir’s Udhampur.

Further, AI can also help deliver higher quality of healthcare to low-income communities. The non-profit Wadhwani AI, for instance, is building a solution that aims to reduce India’s infant mortality rate. It is building a deep learning model that can estimate a baby’s weight at birth by analysing video data, even if shot using low-cost smartphones. It can be game-changing for a country like India, where 83 per cent of neonatal deaths in India occur due to complications from low birth weight but several low-weight births go unreported. Through the AI solution, Wadhwani AI can help improve the quality of neonatal care and primary care for new mothers.

To scale effectively, responsible AI should be top of mind

If applied effectively, AI in healthcare can improve health outcomes and reduce treatment costs by a big leap. However, there is an equal risk of amplifying the inequities of an unequal healthcare system. A 2019 study published in Science illustrated what could go wrong: The journal evaluated an algorithm used to predict healthcare needs for more than 100 million people in the US based on spending data. However, because Black patients have historically lacked access to care and so, tend to spend less on healthcare, Black patients had to be much sicker to be recommended for extra care under the algorithm.

Collecting diverse and local data and being mindful of cultural and linguistic nuances is a starting point to prevent biases from creeping in. Making this localized data available using open datasets is important for developers to leverage and train AI models. Promoting responsible AI also involves embedding a ‘do no harm’ principle at the design stage. We must also ensure that user privacy is protected and that data is used with consent. Stringent regulations are needed to ensure that new solutions are not being tested on the most vulnerable populations. These are just some of the areas that need to be addressed while developing AI-based solutions in healthcare.

While this shift in thinking needs to happen at the level of technology creation, funders have a big role to play, as they can institute a minimum level of checks that investees need to meet to ensure ethical creation and deployment. They also have a role to play in bringing the ecosystem together and sharing the evidence. It is also crucial for implementation organisations such as CSOs, NGOs grassroots experts and governments to partner at the pilot and testing stages. As stakeholders with the highest proximity to the end-users, they can collectively leverage their networks on the ground to ease adoptability.

If done right, responsible AI solutions could help solve our deepest healthcare challenges, potentially qualifying for some of the most impactful technological breakthroughs of the coming decade.

----

Kunal Walia is Partner at Dalberg Advisors, where he leads the firm’s global digital and data practice.