MedTech and AI Innovations in Public Health Systems

Thank you so much. Thank you, ma ‘am. Welcome to all the ladies and gentlemen who have found time to be present here today as we explore the topic of MedTech and AI innovations in public health systems. There’s a lot of AI being branded about here. What we aim to explore during the session is in the public health care and with the three pillars that have been traditionally associated with the public good, public health being a public good. The cost of delivery. That public health care. scales, cost of delivery from the government side and also the cost of public health care for the individual also and what can AI bring in terms of having more cost effectiveness.

The second one will be on the care coordination and how do we ensure that the longitudinal health record get built and the clinicians are better equipped to utilize emerging technologies and AI for better care and the third one is on the efficiency, operational efficiency in terms of how do we ensure that the patient standing the line is treated in the best possible manner and in the lowest possible time with quality being associated. So with these three anchors to public health care, I welcome the panel and let me start with you Saurabh, my counterpart in the government of India. When we talk about population scale deployment of AI systems in health care, how do we ensure that the population is

Thank you, Saurabh. So I would like to inform all of you. Most of you must have also learned about the recent healthcare strategy that has been launched by government of India. It is called SAHI, the Strategy for Artificial Intelligence in Public Health. So as part of that, lots of activities in the field of AI are already happening. So if we see in terms of, we know there is a lot of lack of specialists, even especially in the rural areas. So through these AI techniques, the kind of services that are being provided, whether it is through scanning X -ray images or through diabetic retinopathy, also screening is possible through AI tools. So through that, in the resource constraint settings, we are able to provide good quality healthcare services to the citizens.

We have the eSangevani platform, the teleconsultation where… a person who is there in the PSC, a doctor who is there in the PSC, they can take expert opinions from the tertiary care hospitals. Also, I see artificial intelligence in terms of overall reduction of out -of -pocket expenditure because that is also one of the main important goals of ensuring universal health coverage. So by building more and more such systems, by bringing up trust, safety considerations in the public health system, we are actually also creating public trust in the public health system so that people actually come towards public health system and they rely less on the private health care and thereby reducing the out -of -pocket expenditure.

Similarly, also lots of digitization is happening, lots of records. We have the digital data. And through this digital data also, we can improve upon the overall workflows that are there in the hospitals. We have the resource constraint settings. So in terms of supply chain, management also. So lots of innovations. the health ministry is looking for in terms of ensuring that we will be able to provide the universal health coverage to a person even who is in a remotest of area should get the best quality coverage and also at the least cost. So that is how the strategy of government of India is as far as the adoption of artificial intelligence in healthcare. Thank you.

So you have talked about innovation emerging as a centerpiece in public health and with the strategy of AI adoption in healthcare, the SAI strategy we may look at a UHI movement just like there was a UPI movement where we have the digital public infrastructure in health being set up all the interface layers getting established but that also means bringing a lot of innovation ecosystem to the healthcare. So with the work that you have done Shivji over a lot of time how do you look at the innovation to institutionalization framework in the sense how do we while we every other day there is a health tech startup coming in, how does it get integrated in a structured manner with the public health system.

Thank you and good afternoon everyone. One of the important things which we need to recognize at least in AI is currently solutions are looking for problems not the other way around. Therefore it is important to marry what is the problem which is important for the state and then bring the solution together. So the first step of institutionalization is about how do you apply technologies and who sets the agenda and who is setting the priority. Like the way Andhra government has set up the Center for Applied Technology which has put out a call to say these are the problems we solve for our frontline workers. I think that’s the first step of institutionalization. Second is about the whole although I said solution looking for problem and problem looking for solution these are not either or.

I think both are important. We never knew we all needed a smartphone. But at the same time smartphone has become a problem now. Right in some sense. So I think continuous bridging and very critical element is to taking it to the ground and actually sharing the real evidence because every innovator will want to say their technology is fantastic. Have you come across any innovator who says their technology is not good? All of them will say it’s fantastic, it works the best, it is the best. That’s okay. That’s what an innovator is supposed to do. Whereas I think the state has the responsibility to test it on the ground, to look at the feasibility, to see does it actually change health outcomes.

