On Episode 981 of The Core Report, financial journalist Govindraj Ethiraj talks to Dr. Mayur Shah, Head of the Rare Diseases Business at Sanofi India as well as Namit Chugh, Director at HealthQuad.

SHOW NOTES

(00:00) Stories of the Day

(00:50) EU Invites Canada To Become Its First Associate Member

(03:44)US War With Iran Has Cost $38 Billion Till August And Still Rising

(04:26) Oil Prices Ease, The Markets Rise

(06:00) Will The Tata Sons Ask N Chandrasekaran To Stay On As Chairman?

(07:02) How AI Is Changing Clinical Practices As Well As Hospital Systems

—

NOTE: This transcript contains the host's monologue and includes interview transcripts by a machine. Human eyes have gone through the script but there might still be errors in some of the text, so please refer to the audio in case you need to clarify any part. If you want to get in touch regarding any feedback, you can drop us a message on feedback@thecore.in.

Good morning, it's Thursday the 17th of September and this is Govindraj Ethiraj broadcasting and streaming weekdays from Mumbai India's financial capital.

Our top stories and themes…

Oil prices ease and India's markets rise.

The European Union invites Canada to become its first associate member.

The US war with Iran has cost 38 billion dollars till August and it's still rising.

Will Tara Sands ask N Chandrasekaran to stay on as chairman

And how AI is changing clinical practises as well as hospital systems? Two different viewpoints.

Markets, Oil and the Cost of the Iran War

The question most people would ask now is what does the world after Trump look like? Will the new trade and geopolitical alignments stay or shift once again and what does all of that mean for countries elsewhere including India? On Wednesday, European Commission President Ursula von der Leyen said she would work to make Canada the 27 nation union's first associate member, a concept which does not quite exist in the EC's current rule book.

Ms von der Leyen said Europe needed to tighten its cooperation with like-minded democracies, in particular Canada, to retain the continent's sovereignty and tackle threats from what she called an openly hostile world. She also said that we must urgently reimagine our partnerships and build global coalitions for our resilience and democracies. She was speaking in our annual address to the European Parliament in Strasbourg, France where Canada's Prime Minister Mark Carney was seated next to her.

She said that I would like to work with you on opening the door for Canada to be the first associate member of the EU. Canada has of course been inextricably linked to the United States economically and otherwise forever but with President Trump waging a relentless trade war against it and even referring to Canada as America's 51st state, all bets are off the table so to speak. Now this call by the European Commission President comes on the heels of a more impactful than usual BRICS meeting in New Delhi last week which saw India holding bilateral talks with China and Russia.

BRICS which was set up in 2009 also issued a fairly strong statement on the war in West Asia. Iran and the United Arab Emirates signed up to a joint statement from the BRICS nations that called for restraint at the Middle East war on Saturday with Iran's President Massoud Bezeshkian also meeting with the crown prince of Abu Dhabi on the sidelines of the New Delhi summit and this was also the first highest level meeting since the start of the conflict. Meanwhile oil prices were down with Brent crude futures at $107.64 a barrel on Wednesday morning following reports that Saudi Arabia was offering additional crude cargos via Oman.

The context is this Saudi Arabia earlier cut oil shipments after drone attacks damaged its key export pipeline to the Red Sea prompting top customers such as Poland to rush to seek alternatives as prices dropped $120 a barrel as per the Reuters report. Saudi Arabia had shut that pipeline which is the east-west desert oil pipeline which has in many ways provided it relief from the closure of the state of Hormuz or the relative closure of the state of Hormuz in the last six months. Meanwhile European diesel prices too were near record highs on Wednesday and on Wall Street U.S. treasury yields were lower on Wednesday as investors awaited the outcome of the Federal Reserve's two-day September meeting.

The benchmark 10-year treasury note yield was down more than two basis points to 4.97 that's below five percent while the yield on the 30-year treasury bond fell more than one basis points to 5.347 ACNBC reports said. Meanwhile the U.S. war with Iran has now cost the Pentagon an estimated $38 billion through August 1st and could cost another two to three billion dollars for each additional month of fighting according to a report released on Tuesday by the non-partisan Congressional Budget Office quoted by CNBC. More than half the cost has come from replacing missiles and other munitions used during the conflict.

