Notes

India's healthcare problem is one of scale, scarcity, and uneven quality

A common objection to digital health is straightforward: if you want better healthcare, do you not just need more doctors and hospitals? Indu Bhushan addresses this directly. The argument is not that digital health replaces physical infrastructure, but that it can improve the efficiency and reach of whatever infrastructure already exists.

India's healthcare challenges cluster around three problems. First, scale: any solution must work for 1.4 billion people. Second, scarcity: there is only one doctor per 1,500 people, and India needs at least 30-40% more. Third, distribution and quality are both uneven, with health facilities concentrated in certain areas and quality varying widely.

Digital health's potential entry point into each of these problems is through access and efficiency, not by adding new physical capacity.

Telemedicine and digital records expand access and improve quality

The most immediate way digital health addresses access is telemedicine. During COVID, e-Sanjeevani alone delivered over 10 crore consultations, many from highly specialised doctors. The platform has now crossed 43 crore consultations in total. This is one platform among many.

Beyond access, digital health changes how doctors diagnose and treat. Today, doctors often work from incomplete paper records. A patient may arrive without their full history, and the doctor makes judgments based on a narrow slice of information. With digital records stored on a patient's phone, the doctor gains access to a full longitudinal record: vaccinations, prior illnesses, co-morbidities, previous test results.

Indu Bhushan describes the change this creates:

"Earlier, doctors would be making their judgment or diagnosis based on a very narrow set of data. Now they have entire data, almost from your birth. They can see what kind of vaccinations you've taken, what kind of earlier health problems you had, what kind of co-morbidities you have."

This also enables interoperability. A patient seeking a second opinion no longer depends on one hospital releasing its records. The patient holds the record themselves and can share it with any provider.

One specific cost-saving mechanism is the elimination of duplicate tests. When a doctor can see that a patient already had a particular diagnostic test done, there is no need to repeat it. Repeated tests are a routine source of waste in the current system.

Hospitals, insurers, researchers, and policymakers all benefit

Digital health is not only a patient-side improvement. The gains extend across the entire health ecosystem.

For hospitals, claims processing is a significant operational burden. Today, hospitals file paper claims with multiple insurers separately, and have no reliable way to track where a claim is stuck. Digital health allows a single standardised form to be sent to multiple insurers, with real-time tracking.

For insurers, digitised treatment records mean better auditability. They can verify diagnoses, check what treatment was provided, and assess outcomes, rather than relying on paper documentation that may be incomplete or inconsistent.

Researchers benefit because health data locked in paper registers is practically inaccessible. Once digitised, that data can be used to study disease patterns, treatment outcomes, and population health. Policymakers gain the same advantage: they can track how programs are performing, identify which diseases are concentrated in which areas, and design interventions accordingly.

Indu Bhushan summarises it this way:

"Digital health is a win-win-win situation where patients will benefit, providers can provide better treatment, governments can have better policies, researchers will have data for future research, and insurance companies will have better and more auditable data."

Privacy, the digital divide, and the government's role in getting this right

Two challenges stand out as significant risks if digital health is implemented poorly.

The first is data security and privacy. Paper records stored in a register are, in one sense, secure simply because they are hard to aggregate and steal at scale. Digital records are the opposite: a single breach can expose data for an enormous number of people. Indu Bhushan notes this has already happened at major hospitals in India.

The second challenge is the digital divide. Roughly 40% of India's population does not have access to internet or smartphones. Any digital health solution that does not account for this group risks making the gap between those who can access care and those who cannot even wider.

This is where the government's role becomes critical. Indu Bhushan lays out three distinct responsibilities for the state.

First, infrastructure: the government must build and maintain registries with unique IDs for individual patients, health providers, and hospitals. These registries need to serve as a single, verifiable source of truth, similar to how Aadhaar works for identity. A patient's full records should be linked to their unique ID. A doctor's credentials should be verifiable by anyone.

Second, rules of the game: the government must set standards for data security, privacy, where data is stored, and how consent is obtained and can be withdrawn.

Third, universal access: the government must ensure that digital health does not become the exclusive domain of connected, urban populations.

The analogy Indu Bhushan uses is roads:

"Government will have to create the road and they'll have to create the rules that all the cars are going to be running on the left, this is the speed limit, and so forth. And then all the cars will come from the private sector and will provide the mobility. Not only cars, but also all the dhabas and services on the highway will come from the private sector. Government provides the infrastructure. The use of that infrastructure and all the services will be provided by the private sector."

AI can transform healthcare delivery, but requires Indian data and clear accountability rules

AI sits on top of the digital health infrastructure and, if the underlying data systems are built well, has potential to significantly amplify impact.

Screening is already being transformed. AI applications can now assess X-rays for TB at a scale that would be impossible for human reviewers alone. Similar tools are identifying high-risk pregnancies and screening for cancers. The productivity of scarce specialist doctors can be multiplied through AI-assisted clinical decision support. AI is also accelerating vaccine and drug development.

But the challenges are not trivial.

A major concern is that most large language model-based AI applications have been trained on non-Indian data. Indian populations, disease profiles, and health conditions are different. These models need to be trained on Indian data to be clinically reliable in this context.

The accountability problem is separate and equally serious. Many AI systems function as black boxes: they produce outputs but cannot fully explain how they arrived at them. If an AI-assisted diagnosis leads to a harmful outcome, who is legally responsible? That question does not yet have a clear answer.

There is also a regulatory gap. State governments are actively being approached by AI vendors selling health applications. There is currently no standardised framework for certifying or approving these applications, leaving governments without clear guidance on how to evaluate what they are being offered.

Finally, the equity concern applies here too. AI tools need to work across the full diversity of India's population, not just for the users who most closely resemble the training data.

Supplementary Resources

• India's Evolving Digital Health Strategy by Aditya Narayan, Indu Bhushan, and Kevin Schulman (Journal article)
• The Lancet Commission on a Citizen-Centred Health System for India by Vikram Patel, Anuska Kalita, Kheya Melo Furtado, et al. (Commission report)

The complete transcript file is available to download below.