If we want rapid digitisation of India’s healthcare, grounds-up endorsement of some form of a digital solution by end users right from the grassroots is the way to go

Imagine if we could predict and mitigate public health disasters, predict disease burden at population level and implement preventive measures to reduce deaths and hospital admissions, dramatically cut down the time and cost of drug discovery or provide personalized treatment for every individual at a fraction of the cost. Artificial Intelligence (AI) has created the possibility of achieving them and more, which were touted as wishful thinking just a few years ago.

Now imagine the impact it can have on a $372-billion Indian healthcare industry with CAGR of 39 per cent. It is phenomenal. AI has become the buzzword in the government, industry and academia.

On the contrary, half way across the world, even countries with established Electronic Health Record systems are struggling to scale-up AI implementation. Most of the algorithms that feature prominently in research literature are nowhere to be found on-ground.

So, what do all these really mean for Indian healthcare? As with any domain, bringing AI on-ground is a complex task with many stakeholders and interests. So where do we start? What is that one thing that can make the difference, if focused on?

When we look at different stages of development of AI application, we are faced with three pertinent questions.

How do we increase adoption of digital health?

The fundamental prerequisite to derive benefit from AI is to first generate data through rapid digitization at all levels of healthcare. There are broadly two ways of doing this: top-down enforcement versus ground-up endorsement. The US has used the famous “carrot and stick” method with the carrot being insurance payouts and stick being legislative backing with Health Information Technology for Economic and Clinical Health Act and the Affordable Care Act.

In India, more than three-fourth healthcare services are being provided by private sector and around 80 per cent of them are less than 50 bedded, small mom-and-pop facilities. With a sizeable chunk of healthcare providers being in informal sector, government's control is limited. Therefore, any attempts to enforce EHR adoption by legislative route are unlikely to make it beyond the walls of the Indian Parliament. The Government of India, though, is trying to use the Ayushmaan Bharat health insurance payouts to nudge hospitals to go digital. This approach is limited to in-patient episodes and it misses the rest of the patient journey.

If we want rapid digitization of healthcare, ground-up endorsement of some form of a digital solution by end-users right from the grass-root level, is the way to go. Considering the vast variations in medical practice and consequently varied requirements, different kind of products with different revenue models catering to different market segments would be required. India being a price sensitive market, user fee may not be a viable source of revenue for many market segments. Other data driven streams of revenue are required.

How do we prevent formation of silos and data feudalism?

Data feudalism is characterized by lack of interoperability and data silos, which happens when there is no incentive to share data or incentive not to share data or both. Standardization (EHR standards) is often put forth as a panacea for this problem of interoperability and the Government of India is also aping the west in this regard. However, when we look at the US healthcare system, which is often seen as an epitome of standardization with EHR standards etched into their laws, interoperability still remains a pipedream. This approach has failed miserably in the US despite legislative backing and should be a lesson for India.

Besides, there are various Standards development Organizations (SDOs) who mutually compete with each other and end up creating conflicting standards leading to incompatibility. The process of development of standards is also at times questionable with balloting process being influenced by pressure groups formed by proprietary interests. Also the standards are known to get outdated quickly, especially with rapid advances in medicine and technology.

Skewed regulations can potentially be barriers for new & small entrants. This leaves fewer players in the market, who grow stronger over time and oligopolies get established.

These incumbents will have a strong financial incentive to maintain status quo. Any new player or a new innovative idea is a threat to them and they would naturally lobby hard to create more entry barriers by influencing the standards. This could turn out to be the "Vicious Triangle for Indian Healthcare" akin to the Iron triangle of US politics, often called sub-government due to their power to influence policies.

So, what is the way out? We need an ecosystem that encourages innovators to build custom solutions for wide range of problem statements ranging from simple ones like reducing patient waiting times, automated medical records management, improving medication compliance to advanced once like clinical decision support, automated second opinions etc. This would create a natural financial incentive for all of them to collaborate with each other and they will find ways to seamlessly share data. Therefore, our policies should not let the vicious triangle or Data Feudalism to set-in, by keeping the market open for more new players.

How to extract maximum value from the data?

Data is of some consequence only when we can make it work for all stakeholders. Liberal secondary use of data with innovative algorithms can create more and more use cases and thus immensely benefit the patients. However, there is a raging debate about data privacy and some quarters are calling for a crackdown on secondary use of data with tighter regulations restricting them. This could be a death knell for digital health in India.

Considering a diverse range of users, it only makes democratic sense to provide the freedom to people to choose between products/applications that offer free services by commercializing user data and the ones that offer absolute privacy but for a user fee. A small tweak between opt-in (secondary use only if user chooses) versus opt-out (secondary use of data done unless user chooses otherwise) in data policy can turn the tables around.

Tight restrictions have two-fold fallouts. First, it stifles innovation and creates stagnation in the system and second, user-charges increase as revenue streams from secondary use of data dry up and money has to come from somewhere. This double whammy will hit at the very heart of adoption of digitization that can completely derail digitization efforts in Indian Healthcare.

Data is clearly the common thread that runs across all the verticals of healthcare. Merely creating AI applications will not create the desired change. Indian Healthcare industry is run by incumbent commercial interests that are closely intertwined with existing medical practice and norms in the backdrop of socio-economic and political factors.

Data can have the necessary impact only when it turns into a currency that can create a market for new players and overcome status quo. Only a liberal data policy, of course with due regard for patient privacy, can enable robust secondary use of data, create novel business models and bestow data with the financial firepower to create sustainable changes in the system. Creating highly restrictive data policies can be disastrous with the vicious triangle of industry stagnation setting in.It is definitely the Data, stupid; which can make or break the industry.

(Dr Vikas H is Asst Professor & Dr Naveen Gowda, Resident Administrator, at Dept of Hospital Administration, AIIMS, New Delhi. Dr Atul Kumar is an alumnus of AIIMS,currently serving in the ESI Corporation, New Delhi. The views expressed are personal.)