Hello and welcome to the Indo-German Business Talk. My name is Matthias Catón I am your host. The Indo-German Business Talk is our monthly format organised and run by the Indo-German Centre for Business Excellence and in that format we talk about all sorts of different issues that are relevant to Indian and German businesses; we look at the topic from different perspectives, both countries, we try to find out what we can learn from each other, the pros and the cons and today we will talk about digital transformation and healthcare so two very, very important sectors and we will look at how they change each other and what things are going on there in India and Germany. As usual we have a group of panellists who I will introduce in a moment and they will engage in a insightful conversation here online in a moment but we also want to hear from you, from the audience so there is an opportunity for you to ask questions through a chat function and we use Slido for that. To kick things off I would like you to fill out a little question, a word cloud, where you can type in some phrases or some words that come to your mind and the question that I would like to pose to you is: "What should be considered when enhancing Indo-German cooperation in digital healthcare services and research"? So please go to Slido.com to the page and put in whatever words or terms come to your mind when thinking about Indo-German cooperation in digital healthcare services and research and in a moment we will look at what you guys came up with. In the meantime let me introduce our esteemed panellists, there are four of them and I will introduce them in no particular order, so just random order basically and I will begin with Sandra Barteit, Sandra welcome.
Sandra is a leader in Digital Global Health research at the Heidelberg Institute for Global Health, she uses tools like AI and machine learning to improve health outcomes especially in the field of digital health and she works in Africa or on Africa, with Africa, in Zambia for example, where she uses digital technologies to improve medical education and she's also known for her research in Burkina Faso and Kenya where she also uses digital tools for health data collection and analysis, welcome Sandra. Thank you. Next we have Jan Beger also from Germany he is the Senior Director of the Digital Ecosystem at GE Healthcare and his job involves developing cooperation with partner companies, startups and networks. He has more than 15 years of experience in healthcare informatics and medical imaging and prior to joining GE Healthcare he worked with digital health startups and Radiology I.T companies. He's also the founder of a non-profit called GR4A Academy and that non-profit educates children about artificial intelligence.
Third panelist is Jibu Elias, Jibu is a leading expert on India's AI ecosystem and he holds key positions at INDIAai, which is a national program on AI by the Indian government and he is a Senior Fellow for Responsible Computing at Mozilla. His work focuses on building a unified AI ecosystem in India, reducing the digital divide and creating inclusive AI ethics and governance guidelines; also very important topic that we will address today - the perspective of ethics in this field. And then finally we have Resham Sethi. Resham is the Country Liaison Officer for India at the International
Digital Health and AI Research Collaboration, you can also see that in her virtual background. Her work involves developing innovative frameworks and strategies to strengthen health systems and impact health policy. She is committed to improving access to digital public health services and transforming health systems governance through digital innovation and artificial intelligence. So you can already see that we have a very diverse panel, as we usually do, looking at the same topic from very different, both geographical angles but also in terms of topical areas of expertise and I'm looking forward to the conversation. Now let's see what you guys came up with and apparently you are still a bit shy and a bit rusty because there's only four things that were submitted apparently, but that's fine we're getting into that, let me read them out there is the topic of equitable access that you pointed out, protection of citizens, the question of urban- rural contrast or the divide, potential divide, I guess in the allocation of services and then finally connected to that the general allocation of resources to different sectors of society I assume. Good, there's another one okay people are waking up, digital literacy, good okay
perfect we'll get back to those questions and once again if you have any particular issues that you would like to address or you would like the panel to address please use Slido to send them to us and we'll make sure that we'll get them to the panellists. Perfect let's begin our conversation. Let me start with Sandra, so given your work using AI and machine learning to enhance health outcomes, could you explain to our viewers how are these new technologies changing or maybe also revolutionising the health sector? What are the changes that are most imminent there? Alright yeah thank you. So our main focus is currently Africa but we are also expanding to Asia, for example we started a project, a bigger project, with Malaysia. So for example in Africa, in Burkina Faso, in Kenya we are using wearable devices to, for the first time, generate insights on the individual level because we have no understanding really how people, for example, move in rural areas and this is particularly important for example given environmental exposures like heat which is increasing also the same and similar in India, as it is in Malaysia. So we have those individual exposures environmentally on people but we have
no idea what it means really so this is one certainly important factor for artificial intelligence to generate this data coming from those areas because currently we have a big lack of data to drain AI systems coming from areas of Africa, I'm sure the same is true for Asia so we have this problem also in Malaysia. And we also get to this next step in healthcare provision so understanding individual exposure so that we are able to understand what is happening in the population, to better target interventions and to better allocate healthcare funding to help the people with what they are currently burdened let's say, so this is like one of the activities that we do at the moment and we are also focusing on particular heat diseases, like lung diseases, particularly in dry areas where there's a lot of dust and heat people struggle with with asthma and chronic obstructive pulmonary disease for example. Currently it's a black box, we don't really know what is happening, because in those countries the main focus is on 'The Big Three' they are called: Tuberculosis, HIV/AIDS and Malaria. So this is one aspect of my work, the other, for example, in Zambia what we do is leverage digital technologies to strengthen medical education so we have introduced blended learning, so e-learning, leveraging face-to-face education to strengthen the education of healthcare professionals because we have an infrastructural problem because we need many more people and many more doctors, many more nurses who are able to serve the population particularly in rural areas, so we need to scale up the education but the problem is that it's not possible to build schools, universities as quickly as there is the demand, so digital technologies can certainly help leapfrog that and this is what we focus on in Zambia.