Does it actually save cost, as sir was mentioning. And the third element of institutionalization, sir, is also the use case library that we need to build. I think there is a lot of discussion around this can do that, that can do this. Where’s the evidence? Where’s the use case library? Where has it worked? Has it worked with tribal communities? Does it work for the last poor woman, tribal leader or ordinary person? Right? That becomes very, very critical. And the last part is around the AI policy. Policy and processes. where the guardrails are built so that the state is also able to have a very clear policy towards how do we share the data, how do we ensure that we are able to, all the data that is shared by the community is actually monetized for them.

Thank you. You started with a great point that most of the time innovators come with solutions and they are looking for problem statements. But in Andhra Pradesh, we have articulated the problem statement clearly in terms of how do you at population scale drive an AI -enabled public healthcare system. And that’s where one of our partners is TataMD, which is represented by Saraswati Imam here. I believe you have set a fantastic stall also on the digital system that you have set up for healthcare delivery. So would you want to talk about your experience and how do you bring up a private sector ecosystem into public health and enable a public healthcare system?

Thank you, sir. As the sir mentioned, I represent TataMD. please visit our stall in hall number 5 where we have showcased what we are doing but I will just explain it in simple terms public health system we are looking at AIS assisting the entire public health system so I will just divide it into 3 or 4 aspects one is for the medical officer how can the medical officer gain from the assistance of AI so what we are looking at is how can the so normally when you go to a PHC the doctor would ask for the vitals to be taken the basic he would ask what is the complaint for which the citizen has come but it tends to be episodic it’s not longitudinal so we are looking at how with the help of AI we can share with the medical officer in a structured manner the entire longitudinal data of the citizen so that the doctor knows this is not just an episodic care we are talking about we are talking about continuum So how can we ensure that we understand the citizens?

So if a NCD patient comes, shows, mentions that he has HbA1c of 8, but has it been the same? Has it come down or is it increasing? So that trend will help the doctor to decide the medication. Or normally they would also ask what is the medicine that you are consuming and they would say either continue or stop. But with the longitudinal data, they will be able to say, okay, is this medicine actually working? Is it not working? How do I ensure that the patient is taken care of better? So I think it helps the data to be structured in a manner in which the doctor can use. Secondly, there are sometimes because of the busyness all of, I mean many of, I don’t know how many of you have actually visited a PHC and seen the workload that a medical officer faces.

Many times they are just rushing through the citizens, right? So they do not have the time. Sometimes they do not have the time. Sometimes they may miss an investigation which is required for a particular. So AI can do that prompt saying that, okay, this is the history, this is the data. Maybe we should get a foot examination done for this diabetic patient. He has not done it for last few days. So basically we are looking at AI as assisting the medical officer with the clinical support system so that nothing is lost, there is no oversight. Plus there is a evidence -based treatment guideline which can be shared with the doctor. And finally, the decision maker is the doctor.

We are not here to say that the AI will decide. The decision maker is doctor. The AI is to assist. So this would be more on the clinical side. Similarly on the operational side, right, all of us know the time that is spent in detailing out what the conversation with the patient is. So we are looking at how that can be made in a meaningful manner in a rural public health system. We all know in a closed room probably the listening can be better, the ambient listening can be easier managed. But in a closed room, we are not here to say that the AI will decide. We are looking at different dialects. We are looking at different dialects.

We are looking at different dialects. We are looking at different dialects. different contexts, how can we make that better? So that’s the second part for the clinician. Looking at the frontline workers, I think some of the AI bots, we are looking at how we can help them with their tasks. So if an ASHA is looking at 50 pregnant mothers, how can she prioritize who is the one she needs to look at, who is the high -risk mother whom she needs to prioritize? Because all of them are loaded with work, but AI can help them in scheduling their tasks, do their tasks in a better manner. And lastly, if I had to look at it from a public health system, the government, public, the health department, we are looking at AIs, how can it provide the analysis in a manner that it makes better sense for the government to see the trends?

