 Welcome all, maybe more people will be joining in a while, but since it's live streaming people are waiting outside Zoom as well. So to introduce today's topic of discussion, we intend to talk about India's health data infrastructure primarily on how India has been investing on the digital side of healthcare and what has happened so far, we have with us Ding Wong. Ding is a post-doc fellow and researcher at Microsoft Research. She works at Bangalore and she has been doing some ethnographic study in terms of how technology design is being deployed in hospitals in India. And we have Sudeep Vaidya, Sudeep is the co-founder of credible health. Sudeep is a master's in biomedical design from IT Madras. He has been working on AI solutions and healthcare over the years. And we have Aayush Ratty, who is a policy officer from Centre for Internet and Society. CIS has been working on the policy side and the law side of healthcare data. Today's topic of discussion, as I was mentioning, is primarily around the national health data blueprint, which is what was an evolution of the national health stack which was proposed to the government through an EPIO by industry tech group Iceberg. To give a brief background of what has happened with the healthcare infrastructure, the idea to push healthcare data started with the electronic health data records, EHR as referred to, somewhere in 2013. 2013 is the first time when electronic health records as a draft was notified and recognized when it happened, finally, because of interest in healthcare industry as well as some interest and push by the World Bank Group and the WHO. And EHR electronic health records have been modified further in 2016 and we have the proposals for national health stack and currently we are at this enterprise architecture system as the government is referring to, to the national health data blueprint, which is a sub-system of a giant system called the India Enterprise Architecture. So health is one component of it and part of the health data blueprint, the government wants to build a health data ecosystem. So you will have health data exchanges. So people is going to talk more into that. We'll start with Ding and her technographic study in Indian hospitals and how technology has been used in Indian hospitals. Hello, let me first unmute myself. So what I'm going to talk about today is going to be the fieldwork that we had conducted in a large Indian hospital chain. So as three of us has introduced, I'm a postdoc researcher in Microsoft Research India, so this research is not specifically linked with the health stack or other initiative, but we thought it would be good to share some of the learnings from the ground to give you an idea where what is the status quo that's happening within a hospital in terms of the digital transformation despite all the hypes on all of the data and digitization. So within the, for those who don't know about this, ethnographic study is a type of sociological study that through a close look at, I guess, the environment and the context and specifically the work that goes around using observation and interview, we get a very detailed sense of what's actually happening in the ground. So through this study, what we find is that the hospital is very invested in, you know, electronic patient record or health record that they are not only so they have more than one system running at a moment. They have existing one that they are using and they are also investing in getting a new and updated one there. So one of the things while we were doing this study that's particularly fascinating is that we also find on top of the use of this formal organizational digital tools. Specifically, the group that we observe or studied our nurses, they're also using a lot of WhatsApp to walk around the tool that this organization is using. For the reasons that the tool, the organization is using particularly the electronic patient record is not particularly user friendly to say the least. So while we were conducting this research, you know, this is a very large hospital. They have many patients come in daily and they also have many hospitals across different geographic location as well. So using electronic health record on patient record is a value to this hospital because it's a very structured communication tool, not only just between different departments within one hospital. It also helps to set a structured data and information across the hospital chain as well. And, you know, like any other electronic health record it is designed with multiple user in mind. So, you know, it's the system that's slightly different from what we understand as a work tool or personal tool, you know, it's not one laptop or one desktop per worker. Many of them share a terminal or share a desktop that's within either with a nurse station or some other location within the hospital. So that is something, you know, the electronic patient record have considered in mind, but still there's some dissonance between what has happened. And well, we hope that this sharing the lesson learned today as well as, you know, will we report back to the hospitals to foster rethink about some of the technology design that we are implementing in the ground and how does that affect the people who are ultimately using them. So, currently the electronic health record is not the best system to use from what we can see or will learn from the field. It's incredibly time consuming to use largely due to the user interface design. So for example, the search function in this health record is particularly bad. Like it takes them quite a lot takes nurses or the pharmacist a long time to just search for a patient and then be able to update information about that patient. And also, because the system is rather rigid, it doesn't support in time communication within the tool, and it's often doesn't support dynamic changes. What's happening to it is that it's being used as a data repository. What do I mean by that is that, you know, as we are imagining things happening in the hospital, things happening with the patient, the nurses goes around and input to the doctors, noting to the system and sometimes input to the prescription into the system. It's often done with a delay. For one, that's because the nurses work, you know, have to be centered around taking care of the patient, then later on they can deal with the documentation of it. And secondly, it's not a dynamic system that you can do things as things happen. So that means, you know, what is supposed to be a dynamic system is you used very as a static data repository. And this happens not just, you know, within the word, but also during the times when patients being admitted to the hospital. Again, when when patient is admitted to the hospital, it's yet another very dynamic process that involves many players coordination to make the things happen. Whereas