So to strengthen the healthcare education just to give you a teaser of our work but I don't know how long you want me to go because I can continue. It's extremely interesting, of course I was fascinated by that you said that before you get to the AI part the analysis part, it's a lot about generation of data right so you said the wearables that you're equipping people, could you just briefly explain how does that work? So you just give everyone like a smart watch and that smart watch what does it track? You said heat but I guess also health conditions, what exactly, in practical terms, can we so can we expect from these things? So like everything in life we are financially restricted so we are using consumer grade wearable devices so those are like I guess most people know Fitbits, we are using [inaudible] because they have established a coherent ecosystem of data synchronisation so that we can leverage also a platform which makes it much easier for us to collect data in rural areas where there's no mobile data even, so we need the devices to also synchronise offline. For the first time we went to rural Burkina Faso and the first step was to understand if people would accept those devices because people in those areas they are not very much exposed to digital technologies so we didn't know what would happen if we, you know, go to these villages and put wearable devices on their wrists which emit also regularly those lights; so if you have a smart watch yourself you know they admit green or red light very regularly; but people we found, so we asked them, we went there, we gave them the wearables and we work together with a partner research institution in Burkina Faso and Kenya who also organised the field workers because the field workers need to go to the villages and synchronise the data of the wearable devices and also charge the devices and we all needed to establish those processes to make this possible and we also need to give the devices because the population essentially they don't have any of those devices so we provided them through our study resources and the devices they collect information on daily activities so we know when people move during the day and how much they move, so we found that some people they walk like 40 kilometres per day so we were quite astonished, so it's large distances which we weren't aware of and also when they moved so we also collect weather data so we know it's hot, it rained a lot and so we know environmentally exposures. And they also collect data on sleep so we know how long people sleep and sleep quality so because sleep is a very important indicator for health so if you continuously sleep less than is recommended, it certainly has an impact and kind of fosters also chronic diseases and the other is heart rate which is also very important given environmental exposures. I'm gonna interrupt you there, I think we could listen to you for a long time, it's extremely fascinating the work that you do and you're obviously very passionate about it which I can also understand, I just wanted to get to Resham and ask her how is this situation in India? Is India working with similiar initiatives? How is the situation over there? So there, thank you for asking that question and hello everybody thank you for taking out time to be online with us. In terms of India, India is a very classic example of of an amazing disparity that exists; so at one hand at, one end we have extremely great technologies that a certain segment of the people have access to but we also cannot forget that we do have more than 60 percent of our population living in the rural and tribal areas where accessibility to healthcare services, the quality of the provision of healthcare services, affordability, is a major barrier. So in terms of India, and I like to also give an example of what we are
doing here so maybe that would through a little bit of light, instead of getting into high level technologies and implementing in communities what we're doing is we're getting to basics, we're trying to engage with communities and build trust with them first; because we understand that the digital literacy is a major issue here and so what we're doing is we're focusing on the mental health aspects in Punjab in terms of strengthening the mental health systems through community engagement work so it's a solely research-based project and we're getting into basics in understanding what is the current conditions of mental health in Punjab? What is the current situation of the disorders, the prevalence of disorders, the understanding and awareness of people about mental health and well-being and their quality of lives, so it's asking questions such as do we have enough technological infrastructure in rural Punjab to even implement emerging tech or such as artificial intelligence within all variables within these communities because when we start getting into implementing really extensive technologies we need to keep in mind that there is a digital divide and by digital divide there is an inequitable access to internet, there is an inequitable access to services and that needs to be bridged for sustainable impact of emerging tech; so while there are marvelous hospitals that provide extremely good services we also have to bridge that digital inequitable gap that exists in India and that's what we are trying to do. So that's the situation. Resham, so let me ask you when we discuss all these topics, the impact that digitalisation, artificial intelligence and the like have on healthcare, is it more of an opportunity along the lines of what Sandra outlined that you know we're better able to serve for example the rural population or are you more worried that the opposite will happen is that this digital divide that you mentioned will make you know the differences in terms of accessibility of services even bigger between say rural and and urban areas? That's a very good question. In my opinion I feel they run parallel; while we are moving towards