How can the data show to them that, this is the… key problem in this particular area. We’ve been talking about with Andhra Pradesh government on creating wellness course, composite wellness course which looks at patients, looks at environment, creates a score which can tell them how to look at where the problems are and provide solutions. So basically this is going to strengthen the health department in identifying the risks, predicting the risks and looking at ways to do a proactive preventive care. So this is the way we are looking at ensuring that AI is providing support to all the stakeholders by using the data that is being provided and there’s a lot of deeper work while what I say is on the surface, the deeper work is how do you understand the data, how do you capture data across different geographies.

So that it’s more meaningful and it’s

I’ll come back to you in terms of the challenges you face while working with government, especially looking at the fact that government is probably implementing 25 odd public health care programs and the flavor is from preventive health care to maternal child health to genetic care and so on and so forth. But bringing a different stakeholder into the conversation, Mr. Seth, you represent a social impact organization and have been working on tobacco control. So where do you think is the maximum value of AI in health care driven from your perspective while you’ve heard the other people talk about digital platforms and enablements and innovation? What in your mind will be the biggest AI value generation for public health?

Thank you. Thank you, Dr. Mr. Gaur. You know, large public health programs, TB programs, prevention, tobacco control. adolescent health, the real question is where AI can actually help them day to day rather than in theory. And if you see most of these programs across states, the failure is not because of the design, because they’re not reviewed, but because of variable implementation across areas. Now, the data exists, reviews are also done, dashboards exist, but you find we very often find out what’s going wrong after the event when the failure has already occurred. I have heard so many senior level IS officers lament, the dashboards only tell me what I have not done. They don’t tell me what I am supposed to do.

Now, that is where I think AI can come in and add a huge amount of value. By, you know, helping and telling you where the failure is likely to occur. Identify where it is happening. Who has to take action on it? And then that action, I mean, you know, the person can be informed. And then, of course, you know, where the action is, that’s so important to pushing it. But for this, AI has to exist inside the delivery system, not on top of it. So, in my mind, that is where AI fits into the delivery system.

Typically, a PHC, a Primary Health Care Center… that supposed to see around 40 patients or doctor ends up seeing 60 patients, at least that is the statistics for Andhra Pradesh. You have all been to institution like AIMS where heavily what do I say, the fact that there is no care coordination or absence of care coordination, so everybody seems to be ending up in a tertiary healthcare setup. And there representing a tertiary care unit Dr. Rakesh and coming from AIG hospital, how do you augment, how do you do clinical augmentation for the doctors and in private healthcare, what are the lessons that have been learned and can be adopted in public healthcare also? Dr.

Thanks, Saurabh. In fact, I would start by saying, A is going to reduce the cost both in public and private health not by replacing the doctors but for the early diagnosis and intelligent triage. See, for example, as my co -speaker has said, this is a 1 .4 billion country. And as of now, I think it is going to be around $20 billion. And I think it is going to be around $20 billion. And I think it is going to be around $20 billion. And I think it is going to be around $20 billion. I mean, it’s growing day by day. So any hospital, it can be a primary, secondary, or tertiary care hospital, has got a huge volume of input.

And it’s very difficult for anybody, even a robot also cannot match the human scale. So these need -based innovations are something which we really have to look upon. See, suppose I, in my hospital, have got something where I have difficulty in doing it, and in other hospitals, something else. So need -based innovation we have to catch and then try to solve it. For example, in private setup, as I told you, there are some use cases which we had a personal experience since the last three, four years. I’ll tell you a little bit of economics on this. There is an algorithm which we developed with a pure AI model, costing 500 rupees to pick up a fatty liver, versus getting a machine which is 1 .2 crore and charging 5 ,000 rupees per piece.