this system itself doesn't really support that many users to use it at the same time. So, you know, if we think about what really is what given all of this thing says, is it a really good system to use? I do think so. So electronic house record is this one step, I guess the very first step the hospital has into digitizing the workflow they have. And it is useful for the hospital, particularly if you think about what it does is it makes the billing system, it makes the billing process a lot easier for the hospital and also for whoever involved in making billing work in the hospital. And because that another thing I have to say that I think particularly have been done well with the electronic patient record is how it has been connected with the insurance system. That that again makes this billing system a lot more smooth for a lot of us who have the house insurance that you can just go in with a cashless process that also you know saves time for the patients as well. However, if you think about because it being such a rigid system and people have to work around it, what do they what they did was that then they use the personal tools to basically fill the gap. This tool, this electronic house record doesn't fill the tool people went to is what's up and what's up, you know, as we know is this communication. Well, chat app that has been adopted across the globe in many different work contexts and as well as within the healthcare context. And in the US and UK because of HIPAA and the NHS regulation as well as GDPR it prohibited to use what's up in in healthcare settings altogether in hospital by nurses or doctors. However, in India, such regulation doesn't exist yet and similar things in China so people turn to this personal tools and repurpose them for work. And I have to say from what we can see it's very systematic use of what's up to get the work they need to done, they need done done. However, to relate back to what we're talking about today, is there issues with using what's up to walk around those formal tools. Of course there are. For one, we're using a personal tool for work. There is no longer this clear separation of what's work what's not work if you're using personal tool or personal device for work. And secondly, if we're thinking about a slightly larger scale is that whatever been discussed on what's up as information of this hospital of the patients who goes into this hospital cannot be saved somewhere. Never mind the security and you know confidentiality as well as privacy issues that's embedded in any chat tools. But also as organization they're losing out on some information that hasn't been captured within the formal work tool, formal communication system that just lost within this informal tool. And even if you know we move this organization to adapt to adapt some organizational communication tools you know for example things that Microsoft could provide or what not. That's there's still this caveat that because the electronic patient system is one on its own standalone tool and whatever organizational communication system the the doctor or the nurses use is another system. There isn't much of a data going between the two of them. That still means even if they use an organizational sanction tool there is going to be this data distance between different systems. Never mind there you know multiple systems within a hospital the legacy ones the one they're currently using things they haven't been migrated things on the paper. So basically this is what we're dealing with within the hospital you know despite all the hypes on the digital. I mean the data on health care and all of that it is still not quite there yet. This is both because of the infrastructure the work processes usability and also functionality of all the tools that we involved within a hospital. So I think with that said what I'm trying to set is that I think I think potentially and ultimately we could benefit from having data about specially help data at one place one place if it's secure if it's anonymized to understand a public health at a scale which I think is a good thing. But we're definitely not there yet on the ground. We are using various ad hoc systems and tools to get a work done which is very important at the moment and you know this COVID-19 only stressed on that to deliver basic work is very difficult with limited resource. Never mind this flamboyant talk about you know data and the health stack and all of that. So this is all from me. Thanks Ting. We're going to move on to Sude who has been actually working and volunteering on some of these issues. He has like a biomedical engineering background so he can also tell what is the actual engineering procedure that you should be followed. And how should one develop these systems? Sude. Yeah. Thanks for inviting me to this talk. Before I start talking about the data blueprint and different aspects of it let me just try to give you a quick background of what I do outside in my day job. So over the last about four years or so I've been the co-founder of a company called Pratable Health where what we do is basically build artificial intelligence algorithms for radiology imaging. And so what we're basically doing is you're trying to get machines to learn to read say X-rays, CT scans, MRI scans and so on. The overarching idea being that there aren't enough radiologists who are these specialized doctors to read images enough in the country so if we can augment their abilities using artificial intelligence then possibly we can increase the number of people who can get diagnostic care. Now a lot of my perspective around the data blueprint and how it should work has sort of evolved from a very research or AI perspective and obviously also a deployment perspective of how to get data-based systems out in hospitals today. Primarily because the way we have operated is that obviously when we started one of our biggest challenges was how can we get all the data that we need to build these algorithms? Because we are talking about sensitive healthcare data and these are not available in the open source and even if you get data from the open source they're probably not going to look like what they would in real life. So we needed data on day one, three years for 2016 when we started we needed data so that we could build these algorithms and we went knocking on almost every hospital that's hospital and hospital group and diagnostic center in India going and having conversations with them, hey can you share data? And in the last three years it's been a very interesting journey that we've managed to get, we've obviously gotten rejected by several institutions but we also managed to sign agreements where we actually managed to get a lot of this data. And I contrast this experience with mine with what it could have been if there was a technology infrastructure layer which could have