The war has also consumed as much as two-thirds of America's missile defence interceptors since June 2025 mostly because of the U.S. helping defend Israel against Iranian attacks leaving the U.S. with reduced inventories for several years the CBO said. Increased flying hours cost another 10 billion dollars higher fuel prices added 2.7 billion dollars according to that CBO report quoted by CNBC and of course the war has resulted in fuel prices around the world including in India and that includes gas by the way to rise since President Donald Trump launched the war on the 28th of February along with Israel. With all of this in the backdrop oil prices being softer on Wednesday India's benchmark indices rebounded from a two-day fall the SENSEX was up 332 points to 74,336 and the NIFTY 50 was up 99 points to 23,217.

The broader market saw the NIFTY mid-cap and small-cap falling however at 0.01 percent and 0.18 percent. Now the rupee has meanwhile hit its lowest since late July including over expectations of a U.S. Federal Reserve rate hike in the afternoon U.S. time which weighed on the currency said Reuters adding that it closed at about 95 rupees 95 paisa per dollar after touching a low of 95 rupees 97 paisa its weakest level since July 27 gold prices were up one percent on Wednesday as the dollar weakened and oil prices slipped as we told you as investors once again waited for the Federal Reserve's decision on interest rate spot gold was up slightly to 4,346 dollars per ounce after touching a more than one month low on Monday as per Reuters. Meanwhile the government has eased export rules by waiving the requirement for registration certificates for shipments valued up to three lakh rupees a statement said on Wednesday a move that should help reduce compliance for small businesses and first-time exporters this will cover a large number of low-value export transactions and could facilitate exports through couriers postal and e-commerce channels.

Could N. Chandrasekharan Come Back as Tata Sons CEO?

The board of Tata Group's holding company that Tata Sons will meet today to discuss how to handle a listing mandated by the Reserve Bank of India and also discuss if outgoing chairman Natarajan Chandrasekharan should be asked to stay on according to several media reports. A last minute addition to the meeting agenda is a discussion of the nomination and remuneration committee's recommendations Bloomberg reported adding that the so-called NRC is expected to recommend asking Chandrasekharan to reconsider his decision to step down. The Reserve Bank earlier refused to relax listing rules for Tata Sons and Chandrasekharan said last month that he plans to step down when his term ends in February pushing the Tata Group to find a suitable substitute.

His exit also follows friction with Tata Trust's chairman Noel Tata over the listing issue and capital allocation across the Tata Group. Tata Trust has a 66% share in Tata Sons. How does a physician diagnose a condition they may never encounter in their career?

How has AI facilitating better Healthcare?

In India, up to 100 million people could be living with rare diseases like gorcha disease whose early symptoms often mimic common illnesses like dengue and this typically traps families in a diagnostic maze that can run into several years.

And despite this large footprint, 43% of Indian doctors have never seen a rare disease case according to French multinational Sanofi. This stark divide underscores a critical imperative for healthcare equity and finding a way to extend knowledge that's gathered not just from India but globally to doctors directly across the country. Now AI is the solution provider here fast emerging as the catalyst to bridge this gap going beyond basic administrative automation into active clinical decision support.

Interestingly, AI's role is also being felt across hospitals as more and more hospitals lean on AI to understand their patient records and their analysis of it much better. On the core report today, we have two guests from different walks of the medical spectrum to speak on how they are seeing the application of artificial intelligence from their vantage point. My first guest is Dr. Mayur Shah, head of the rare diseases business at Sanofi India who spearheaded India's launch of Accel Rare, an AI-led initiative to democratise rare disease expertise and put world-class diagnostic intelligence as the company says into the hands of every doctor across India.

I began by asking Dr. Shah to talk about the spread and incidence of rare disease in India. India's healthcare ecosystem backed by 35% electronic medical record or EMR penetration is uniquely primed for scale.

INTERVIEW TRANSCRIPT

Dr. Mayur Shah: This is one area which has often been taking a backseat versus some common disorders, but individually is rare, but as a whole it's not rare. And why I'm saying so? We have around 7,000 rare diseases collectively, and in India 450 of them have been majorly defined.

You know, as we speak there are varying definitions of rare disease. Just to talk about the prevalence, the rare disease prevalence has been defined anywhere between 1 in 5,000 to 1 in 10,000 population, and that's what we are talking about. So when we talk about rare diseases, some or the other times these actually get masked behind common symptoms in science, and which leads to a delay in diagnosis.

And that's one of the major starting points of problem, that there is a delay in diagnosis. Just to give you a figure which was academically quoted in the literature, an average duration of delay in a rare disease diagnosis is anywhere between 4 years to 7 years. And that's what it means for a rare disease patient that I don't even know what I'm dealing with.