So this is something which is a need -based innovation for me as a gastroenterologist. I’m a hardcore clinician, and I look upon any metabolic disorder which is the crux of the entire metabolism. And if you have tools like this, that will give you a lot of value in terms of fast diagnosis as well as the economics getting scaled down at your level. Then there are other use cases where you have the EMR and ESR. In fact, Sarabhiji, there’s a lot of chaos which happens when you have the admissions in hospital. So in that, we have a use case where the patient, they stand there, and the discharge summaries will take 8 to 10 hours for them to come out of the hospital.

So we have an AI -enabled system where the discharge summary will happen in half an hour max once I say my patient has been discharged. Vis -a -vis when you have electronic medical records and you want to have the patient bed management made. so one patient there’s a huge line where you start it’s all a personal experience in this hall everybody goes and you’ll be standing in the queue and even it’s not that you have to blame the hospital authorities but again that is something where in those areas you need these AI enabled systems so it can be digital health or a clinical related AI system so that’s where we have to concentrate on

so while you do that my question is again to you only go around the panel in a reverse order now and so while you do that and look at private sector efficiencies being coming out how do you think you can collaborate with state governments or governments at large in terms of the fact that you will be an early deployers of medtech solutions and the fact that you will have built it in your cost economics to use them faster how do you think you can accelerate their adoption in government ecosystem also

it’s a very valid point in fact the the see as a private sector person we have the early adaptation and adoption compared to getting into the public but in fact on that note i would say i couldn’t bring the aim foundation which we have which is working closely with the government of anupadhyay and other governments so what we did is we formed a platform with triple it hyderabad indian school of business iat delhi the fit and it is a neutral platform where anybody can come and then pitch their idea we handhold them nurture them and then make it validated at our clinical level and once we have the products for example the journey mitra which is launched in the government of anupadesh as the co -speaker told so that is something the asha workers can pick up with the a enabled system about the high risk pregnant mothers and then the nutritive value to decrease imr mmr so tools like this which we can do at our level and once we validate and we feel confident then we can give it to the public systems so there should be a public private integration, which is the main strength for this country.

And then only this will scale fast because time is running fast and nobody waits for us. And we have to keep up with that and then try to get the solutions because we cannot adopt the Western world solutions to us. Ours is entirely a different system. So we can never take any Western AI algorithm and then try to adapt. We have to have our own algorithms and we can do fast because of the population we have, the volumes we have, and of course the zeal we have.

That’s great. In fact, we are working with the AIM Foundation and then looking at setting up a biodesign lab in Andhra Pradesh with the AIM Foundation, with all the other institutes that you’ve talked about. And I see a lot of facilitation of deployment of Meta -X solutions happening through the CAT, the Committee on Advanced Technology. And the biodesign lab. But while we talk about all these Meta -X solutions, the core is something that… we believe as a state also that preventive health care has to be strengthened. And that’s where for prevention as an entry point, where do you see AI playing a role in terms of preventive health care being strengthened, Sanjay ji?

So I think prevention programs, as we all agree, prevention is better than cure and preventive programs will have the highest ROI. Unfortunately, preventive programs are not politically supported. Right. And that is where AI, I mean, if you take adolescent health, you know, student health, nutrition, you know, and I mean, if you take non -communicable diseases, we are talking about behavior change across entire populations, and that’s become the most important, you know, today maximum number of deaths are taking place is because of NCDs. That’s where preventive. So, you know, health comes in. Now where AI can really, I mean why AI really fits into this, because these programs operate at scale. They require continuous and repetitive activities to be done.

And they also show very predictable gaps during implementation. Okay, where the drop -offs are taking place, where the failures are taking place, and actually, you know, prevention, or sorry, the program implementation being done. And the number of variables are also very large, because as soon as you talk of behavior change, you are talking about, you know, huge amount of different cultures taking place. Different cultures respond differently to behavior. And if you take the mass of data, this is where AI can really support the programs and bring the, you know, not just the cost down, but the effectiveness of the programs to take place. But as I said earlier, AI needs. You know, I’m going to do this, and I’m going to do that.