provided access to startups like me building artificial intelligence to have data accessibility not as such an issue that needs to be solved over the course of one year or so but maybe in a few weeks or a month. And so that's sort of a quick introduction about the background and hopefully that helps give you, gives you some flavor of the perspectives that I will share. And so just to give you an idea when I write about the national health data blueprint and I started talking to some folks from my spirit and so on. The general idea that I got which seemed exciting to me is that you know is the entire concept around how data today is in silos within hospitals. Nobody has access to their own data. And if you go to another hospital today, you don't have access to, there is no easy way to provide them access to your data which is stored in some other hospital records. And it's either a paper based record which is very difficult to get out or you have to get them written in CDs, DVDs, it's a nightmare and standards are not easy to come by either. And so if you look at this from the view from the lens of what our banking system looked like before UPI and all the interconnectivity happened that exists today from core banking to UPI to what we have today. I think there are some battles that that I could see which could potentially bring be brought to the entire healthcare data ecosystem. And that's that's that's primarily you know one is can is it possible that patient has direct access to his medical records irrespective of which hospital his data is saved. Can I provide access to a new hospital instantly without having to go through all the hassles. And can I do this in a way that does not involve me, you know, involve the state having to aggregate all the private players coming having to come together and aggregate all the data in one center centralized repository but but do it in a federated way similar to how maybe the how UPI acts today that the data is still stored within the bank servers and you have a and you obviously have some fiduciaries in between who sort of enable this access. But can you have a similar system because healthcare data can easily get very large. So it's not just you know very transactional data or text data that you need sometimes just heavy images and signal data. It's not that easy to share or aggregate. You know, even if you wanted to. So what is a way that you can build it so it can be federated and and and and most importantly, can you move the consent of sharing this data away from the hospital to the patient. So today, you know, for instance, in my in my case, even though I was using so many patients data we in fact today are using so many patients data to develop our algorithms be the core. The discussion we have about you know the ethics clearances the commercial terms of how and how we use this data all happens with the hospital and the patient usually never involved and because it's mostly retrospective data. Consent is also most likely most often made off is because there is no intervention or adverse event to affecting the patient at all. And so this this you know this is similar to you know, if you think about it, it's taking it's grieving the entire responsibility of the data and the opportunity of commercialization and and and figuring out all the terms around the data with the hospital. Instead of the patient, which which again I think is is fundamentally not the way it should, if you think about the system where you want in data interoperability, the data rights and who gets to access this should not be ideally by defined by the hospital to be directly by the patient. And so so basically, given this entire context of how for the success we have around being able to integrate the entire financial ecosystem through API and being able to bring interoperability. The idea was is that is there a way that a similar kind of a data intra connectivity model could be established in the healthcare ecosystem. And it's a much much more difficult problem. If you think about it, because, because one, you know, EHR, because one the data standards are not established. So I think, like I said, I think it was put out. EHR standards came out several years back. But but the implementation of it is still challenging. If you go to any standard hospital today you will see that most of them know I can give you a large large person maybe 90% 95% of them do not have do not have digital records, they still use and written records, very, very minimal limited information, only the information that is required for billing is usually just entered in the HR everything else is still managed on on some sort of records newer hospitals are increasingly increasingly entering but but you know you have to understand the Indian healthcare ecosystem is very different from from our counterparts in developed countries because they have time to see every patient here you have 30 seconds for every patient so doctors don't have the time to do administrative work. So so EHR even though standards exist you have a fundamental problem the data does not exist digital data does not exist you need a way to be able to collect collect this information, you know, from the hospital get doctors to enter it in the first place. And the second part is, is that you know clinical notes is just a small tiny portion of the entire ecosystem of data. There is far more data that that exists beyond clinical notes you look at radiology imaging for instance you look at ECG, you look at other you know sort of temperature blood pressure, there are just so many so many biomedical data vital science that can also potentially be used. And, and, you know, can give good information on on the patient record, whenever the patient needs to go to another site or another second opinion, or even come back for follow care very important information. So, so it's not just about clinical notes that you need to talk about and I mean it's not just about the EHR standards, but it's rather standards for all the different kinds of data, you know that that you can talk about, you know, starting even thinking about your variables, your home health monitoring devices. Now how can you get all of this data ecosystem into into a centralized place. And, and is no while while a lot of the hospital data has you know has some place where you can store them maybe the hospital infrastructure. No, if you have home health devices you have, you know, other kinds of devices that you enter it how do you how does it come into the system, you know all that are a sort of questions that need to be answered and potentially opportunities that probably vendors can can come build around the ecosystem itself. And, and, and you must understand that, you know, the EHR system fundamentally came up in developing countries, primarily because of the strong incentives around insurance and payments