How AI in healthcare can accelerate the journey towards the UN Sustainable Development Goals 2030
By Ajit Ashok Shenvi
“Access to care is a fundamental human right”. United Nations’ Sustainable Development Goal-3 is about ensuring healthy lives and promoting well-being for all at all ages. All UN member states have committed to achieve this universal health coverage (UHC) by 2030. The path to UHC is not easy though, as India and developing economies are constantly plagued with the hurdles of Availability, Accessibility, Affordability and Awareness of Healthcare. Conventional approaches alone will not help address these challenges. Connected solutions enabled by AI will disrupt healthcare making the dream of UHC a reality.
Below are some categories of use cases (not exhaustive) where AI can and is playing a significant role in accelerating this journey.
- Augmenting the expertise of HC providers:
The sheer volume of cases in a Govt Hospital like AIIMS can make life hell for a scarce resource like the radiologist. In many cases, the images be it CT, MR, US would be normal, but the radiologist would have to go through all of them with the same level of granularity/detail. Human fatigue and monotony can lead to mistakes and errors. AI models integrated in the workflow can already identify abnormalities, or atleast put the risky cases higher in the review order. In other words, AI can be “the intelligent” assistant augmenting the expertise of the radiologist, enhancing her productivity, allowing her to cater to many more cases than what she would do normally.
Generative AI can take this much further where it can sift through images, reports, correlate it with vast medical information on the internet/research journals and suggest options that the doctor can use to decide the treatment options. While this may seem like science fiction today, it may be reality soon with advancement of tools like GPT4.
- Population Health Management:
Burden of chronic diseases on the Indian healthcare ecosystem is enormous. For e.g. Consider TB (Tuberculosis) – As per Global TB report 2021, around 1/3rd of global TB burden and corresponding mortality is in India. When diagnosed in time, TB is curable in 6 months, however the biggest challenge in this journey is early screening. World Health Organization (WHO) recommends the use of digital X-ray as a rapid low-cost TB screening method.
India has an ambitious Kshay Mukt Bharat program (NTEP–National Tuberculosis Elimination Program) that aims to make India TB free by 2025 (5 years before the WHO/SDG goal). A solution that is fast, affordable and easy to implement in resource constrained settings is the need of the hour. This is where AI can come to rescue. A screening van or a Ayushman Bharat Health & Wellness centre equipped with a Portable Digital X-ray can screen individuals. The AI model integrated in the workflow can analyse the scans indicating abnormalities. The high probable cases can then be directed to nearest Primary/Community Health centre for confirmatory tests and treatment.
Similar workflows with integrated AI models can be implemented for Cardiovascular, Breast cancer screening – the next biggest challenge for our nation.
Every year, India gets infested by number of communicable diseases like Malaria, Dengue etc that follow cyclical patterns. While there is a strong IDSP (Integrated Disease surveillance program) run by Govt, there is an opportunity to make it proactive by correlating it with historical data, social media data, weather data (e.g. rainfall etc). AI models based on these can predict onset of these diseases that can help Govt/private health infrastructure be prepared for such outbreaks. IET HC working group demonstrated a feasibility of such a possibility based on social media data – Link to SARTHI
- Efficiency and waste reduction:
The AI discipline is in its infancy with low trust on its efficacy considering bias, explainability and other factors. While the use cases directly dealing with patient care may take some time to mature, there is a huge opportunity to ensure uninterrupted operations using AI based predictive and proactive maintenance. Downtime of equipment’s like MR, US, CT etc amplify the woes of already choked healthcare system. AI based maintenance systems can ensure uptime so that care is not denied to needy patients.
Availability of hospital beds (especially ICUs) is another daunting challenge. Here again AI models to determine discharge readiness, early warning score to gauge patient deterioration etc can ensure efficient flow of patients, so that beds are available for the needy.
- Wellbeing and preventive care:
The holy grail of any healthcare system is to ensure that people are kept outside hospitals – in other words focus on well-being and preventive healthcare. It is encouraging to see digitally savvy Indians engaged in their own health with wearables/trackers etc. AI can mash up data from daily activities, sleep patterns etc to give personalized triggers for individuals to manage their health proactively.
All these use cases demonstrate the immense possibilities that exists to decentralize/democratize healthcare. The Covid pandemic although a nightmare, did have a silver lining in alleviating the inhibitions of embracing digital technologies especially in a highly regulated healthcare industry.
The siloed/fragmented healthcare system in India makes it challenging for AI as data is not shared amongst healthcare providers making it difficult to get a longitudinal view of individuals. Ayushman Bharat Digital Mission is developing the backbone necessary to support this integrated digital health infrastructure. Strong regulations including the security/privacy bill will go a long way in allaying fears and trusting the reliability of these digital health technologies.
In conclusion, AI is neither an end-in-itself nor will replace healthcare professionals in foreseeable future, but it is an indispensable piece of solving the UHC jigsaw puzzle in Digital India.
(The author is Ajit Ashok Shenvi, Member, Healthcare Working Group, IET Future Tech Panel, and the views expressed in this article are his own)