And especially in a complex healthcare ecosystem like India, there is a variance in terms of also the capabilities which are required from a healthcare facility perspective to identify, even have a suspicion of rare disease. So, you know, that's the kind of a problem that we are speaking about, but if I apply this to India cumulatively and collectively, rare disease might be something which will be present in more than a million population within India, collectively, if put. So that's the kind of solution that we will have to address too, if we look at this challenge.

Did you say 1 million or 100 million? It can range from 1 million to 100 million, in fact, and that's the kind of a disease bit. But as we speak, there is no such, you know, registry-based strong data and evidence that we have and which is what the government is also trying to tackle along with some of the other institutional organisations.

Govindraj Ethiraj: Right. I know that by definition, they're rare and therefore not prevalent, but what would be the top 2 or 3 rare diseases that at least have been seen in India or noted in the last few years?

Dr. Mayur Shah: So basically, I would say the government did a very nice job of bringing out a policy of rare diseases, wherein they actually quantified and registered 63 notifiable diseases within that. Just to give you an example, lysosomal storage disorders, okay. These are one of the common rare diseases.

There are other rare diseases which are there in terms of genetic disorders which are there. All of these rare diseases are predominantly, if I were to tell you, 75% of them actually are seen in children and that's where it's again an important issue that we are dealing with because it adds to the complexity. Also, the fortunate part is if we are able to diagnose on time, some of these people have gone ahead and we have examples of that, have gone ahead and lived normal lives.

Just to give an example, I was saying lysosomal storage disorders, they have a deficiency of an enzyme and if you identify the enzyme deficiency at the right time and you provide the deficient enzyme to the individual, the individual technically leads the normal life. Right.

Govindraj Ethiraj: If I were to ask the question the other way, let's say there are, as you said, 65 notified disorders or conditions, what would they predominantly be in the area of?

Dr. Mayur Shah: These are all genetic disorders predominantly and they would be in the area of across different systems. They can impact gastrointestinal, they can impact blood disorders, they can impact even developmental milestones etc. So, these are spread across different, you know, disease areas and therapeutic areas but then predominantly as a whole, when you look at them, they are genetic disorders to the large extent.

There are some of the other disorders also but because they are lesser in number, they are called as rare. Right.

Govindraj Ethiraj: What is the solution that you have worked on or are testing right now and which also uses artificial intelligence and how is that playing out?

Dr. Mayur Shah: So, as I was referring to my earlier conversation point that rare disease, one of the biggest starting point of challenges is delay in diagnosis and even if I dissect that further, the first point is suspicion. I mean, if it's not even in the suspicion criterias or a consideration set of a doctor when thinking of a differential diagnosis, that's where Sanofi actually came up with an AI-enabled solution which is called as Excel Rare which is an AI-powered pre-diagnostic solution which basically takes around 5 to 10 minutes for a doctor to put in their difficult cases which they have not found a confirmed diagnosis to and once they put the signs, symptoms, some of the test results etc., what they are provided is with a differential diagnosis and the differential diagnosis is like a supporting tool which throws up 5 to 7 potential differential diagnosis of rare diseases which probably might have got missed. The kind of accuracy that it shows is 89%. I mean, 9 out of 10 times it will show one of the confirmed diagnosis out of those 5 or 7 potential differential diagnosis and that is something which is easy to be used across computer or the phone etc.

because this is a web-based tool. So, that's the benefit that it gets to the nooks and corners where the doctor potentially has a patient, is struggling with a difficult diagnosis and can throw a suspicion.

Govindraj Ethiraj: And the database that was used to build this, is that India-specific or is it global? And secondly, do doctors have to pay for this or is it a free access?

Dr. Mayur Shah: So, the database which this uses is currently global. However, more and more the usage starts increasing across the world. They also have a mechanism of updating this data and that's where the more and more that we use in India will be able to build in more and more cases from India which can then again throw back relevant data.

Number one, this has been validated by 13 rare disease networks across the globe. And to your second question, this is a web-based tool, completely available free of cost and it is specific to be used by healthcare practitioners because this is the kind of details that anybody would require would be possible for a healthcare practitioner to provide that.

Govindraj Ethiraj: And conceptually, if this is how you're approaching, let's say, the diagnosis of rare diseases, could this be extended to other kinds of conditions, diseases?