that were linked to it. And, and it's no, no way if you think about it's no way the most optimal system because it's, it's, you know, it's, it's rather you talk to any physician in the US and they will have nothing good to talk about the MR it is very clunky interface like being just just talked about, but also, you know, it's, it's, it takes a lot of time, you know, doctors always almost have, you know, backlogs of patient records and data to enter and then it's not so very fundamentally there is a reason to go back in question why do we need EHRs what is critically important in these EHRs that we need to collect the raw data in many cases available do we still need the analysis and the reports to be recorded in the report itself and and how do we, and how do we rethink from first principles what all we need digitalized and digitize in the data in the first place. So I think, I think that thinking is very important for the Indian ecosystem because very simply just coming down and saying every doctor every hospital should start using an ESR system is very impractical and most doctors will tell you the same thing that they're still they're just not going to implement it it's not interoperability it's not the technology that that and the interoperability technology that's limiting, you know, the adoption of ESRs but rather know the workflow of the of the software itself. But but let's let's assume that problem is sort of solved and you figure out a way to come up and assemble all the data, whether it's raw data, whether it's the bare minimum discharge summaries prescriptions treatments. You know very limited data that you that you collect for the purpose of insurance and you know purpose of building insurance and let's say you know for whatever is needed to for clinical care itself you know so that the patient can get a second opinion the patient can come back for follow up and lack of data is not going to come in the way of him getting best care. And in all of those places I think the the the the the work of or what the data blueprint right to attempt to build in some way or form is to have some kind of data in these very basically anonymize or basically dealing with the personally identifiable information from the information from the data, the raw data or the clinical notes, so that those who have access to the personally identifiable information don't have access to the clinical records and those are clinical records don't have access to the to the identifiers so that you know that in that way you probably can can there is a data provider who matches the data identify patient identifiers provides a clinical records in encrypted format and the user and the application builder or the vendor on the other side who gets only the patient records for his usage or for the purpose of transparency but does not have access to the patient records it's it's it's there is a lot of anything previous work that's happened in the financial sector, which is sort of a good example to follow and I'm sure there are loopholes that many of you might be aware in this space, which already exist. I am now expert in how to exactly manage the data exchange so that it is federated and spread across all these, all these hospitals. The challenge, the biggest challenge when you compare it with what has been done in financial sector, for instance, is that, you know, the banks, there are about, you know, 10, 20 banks in all of India, and and you know, maybe five, six large ones which command 80% of the market, and they already have a core banking system but but if you look at hospitals, there are, you know, 5000 hospitals with more than 100 beds each and the total number of hospitals could be you know nearly 10 to 20,000. And each of them have a different IT system so how integrating who's going to go to each of these sites and build a federated sort of integrated federated data access system which enables access to all of these hospitals. So these are open ended challenges that are going to come in the way of implementing such a large scale system so that you have data access but what is important is that this plumbing needs to exist because without this plumbing, you know, there is a lot of a lot of potential applications and technology evolution, technology opportunities that I think he will miss out on. You know, for instance, there are, you know, countries like Israel, for instance, which have a very strong integrated healthcare system today give rise to many of the leading artificial intelligence providers in healthcare and that's purely because the government opens up the data for all these providers themselves and and you know from from day one they they have access to several million patients, which they can use. If India had a system where you could have data access for for both development but also deployment of solutions and the plumbing layer, if I may call it, of enabling hospitals to all be able to talk the same language so that vendors can integrate so that data interoperability is managed. It really opens up an entire, you know, playing field or ecosystem of research in sort of data driven models, evidence based care, how do you improve care, what is the most cost effective way of care and and really gives rise to a lot of innovation among startups who can potentially leverage this data and and you know come up and and maybe not just get them deployed in India but potentially be global players because you know India is is literally home to one of the largest populations if you have access to this kind and the scale of data, the potential of the kind of applications that you could build is is is you know is comparable to maybe only China and and the Chinese are a way ahead in terms of that mandate to open up data to AI providers who want to use that data want to use that data for for some kind of commercial interest. And and and you know I'll sort of try to conclude with that, but but I think what what is what what is sort of open ended and this challenging is is obviously the as we talk about all this interoperability and connecting all the hospitals to get access to data. It is obviously, you know, we're talking about healthcare data biometric data by design is very very difficult to anonymize and you anonymize with so you can know sort of within reasonable limits, try to do what is, you know, recommend what what should be standard practice for anonymization. But if bad actors at one any point want to exploit the data and sort of try to identify it's it's it's it is a challenge and there is no mechanism for anonymization. So it is it is not something that I think requires a lot more work a lot more a lot more discussions to sort of arrive at what is acceptable, very simply ensuring that bad actors don't get into the system in the first place to try to maybe simplify the solution of it. And and and and and secondly, I