And I’m going to do that. be within the delivery thing, not as a layer on top. And if you then, if we focus on, you know, how these schemes or programs result in outcomes, all right, this is where I feel that AI can give a very vast feedback. All these different entities, facilities, units are feeding a huge amount of program data coming in. AI can analyze where the likely failure rates are. It can escalate it to the appropriate level, bring it to the attention of the senior people, and that will result in far, far better delivery outcomes.

So can you be more specific in terms of, for example, the tobacco control program that you run, Sanjay ji? Right. Where is it, do you think that the fact that if we are doing it across, say, 20 ,000 schools and I do not know how many schools you are doing it with, are you able to do those kind of predictive outcomes? Outcomes or prediction in terms of where the program is. is probably bound to fail or is looking at a failure condition and the actions that need to be taken.

Oh, yes, we are getting, we are running tobacco control program in Antara, as you know, more than 20 ,000 schools. And each school is supposed to do a standard set of nine activities. All right. So very early, we are able to see which, you know, schools are not doing some certain activities. All right. And if we manually, if we start analyzing across 20 ,000 schools, there’s no way we can do it. All right. So AI helps us to analyze the data and say this block, this district, this area, there is a failure taking place. These schools are not taking action. All right. And then when we in terms of after the analysis, what we are doing is also the schools which are acting when they do the activities, they upload the activities.

Now we do image recognition. And decide whether the activity is done correctly or not. And we’re very high, 98%. accuracy we are able to see whether the activity has been done correctly or not and that tables enables us to give feedback immediately look within I mean as soon as he enters it within shortly after that he gets feedback you haven’t done this properly please repeat it all right then once we are talking about you know informing people so we are now sending out in Andhra Pradesh for instance 40 ,000 teachers get messages from us personalized messages you know for each person in the language he prefers in the language you know in the tone he prefers and that makes the motivation or the you know the way they act much faster than they used to earlier so we are seeing orders of magnitude improvement in terms of effectiveness of the program taking place I believe is you know works for I mean not in Andhra but in other states we are seeing this same thing happening across other programs which we have been working on

This is very heartening to see. So while for example you may be doing a tobacco control program with us in Andhra Pradesh, there is a cancer care program happening in Tamil Nadu that we got exposed to in one of the workshops. There are other states doing fantastic work. I saw Odisha for example, the stall today. So while and I probably picture to you Shivkumar ji that while we have all these islands of excellence and innovations, what is it that prevents them from scaling up and what are the structural barriers that probably government is not able to while we talk about ease of doing business, what is that ease of doing governance, public governance and public health care system that will make them scale up?

in place and the data quality actually improves. AI models really can’t work on top of it. And there is an exception in terms of, you know, you’re doing surveys and various other data points are there. Most states don’t have it. Right? And therefore, the processes unless they throw data out, I think we can all dream about AI, but really having the kind of value that we are talking about is going to be very, very difficult.

Thank you. Thank you. That’s a great point. And while we are at a certain maturity level in state government of Andhra Pradesh, there are other states which do equally well and there are states which probably are lagging. But with the national framework being put and with the ISHMA and Bharat Digital Mission, I bring it to Saurabh, my colleague. Where do you think in government of India, how do you facilitate all state governments to at least come on par and how do you see AI within the national health systems which become say gold standards or standards at least for all the other states to follow?

So as we know that health is a state subject, so ultimately government of India works in collaboration with the state government. And we understand that the kind of AI systems, the algorithms, the applications that have been developed for AI, ultimately the quality of output that comes out from those systems depends upon the data on which it is trained. And that is why it is very, very important that the data should be representative. It should be from every region because every region has a different kind of disease profile, every kind of various kind of demographic profiles. So this is very, very important that the data should be representative. Data quality, as was mentioned, should be very good.