Dr. Mayur Shah: Technically speaking, the objective of bringing this tool was to specifically focus on rare diseases. There are other tools available in the market which looks at general conditions but when you look at them, you tend to really miss out on the rare diseases. So that's why I said that if there is a difficult diagnosis which is not being able to address within the first, say, 8-10 weeks or 3 months, you have still a patient who has not been able to reach a confirmed diagnosis.

I think that's the time to bring rare diseases into the considerations set rather than waiting for a long time. So it's a choice between whether you want to be specific with the rare disease differential or whether you want to keep it open and then probably this gets missed out. So the decision has been to stay focused on the rare diseases.

Currently, this tool addresses 310 rare diseases and the number keeps on growing up year on year.

Govindraj Ethiraj: Last question, I mean, according to, I guess, or rather for parents will be, I'm sure, listening or watching. You said it takes 4-6 years sometimes to diagnose. Now, at what point should, let's say, a parent, because since you said 75% are children or parent or even for that matter an adult, go and test for a rare disease diagnosis?

Because obviously for every condition you can't be saying, okay, I want to run this past a rare disease diagnostic template.

Dr. Mayur Shah: So a fair assessment, the general rule of thumb is as early as possible. But a fair and a realistic assessment could be somewhere, as I mentioned, if we have not been able to reach a confirmed diagnosis through the battery of tests that a doctor highlights, at least in a span of 3-6 months, I think it's a fair time within the India ecosystem to consider that there could be a potential rare disease which is getting missed out behind the common symptoms and signs that I'm seeing.

Like classically, you know, Hepatospinomegaly, right? It has multiple reasons that a child has a bulging abdomen, protruding abdomen and there could be a large spleen and a liver, which is common. But it's important to rule out the common disorders first and which generally takes, in an average setting, 3-6 months for a doctor by and large to rule out all the acute and potential chronic cases.

But even then we are still not able to pinpoint a confirmatory diagnosis. I think that's the time we should think about that. Am I missing something?

And consider at least that potentially there could be a suspicion of a rare disease. My point is, have a suspicion. Then the rest of the things can follow.

And the good part is this tool also, you know, just doesn't give differential diagnosis. It also gives potential tests that are required for that differential diagnosis to be ruled out. So A, it enables the doctor.

It does not throw one single solution which can challenge the doctor's assessment. It enables and supports the doctor. And B, it also gives a battery of tests which can help you reach there so that you are able to optimise the right tests in the resource constrained setting.

Govindraj Ethiraj: Right. Dr. Shah, thank you so much for joining me.

Dr. Mayur Shah: Thank you so much.

All cheers for rare diseases. Let's ensure that we reach out to the last patient who needs it within every nook and corner of India.

Govindraj Ethiraj: Let's wish for that too. Thank you.

Dr. Mayur Shah: Thank you.

—

A new joint report by Bain & Company and private equity firm HealthQuad which invests in healthcare companies has said the report titled AI in Indian healthcare delivery says that most providers are still testing AI in control settings and the immediate challenge lies in moving from operational pilots to deployment ready clinical tools especially as generative and agentic systems rapidly double their capacity for expert level work.

I spoke with Namit Chugh, director at HealthQuad, the private equity fund and I began by asking him how he was seeing the role of AI in healthcare systems including where they were investing right now.

INTERVIEW TRANSCRIPT

Namit Chugh: One of the key aspects of our report was to see what's happening on ground because it's such a vastly spoken about subject that sometimes when we talk about technology, we often get caught up in just the models and the breakthroughs and the benchmarks rather than looking at its real on-ground adoption. So in the Indian healthcare delivery context, what we have realised is that ambience cryo is something which has started to see a lot of adoption. Just to simplify it a little bit, hospitals have started integrating ambience scribes within their workflows so that the entire documentation process around consultation, so that whenever the doctor is having a conversation with a patient, all of that can be automatically scripted and they just need to have a quick glance and approve it.

So that is something which ambience listening plus scribing does together and that has seen the maximum adoption. What it also leads to is saves at least 30-40% time on account of doctors which they can use to see more patients or have a more meaningful present conversation with the patient rather than thinking about writing a consultation post the consult. So that is something which I feel has had the maximum impact.

It is also very easy and flexible to use given doctors just have to do what they do best, which is see the patients, talk to them and everything gets automatically picked up using artificial intelligence.

Govindraj Ethiraj: And when you say ambience scribe, can you define that? Is this a specific product or is it a generic?