think, you know, even if you do build all this interoperability and data interoperability framework, what are the incentives for hospitals to actually implement system what do they get. How does it really affect affect them being able to bring more patients can we use this data so that hospitals can project themselves better in terms of outcomes in terms of patients, in terms of patient experience in terms of in terms of anything else. So so the entire ecosystem around how this data blueprint can actually improve, you know, the experience and and and the business case for hospitals is also something that needs a lot of work a lot of selling a lot of thought around. And and that's that's that's about it. Happy to sort of answer more take more questions as as we go. Thank you. I'm going to come back to you, but we move to Irish to explain the complex problems of how do we even solve it and how the consultations are happening around this, because while the keys and practitioners and doctors all of them have a different set of issues hospitals and data ecosystem. I can come and probably tell us more about what's happening on the consultation and what's the government doing to enable all of this and where are the concerns, especially privacy anonymization and what's happening with the health care privacy laws. I shared up. Thank you. Thank you. And so third for for sharing everything that you did and for raising tons of person points lots to unpack over there. I'm going to jump in and talk a little bit about sort of how we've come to this point where healthy like health systems in India are being increasingly data fight and my work specifically sort of looks at it from the perspective of welfare systems and provisions by the state. And that is that a lot of story that I want to start with digital digital which is the NDHB and talk about a few things sort of that came out in in the consultation documents and at the consultation with the ministry itself. So to sort of begin with how welfare systems have come to this point where increases and this is sort of it's the tool and this exit move of India in move by the Indian state in sort of adoption of neoliberal governance practices and by while the term is often very loosely used for me it specifically entails a certain withdrawal of the state from the provision of public from the provision of public services that includes things that we are seeing today such as the linkage of the other with the PDS system tons of welfare programs across ministries both at the center and the state being sort of sort of made into data based systems and searchable database systems to sort of merge the silos and be able to create this perfect profile of beneficiaries and that's really where the there's so many of the health systems that that I'll be talking about come in and these these are not new. So one of the oldest systems that we've been able to sort of look at is called the mother and child tracking system which started in around 2003-2004 under the sort of National League Governance program that was created at that point to use ICT based tools for to reach quote unquote underserved communities and to sort of rationalize India's massive sort of social security and welfare programs expenditure. So a lot of the discussions that we then see and this is again linked directly to the concern that has been raised by several economists, sociologists, activists and like large members of civil society at that point itself which was that this this rationalization of expense is the sort of underlying intention behind the focus that has been placed on leakages and corruption that are that are sort of in that in here in India's welfare systems and and that is when technological tools started being proposed as solutions. So as far back as 2007 that is when the right to food case was being argued in the Supreme Court that is the PUCL case in one of the interim orders that the court had laid down building upon recommendations of a committee that had been set up. They had sort of they had explicitly indicated that the welfare programs need to be quote unquote automated to to make it more efficient. So this is again the language of so we are sort of approaching the language of privatization liberalization and have moved pretty far away from the explicitly socialist approach of the Indian state prior to 1992. And this is where the NDHB the National Health Stack and in its in in their previous reforms come in but before that there was something called the integrated health information platform. There were policy measures the EHR standard documentation there were policy measures such as the Disha Act which still haven't seen the light of day. And so with while initially in this sort of datafication process there were there were both national and central level health programs that were being datafied but they were still very they were still discreet and did not speak to each other and the National Health Stack and the NDHB then seek to fix that problem and they want they seek to create this overarching pervasive database. In their words it will be the single source of truth and it will then contain sort of deeply sensitive health information of if it's if this imagination is given sort of its full extent then of every citizen in the country and will record every interaction of each citizen from cradle to grave again quoting from the policy documents around around a lot of these welfare programs and also specifically the NDHB. So the and this comes hand in hand with the National Health Policy which again expressly stated the leveraging of big data tools to ensure to sort of have a wellness approach to health care in India and to sort of to and again quoting F. Jenny Morozov using a techno-solutionist approach uses a techno-solutionist approach to fixing sort of fixing underlying structural problems. So there is never a thought given to how sort of let's say issues of access issues of poverty need to be fixed. But it is assumed that the mere placing of a technological infrastructure will paper over these cracks and and then potentially solve for these problems as well. And so the National Health Policy then comes in places this incredible primacy on health insurance which also the previous discussions have mentioned and that is the situation in which the Ayushman Bharatiyoshnath, the ABY comes in and sort of seeks to and is the latest iteration in the Indian government's attempt to provide universal health care. But this is still done through a focus on health insurance and thus what is then being that is being actually ensured is this this indirect targeting of demand side financing. So this is trying to fix problems that insurance based models will have not really health problems as insurance model insurance based models are characterized by problems of information asymmetry, model hazard to