And in fact, through this Aishwarya Bharat Digital Mission with more and more of digitization, we are and with every person now being provided with the ABAID, which is a… Actually, an ID which is linked with the health record so that the health records can move. with the person. So with all this digitization, with all the data that is being generated, we are able, we can do lots of usage of AI in terms of disease surveillance. We can use it for modeling of various diseases. We can use it for imaging. Lots of MRI because as I have mentioned earlier, still there is a lot of issue about the availability of specialist doctors, especially in the rural settings.

So at least if the AI solutions are available, the basic, at least 90 % of the imaging can be taken care of by the AI. So that only the most suspected cases can be referred to the tertiary care hospitals and the basic healthcare can be managed at the facility level. And as my colleagues have also mentioned, one of the issues in public health delivery is totally preoccupied with lots of administrative work. Lots of data entry, lots of portals that they have to enter data into the portals. That takes a lot of time beyond what is expected out of them, which is their clinical duties. So with this more and more, the AI application and more and more systems getting digitized, we can have a system where the data which is fed into one portal can be automatically populated all across the portals and the administrative work of these healthcare workers, which are our frontline healthcare workers, can be substantially reduced so that they can focus more and more upon the actual clinical work.

So AI is ultimately, it’s about improving efficiencies. It’s about improving the workflows. We have the supply chain management also. It is about optimizing of supply chain management. And in this entire journey, in the adoption of AI, we take states as our partners. Because ultimately, when both government of India and states work together, only then we can have a very robust AI system which can actually deliver quality care to our people.

We are here to work with the government of India very closely and establish those models. But the point you made, Kher, I actually cannot think of working as a NM myself also, despite being, adding a state health department. The sheer fact that a poor NM or MPHA male, the multi -purpose health assistant or the nurse on the field has to work with 25 programs. And while there is this national architecture coming up, there is so much of digital literacy challenge, not just digital literacy but adoption and using all these apps. This is a real challenge we face at the state level also. And with TATA, when we are building the digital backbone through project Sanjeevani that we are doing together in a collaborative approach, in an example of public -private partnership, I would want, Saraswati, you to play the devil’s advocate role and tell us the three key technology integration challenges that you see.

And please be critical of the system. But tell us that what is it that you would want to see. what are the challenges that you face day in day out when you look at building this care coordination oriented digital backbone for public health in Andhra Pradesh

Thank you sir, tough question to answer especially in a public forum but I will do my best, so one of the things like we have spoken about in a PHC there are lot of things to be done and like you mentioned sir, all of them have a lot of activities lot of programs, lot of reporting that they are doing introducing AI as something like Sanjay ji said as something additional or bringing technology as something outside is definitely a challenge, so our aim and what we have realized is if it is not integrated in their workflow, if it is not something that they find value in, the adoption is going to be a challenge like Shukumar ji mentioned that people are collecting data and just sending data But is that data really helping them?

Is it helping the people who are collecting the data? Is it helping them to do their work better? Are they able to benefit from the work that they are doing? If the answer is no, definitely they will not take it up. So one of the things is how do we integrate in the workflow? And they find value for what is being introduced. The moment that we are able to reach that sweet spot, I think they will start utilizing it. So one is how it’s integrated in the workflow. Second, I think, is it’s less of technology management and more of change management. Whenever there is something introduced, people look at it. I mean, there’s a cycle of adoption similar to what even we all face whenever we get introduced to anything new.

There will be a set of people who are ready to adopt it and are forthcoming. Then there will be a lot of people who are resisting. Slowly, the pattern changes. And people start seeing benefits. So what we are trying to… get at is who are those people who are seeing value for it who are those early adopters how can they bring in and probably with them we train the model like shikumaji rightly said if we do not train the model correctly you’re not going to get the good benefits so who are those people who can be utilized to train the model and who will not resist and then you give the trained model to people who are resisting so that they can see value so i think it’s a lot of change management related resistance which is what we need to address and lastly while i mean andhra is a very progressive state and we see this not as a challenge here but generally the connectivity the power availability all these tend to be one of the other challenges that for doing it a system wide change how these could probably be the this as i said in andhra thankfully those are not issues that we have seen but finally i think it’s about the incentives, right?