Namit Chugh: No, it is not the name of a product. It is just a feature that essentially ambience setting is something which means that there is something happening at the backdrop. So it is not interfering with any conversation and scribe essentially, as the name suggests, is scribbling and writing.

So when you instal a feature on the laptop of a doctor or whatever EMR or EHR they might be using, someone is just listening to that conversation ambiently and that is AI listening for it. So it is a conversational AI feature. It listens to the conversation, drafts the summary, converts into a proper note and then flashes in front of the doctor once the consultation is over.

So the doctor or the nurse can just have a quick glance and make sure because given the vernacular languages, given our specific diction, our pronunciation, sometimes AI can make an error. So just to make sure doctor can have a glance and approve or change a few things and then there you go.

Govindraj Ethiraj: Right. So this refers to the conversation within a clinic or in with a patient. So is that the main area that we've seen application of AI or is there anything else outside of that, including inside or the operating theatres or other parts of the hospital system or medical system?

Namit Chugh: So maybe if I take a step back, what we had done is in this report we tried to dissect the entire patient journey from their appointment, booking, scheduling, doctor visit, inpatient admission and then as far as their post discharge journey. So within that patient journey, this was one use case which was seeing the maximum adoption. As far as claims management is concerned, which is again a large problem statement in the Indian context, a patient gets a clean shirt from the hospital at let's say 11 in the morning, but it still takes four to eight hours to get the patient discharged from the hospital because of the processes involved in the back and forth.

So that is one use case where again we saw that a lot of agents have started to work. So that's an application of agent-picked AI where agents are starting to work and see all the documentation, the process flow, day-by-day summary of the patient and then documenting it in a fashion so that once the documents are sent to the insurer, there's no manual intervention. It's done by the rules, by what is approved.

So that is another use case. Both of them are seeing adoption and then there are a bunch of things happening on the procurement side, inventory management, supply chain, finance. So those are some back-end use cases, but the purpose of our report was to sort of focus a little bit more on the clinical aspects of a patient journey.

But those are four or five things which have seen a lot of action.

Govindraj Ethiraj: Right, and you talked about electronic medical records or the conversion into electronic medical records which is now happening in India. So what's the adoption rate like including amongst the hospital chains that you're either invested in or talk to?

Namit Chugh: So EMR adoption at the India level itself is 35 percent while in western counterparts or western healthcare it is close to 90 percent. So that itself is something which is very unevenly distributed as far as Indian hospitals are concerned compared to globe, which means that we don't have the backbone of digitisation in our healthcare system which makes application of AI a little tougher at this stage. So with the ABDM guardrails coming in, this will ease out in the next few years.

And just to clarify, at HealthCore we don't invest in hospitals. Ours is to do health tech, med tech, biotech and novel healthcare delivery. So slightly more newer areas of healthcare where technology is either playing a core role or is an enabler.

So hospitals is something that we don't invest in. But because of this report, given the biggest consumer of AI in healthcare is going to be doctors and nurses. As a part of the report, we spoke to dozens of CIOs and hospital CEOs to get a sense of what's happening on ground.

Govindraj Ethiraj: Right, and as you look ahead in the universe that you invest in, which you said is not hospitals, what are the kind of AI applications that you're seeing which are interesting and of course, relevant to customers or patients more specifically?

Namit Chugh: So remote patient monitoring is something that we are really excited about. We see a real potential. We've also done an investment in the space in a company called LifeScience, essentially.

But just to walk you through what that company does or what the business model is, it's sort of using a variable device, collects patient vitals when the patient is admitted in a hospital and at the back end, an AI engine ingests all that data and then uses analytical insights for the clinician to predict any deterioration like sepsis or cardiac arrest or respiratory distress, which is not possible using a naked eye for a doctor to make a decision on.

So this is helping the doctors avoid any last minute changes in the patient's health or deterioration in the patient's health. So this avoids code blues. Similarly, what we've also seen, a couple of other interesting companies in the ICU space, because India is so short of intensivists or ICU specialists, can we use AI models to play some of that role by looking at patients closely inside of the ICUs?

But operating theatre and ICU is something which is slightly more clinically complex. So this is not something which is commercially scaled as of yet. A few hospitals have started using it.

So I think it will take some time, maybe 12 to 18 months, where all these business models start to make more sense from an investment standpoint.

Govindraj Ethiraj: Thank you so much for joining me.

Namit Chugh: Absolutely. Thanks, Govind.