sort of curtail the incentive to overuse health systems and thereby preventing cost inflations and so on. An important point to be mentioned is that the UK had attempted to digitize its health care service similarly called National Program for IT and it was discontinued in I think six to seven years. Like it was a complete failure. There was and there was complete for those absolute political consensus around it and the biggest problems that it and there's a sort of bunch of literature around it. Some of the key problems that were highlighted whether it was top down in nature did not factor in for local decision making. And this is again a problem that the that is that in here is in the NDHB as well and also other health data health health data systems such as the are very productive and child health portal or the MCTS where so health workers and data entry operators are already overburdened overburdened in terms of the data entry requirements itself. But while they are the ones who are still do the data entry, it is that they are it is that they are never involved in the decision making around what data has to be collected and how this data can be better processed or used for making reflective policy decisions that actually sort of contextualize the local needs, local needs of diverse populations and underserved populations in the country and especially given the sort of unique contours of the of the Indian demographic in terms of both density of population and the geographical spread and sort of the inability to be accessing health welfare systems in the first place. The problems around access are especially pertinent because this is an assumption that all welfare data vacation measures and explicitly in the NDHB is the assumption is that of new universal coverage of phones. So and in something like the NDHB which has like a complex technological framework where where patients are or I don't know consumers of the health system are supposed to interact with this infrastructure and that is where the sort of consent framework lies. It is it is not just phones that you need to be able to access, but also regular connection to the internet and then access to smartphones. So, so one problem is that of access and then this then this access is so on is strictly mediated around along gendered lines in India so while there may be access. There may be phones present like you may have non smartphones present or feature phones present in most households in the country, but it is that women for example do not have do not have access to their personal phones and this is a major privacy violation, especially in especially when patriarchal gender norms are are continue to be rampant in continue to be rampant in the country and then also in mediating access to to the online. And there is like, again, there's several surveys that have been done one, including that where digital empowerment foundation that talks about how that how while we may have made slides in literacy but it's digital literacy is severely lacking the packet at about 90% of the population not being digitally literate. And again, this is a severe challenge, especially in terms of the infrastructure that systems like the NDHP propose and this is also a function of how the decision making is happening. It's happening in our urban centers. The decision makers in ascribed to a certain Silicon Valley is Asian ideology and this is this is sort of an in an environment in a context which is strictly male dominated and these all feed into the design of the systems and that instead of making sort of making welfare systems reach so underrepresented communities, it actually ends up further excluding them. Now the next thing that I want to quickly mention is the standards of interoperability and content that are proposed in the in the in the NDHP. So there is there is a strong focus on open standards and interoperability and that is truly commendable. But it is there's we have again tons of tons of experience around how it's it's incredibly difficult to operational operationalize interoperability given the sort of menu of open standards that that are there to choose from and in the NDHP then specifically talks about decentralized stores. So it's going to be an incredibly federated architecture with a central repository and tons of regional centers or at the sites of service providers while on the one hand it's already unclear what information will be stored in these repositories. So, for example, what will be in the central repository and what will be contained in the regional repositories it's it may it may be the case that these problems further get amplified if as mentioned earlier that health data can get exponential very quickly and it sort of it may become a problem that becomes like really difficult to handle later on the the next thing I want to talk about is that of anonymization. Now anonymization or the identification has been a constant feature in data in data protection data protection communications. The data protection discourse generally over the last few years so right from the Shri Krishna committees data protection report at that point itself they had mentioned how methods of identification the identification of anonymization have a strong have a possibility of failure earlier so that mentioned that they are not foolproof as well. And the key reason why this may be the case is because quasi identifiers are used are still used to link seemingly anonymized data to respective individuals and this is important this is for any system for any welfare system to work that focuses on reducing errors of targeting the identification is going to be is going to inherit in some in these systems in some form or the other and how this is typically done in even enough in a framework where anonymization is thought of from the very beginning is that of using posse identifiers to to link this and to sort of link it to link it to individuals. And it's especially pertinent to mention in that why a further tension sort of arises in in so far as the end HB seeks to also create this ecosystem where private actors and startups may be able to leverage this information through the use of open standards and open API is that has been a constant feature of the stack environment so India stack health stack and so on and this is this has been a central feature so this is and that is where innovation and that is regarded as a site of innovation but the Disha act and the Disha bill sorry explicitly stated that digital health data whether identifiable or anonymized shall not be accessed used or disclosed to any person for a commercial purpose and in no circumstance be disclosed to insurance companies employers etc or any other sort of entity as the central government may specify and and in doing so the Disha act also squarely places the focus places the ownership of their digitized