What are the incentives for people to adopt? If there are incentives for them to adopt both from the state side and from their personal side, the adoption tends to be easy. So, it’s a lot of work that we need to do to make sure that this is taken at scale, sir. Thank you.

So, I think we have a round time for one more quick round and I want to keep it short. Thinking aloud, what do you think? And the question is to all of you in the panel. What is the one maximum impact zone or maximum impact innovation that you feel based on your engagement with public health or with healthcare that should be happening? And I start with Dr. Akesh. What do you think will be the one most impactful thing that we can do? That’s a

very difficult question. There are many things to do. But what I would say in a nutshell is with the current scenario where we are, so the clinical insights from the doctors, the engineering capability from the bioengineers or the AI engineers, and the policy support from people like you, so this is something which will make the AI -related or metric -related innovation to go from the lab to lives. So that should be a collective holistic approach which we have to do and join hands together. I would say in that way that’s the need of the hour. Thank you.

I’ll make it simpler for you, Sanjayji. In preventive healthcare, which is one most impactful innovation that should happen?

Since I’m working in that area, I guess that is where I will obviously state, but if you look at it, Andhra Pradesh, 48 ,000 deaths every year because of tobacco usage. If you take any of the adolescent health, the future of our youth is how well our adults are. lessons grow. As I said, preventive is the highest return on investment, and it is not glamorous. It is very dull. It requires enormous amount of day -in, day -out discipline. But as a state, if you’re looking at what can really give you the maximum amount of benefit, I’d argue for preventive health. Thank you so much.

To you, Saraswati ji, in terms of engaging, and in the public -private partnership board, which is the most impactful thing that can be done?

I would probably respond slightly differently. I think in terms of bringing back the trust in the public health system, that would probably be the focus, and that hinges on quality of care that we are able to provide in the primary care, and that is what is going to ensure that the need for tertiary, secondary, and the disease burden that we are envisaging, that would probably… probably be managed if we strengthen the primary care with the trust in the public primary care. Thank you.

And Shivji, with you heading our committee on advanced technologies and doing all the work with innovation ecosystem, which do you think is the most standout innovation that you have seen that can be impactful for public health?

Sir, I’m going to be a little controversial on this. I think technology is just an enabler. I think our single biggest problem is going to be work culture. Nature. Work culture, incentives, and today every officer feels that they need to see a dashboard and tell their team what to do. I think if we have to really make AI help everybody decide, I think the work culture around evidence, the work culture around data is going to be the biggest one. But I will answer your question. The biggest innovation should be people’s data should be owned by data cooperatives. Nellore is a district. In Andhra, Nellore people should own the data through a data cooperative. and we should have reverse tokens where people pay for their data.

And we are feeding the AI engines and I think our people should gain from that. When we reverse that, sir, and when we reverse the incentives and the work culture of use of data, I think automatically you will find people coming and telling you this is how I am using it. Thank you.

That’s very interesting. And what about you? What do you think that can be the most impactful at a national scale also for the health innovation that can be there?

I would just also like to address the work culture issue that you have mentioned. In fact, if we can just sensitize, if we can make our doctors, our health workers confident that the outcomes are predictable, outcomes from the AI systems are good. And actually by adoption of AI systems, their productivity is improving. The kind of work that they have to do in less number of hours, in less time, they are able to use that. Use this. Use the same. Do the same kind of work. And they can do better in terms of their clinical approach and their productivity approach. I think. I think our health workers, they have adopted very swiftly to the technology. And if we can show them the reliability, the outcome that is certain, and overall improvement in their productivity, definitely workforce will adapt to this technology.