health data upon the individuals so on the other hand the the end HB sort of possibly intentionally chooses to to sort of ignore some of the recommendations that were made at that point in time and thus in and in doing so sort of it talks so much about this goes around big data frameworks as well so the more the data the the meteor it is and and and and thereby so this linkage is crucial to driving innovation and sort of providing sort of providing welfare services to to Ben to potential beneficiaries then we also have the problems with the HR that both so there's an thing have already have already mentioned one thing that I just want to point out additional additionally is that while their learnings were still focused to like one major large one major healthcare healthcare chain in India and in and so there's in some in the case of private actors this these problems get further heightened in the case of already underserved and under-resourced settings and the whole premise of the functioning of the systems is the requirement of more data so those already with reduced access to digital tools cannot frequently be using the HR or pH or personal health records as the end HB terms it and as a result what then happens is that since these these information systems work best with greater frequency of their use it then so ends up happening that those who are already on the margins then further get marginalized let alone the problems that doctors and doctors nurses and other practitioners in hospital systems already face in terms of being using in terms of being able to use these systems and these are in settings with lesser constraints to resources I'm going to stop with that and I'm happy to take any more questions on anything I can clarify for you to sum it up Aish is essentially saying that technology can't really solve all of the welfare side of the challenges which healthcare is of like considering the pandemic that's going on right now it's increasingly clear that privatization of insurance as a model there is not enough to submit it in that aspect of the welfare side of things we're going to open up for questions I think we have around 20 minutes if you have any questions please type it in the question and A we are also running few polls taking some feedback about how the events been kindly answer some of those questions and if you I will let people ask the questions themselves I have one question to Sudeeth from Srikant Srikant I am turning you I'm allowing you to talk and you can ask the question directly to Sudeeth you need to unmute yourself yeah so my question was Sudeeth was mentioning that currently for research purposes the data acquisition happens only through hospitals and patients are not even aware because it's past data but how do you see with this digital health data blueprint and how things would change would say patients be asked for consent for say some procedural data that happened for five years ago and would they have say a chance to not be part of the research and so on good question and one that I've had multiple thoughts over but I don't think as far as the national health blueprint I might be wrong but I do not I'm not aware of this being talked about in the blueprint as such on how these finer elements will actually be taken care in my opinion I think doing something like data access dividends is very very ambitious and I don't think it's feasible in the near term but I do think something like opt out of clinical research or allowed or maybe in the future being able to use that kind of data infrastructure to ask patients to voluntary submit the data could be sort of frameworks that the government could explore but I don't think we have access to this question at this point today the way it stands is that the hospital owns the data and tomorrow maybe the government could for public health use cases and date that hospitals open up data for you know say something like tuberculosis or coronavirus right now the government is I know putting their way behind hospitals to get them to share their data but beyond public health issues you know it's difficult to see how this is going to ban out and how you know what role the digital blueprint can play in ensuring there is some patient consent involved in the entire process you have a follow-up Sringan? No so you have some more questions on the Q&A sorry it's from YouTube where we have live streaming it I think someone was asking if there are any good resources to learn more about the Chinese approach and if there is any startup activity in the area of enabling EHR adherence like from a compliance standpoint of you do you know any anything about that Shukri? Good question these sources see Chinese I think EHR adherence or implementation I think there are multiple books on the US system and what is wrong with the US system I personally think there are good books like Digital Doctor and multiple books by Eric Topol which talks about how the digital revolution in US and how bad the EMR systems are I think the most you know commonly cited literature around this is the entire implementation of the high tech act in the US which basically gave rise to the implementation of EHR systems across all of US China I have personally not myself come across literature which talks about how they are doing it what I do know is anecdotal data that many startups in China claim to have access to data from 700 hospitals and a scale of several millions of patient studies which is basically unheard of for anyone in any other country so I only have anecdotal I have not seen any resources if I find I would I I myself am looking so I haven't found anything so far Khudai Chai Main here? Yes, yes please so yes I think you know like how we talk about any digital story found outside looks pretty wonderful but I also happen to have done fieldwork in Chinese hospital as well the current what we are talking about in terms of the issues with electronic patient record is being faced in China as well as in as in in India or any other country and you know the the issue that Sujuan talked about in terms of the data being held by hospital it's also the same case in China yes there are initiatives to push for more open data sharing but that's currently not there so just to give you an example if a patient want to have their digital patient record out of hospital impossible the hospital itself won't allow it currently the most common practice is that the hospital would give a printout version of the digital health record and then for the patient to take to take it somewhere with them even for after the patient is discharged the store the storage of the digital health record is not is not the storage of the old digital health record again it's the printout being scanned and put back into the system and this is not just within one hospital there are several hospitals that are doing this so yes there are initiatives pushing for this and especially you know along the coastal line of China where the digital development is further and more digital literacy again a lot of hospital because they are national owned hospital that's the most common type of hospital in China they are under resourced so investing in digital infrastructure is lower on their least compare into the procurement of the MII machine EPG all of that so this is my two cents thanks Ting I'm launching the second fold if you want for the feedback meantime I'll ask one more question that's from YouTube it's on consent I think it's Karthik Vishwanath he's asking how can we build effective and meaningful consent standards from patients while we also move towards data data sharing especially during a pandemic when swift action is key to design interventions it's open to anyone Ting can you tell us about the current concern practices in Indian hospitals um the concern in terms of what data sharing how are the collecting data are they informing the patients um I don't think I have a good answer for that that is something I didn't look into because when we were looking doing field work in in the Indian hospital we actually specifically limited ourselves the inaction we could have with with patients because the focus would be on on a hospital side that for one and secondly if we were to conduct field work with patient that process in the south is yet another consent nightmare so we sort of for the initial study we decided to not get too involved in that so I don't know I don't have a good answer for that however if it's a research hospital that might be slightly different while consenting to the hospital of receiving treatment there could be a subclass on consenting the hospital itself um using the data for research purposes but that's a for research hospital and that's a slightly more common practice across not just in India in China as well as in the UK I mean right now whenever I go to a hospital like if it's a fairly modern hospital they actually create a HR record of my data based on my phone number so you have to go through it and whenever the doctor actually gives you a prescription the nurse would not let you take the prescription without scanning it so they have scanners everywhere and they just scan it I don't know if they're doing well the concerns like in flight that we will take your data for now but Sudev do you have any understandings of how people can build any consent based solutions or is it going to remain this way? I think it's a tough problem to solve purely because of how you get access or how you enable the patient to be able to provide you that consent either you're going to end up with a system where you need identifiers all the way up to mobile number which provide you some kind of consent in many cases it's not the patient who is in a position to be able to give you consent but often the caregiver who is going to be a family member or someone else so I think it's a very tough problem even in cases where you want to do a lot of trials and clinical trials research projects the consent is usually done telephonically and the percentage of people who are able to reach and get that consent itself is usually you have such huge percentage drops and it's never possible to reach out to them so I think it's a fundamentally I think broken problem and needs a lot of thinking to get to a stage where we can talk about a system that works I wanted to add a little thing Yeah this coming from the citation in the Shree Krishna committee around from Arthur left article called contract as a thing so the basic proposal is that contract should be treated as a product so then the regime of product liability gets invoked and data fiduciaries are liable as if the consent form itself was a product so then sort of data collectors would have the obligation to sort of take on more substantive obligations such as showing the notice before any practices that are communicated in the notice requiring affirmative consent without any prechecked boxes not having boilerplate clauses ensuring that that sort of that the use of the data is communicated read and understood before so that informed consent can be provided and this sort of adds in to again use a term from that is commonly used in sort of the FinTech and the app building space is that of that of introducing friction and there is an inherent challenge here because this introduces friction but it's important to introduce this friction but there is a lot of resistance to do this and the other thing that I want to quickly say is that so while the current outbreak has thrown up solutions that data data heavy and since I do not have the expertise to comment on there being like alternate epidemiological models that can address these one thing to still to still be said is that it's not this dichotomy is a false one that that of being able to only that only being data subjects allows would allow analysis of analysis let's say in any epidemiological crisis it's that these systems should have a shelf life which is what usually isn't the case that these then end up taking the shape of pervasive surveillance systems that extend to far beyond what their initial use was and this is what we see with the technological system generally as well that get implemented by the state in terms of like shifting goalposts when either they use changes or they are rendered not particularly useful for the use that they were initially thought of and that is something which I can anticipate happening with our we say to app for example, so the point being that this one is both a consent is something that needs to be addressed both from the prong of consent and as well as the from the prong of having sort of accountability in terms of un-purposed limitations in terms of how data systems can be later misused. Thank you, Aish. I think that was a question as well. How does are you safe to have kind of relate to the national health data blueprint and what happens in healthcare emergencies? I guess you've answered that. I don't think there any more new questions. Maybe I can just wrap it up. If you have any finding comments that you guys want to add please jump in. Okay, I guess that's it then. So to simply sum it up this we can't look at the whole health infrastructure in isolated way because these are very similar to other infrastructure rather digital infrastructures that are coming up. For example, consent for healthcare will be similar for consent for some other online mechanisms because the electronic consent form is the standard that is being built for any digital India solutions. So we can't look any of these systems in isolated way especially in terms of the society and the law and how technology can affect it. On that note I have one final question for the poll and please take that and we will end the session.