And as far as coming to your question, I think diagnostics will play a very, very important role in terms of the adoption of AI. And we are seeing it in the area of tuberculosis and also in diabetic retinopathy, where through the scanning of these images, the doctors can make a very evidence -based decision in a very less time. So in the same time, if they were seeing 10 patients, now with the support of AI, they can see 20 patients or 30 patients with much more accuracy. So the kind of shortage of doctors we have and the kind of patient load we have, especially in the tertiary. I think diagnostics will be playing a very important role.

huge role in the field of AI. Thank you.

I think that’s all the time we have. We still have time for one or two questions from the audience. If somebody would want to, the gentleman at the back actually raised his hand first or I spotted him first. There’s a mic behind you.

For mental health perspective. Because that requires additional safety, security as well as sensitivity. But I have not seen anyone touching on yesterday also as well as today. Mostly we talk about medical imaging and that takes because radiology as well as radiology all the innovation. there are developments also but I was thinking I will get some insight but so far

Dr. Rakesh you want to take it?

I think I have a point on that it’s a very nice question so on the mental health there are people in the western world who have got some apps and they are doing it but in India unfortunately there is no robust system to collect the data in fact if you have suppose you are working in a private hospital you have a robust EMR, EHR then you must be having a questionnaire on which you can build up these things but that uniformity to come it takes a little more evolution but there are people who are working on it in Indian sector probably it will take a little more time in fact you will be the one who can start that at your level

no we are working actually we are struggling I am at Eames Bhopal and what we have audio recording like we don’t have medical imaging either you have mental status examinations or video recording and based on that voice recording similar like detection of suicidal ideations, detection of depressions, anxiety and those kind of things. So there I was seeking if some assistance or guidance can

So I will probably just respond to this. What we have done in Andhra Pradesh is we have worked with psychiatrists and so there is a methodology called QPR, question, persuade, refer, which is actually proven in the sense that it is patented, where we worked with them and said that okay, especially with let’s say our students who go through high pressure, people who are in intermediate education, 11th and 12th, and there is pressure to perform an examination, parents are pushing them. So out of those 10 lakh students who are appearing for examination, which are those and our estimate is say around 15 % need to have special focus being paid. So taking all of them through this QPR methodology, working with this organization called Suicide Prevention Foundation of India, SPFI, and working with them, we have been able to at least look at which are those students who need to be given specific focus who are having those kind of ideations or having those kind of vulnerabilities and what kind of messaging needs to go for them and while it is a challenge and I am not saying there is a lot of AI into that because there is lot of what do I say privacy issue also associated with this but the other point you said about a scribe or essentially since people are talking you can actually get into the behavioral insight and understand whether what kind of ideation is happening and getting answers out of that I think that is a great point and would love to work with any innovator who would want to do it as a sandbox with us.

Thank you. One more question. Yeah. No, no, we will go to. You are an in -house person. Yeah, please go ahead.

So first of all thank you, Saurabh Gaur, sir. The first thing about that MedTech challenge I think this is the first time we have seen state government or government is opening up and telling that why don’t you guys innovate as come with your solutions including small startups like us. and we will give you a platform to pilot it and then finally help us in scaling up. My question more is to Saurabh Jain sir because we need to replicate something like this at a central government level. At the start -up, we definitely cannot go to all states and keep on doing pilot while MedTech is a segment where almost zero VC or private investment is there.

So it’s largely we running on either government grants or our own save money or loan or everything. So is it central government can create a platform which Saurabh Gaur sir or other state government can take those validated solutions and scale up them and we don’t just keep on repeating the same thing?

I think yes. ICMR is developing this kind of sandbox in which the start -ups can come up with their innovations. You can test it in the sandbox. So ICMR is actually developing this kind of mechanism to test the models. And ultimately, it’s about replication as you have mentioned. So once it is tested and it is tested across various settings depending upon the… outputs… Definitely it can be scaled up.

Thank you so much. The audience, you deserve a round of applause for being very patient audience. And I thank all the panelists also for their very valuable insights. Thank you. Thank you, everyone. There is a moment to give also. We can quickly hand over the moment to Dursu. Thank you. Thank you.