transformational work we also have to keep in mind that our transformational work is equitable; so in that sense I think even though there are amazing opportunities there are definitely challenges and there will be pros and cons and challenges and opportunities in everything that we do. So while we are moving towards addressing these opportunities we also have to look at challenges and take care of them. Let me move to Jibu who is a leading expert on AI in India and also someone who looks at these things from an ethical perspective which I think is interesting, so what's your take on this? Is that an opportunity? Is it a risk and how do you judge the ethical component of that? Because obviously when we're talking about data and healthcare we're talking about a potentially very, very sensitive area and one where there's also potential or at least the risk of abuse of that information, what's your take on that? I think that's a very important question and I want to say Sandra said some wonderful things and you know I wish, I could, she could have continued beyond the allocated time; at the same time I think Resham also pointed out some interesting points. Before also I get into the point I'm very happy to see some of the familiar names in the attendees, you know Avneesh, Amarjit and so on, they are also experts in healthcare from India.
Now coming to India and AI and Healthcare right, in India we see AI I think in like three importance: first accessibility, improving accessibility. I will give you an interesting statistics there is only one eye doctor for every 200 000 Indians and you understand that 70 percent of blindness in India are preventable, so most of that happens because of diabetic retinopathy or cataract, especially considering cataract is very common. So that's one area: how can you ensure that there is accessibility in this rural areas, areas beyond big cities. Secondly how can you bring down the cost of diagnosis right? For example, another statistic I think 60 - 70 percent of cancer-related death in women in India happens as a result of breast cancer and breast cancer is a condition if you're diagnosed in very early stage you have a very high rate of survival and then later you get the your survival rate like drops off, like a ball drop in from a cliff. So how do you bring down the cost? The cost of mammography is very high in India and
at the same time you have to go to big cities to have that scan. So this is one area right We have startups which is thermal imaging and machine learning to provide non-invasive breast cancer diagnostic or E-Paarvai, the government initiative from a state government in India which looks into cataract which is a mobile-based application that can diagnose cataract. Second area is in electronic health records; now we Indians are very notorious when it comes to our health reports I mean because it doesn't exist to be honest. For example I went to a doctor in a particular hospital to show my back then I go to another hospital for a cold right and there is no uniform space to keep all these health records and considering that we have this advanced machine learning models or language models or things like that, these health records can be used to you know gain information, you know find data and things like that. So that's where the Ayush you know digital health mission program, the Indian government is building on which will unify the health record, it will make sure that there's a token system through which maybe a particular hospital or a doctor can seek access for the self-record and things like that, so that's a second area right how we uh focus on addressing uh this health record and things. Now coming to the ethical and private question, yes there's a lot of question right and the ethical questions doesn't happen in healthcare just because of AI but it hasn't been existing right, for example in India there are many multinational companies who've been using India as a test ground for pharmaceutical trials so those things are happening, now the question with AI is first of all the protection of this private information I think the government is working on Digital India Act and the personal, privacy protection act and similar legislation so that's going to be very crucial here, the legislation will be very crucial, how this data is stored and how this data recorded, e-records, you know, health record database and things like that. Secondly there is a question on where and
where to use AI and where not to use AI right and that has to come from the public trust also. Right now what is happening is, I think Resham pointed out, that the lack of digital literacy so that because of lack of digital literacy the whole debate of AI interest is not in the public domain yet as we might have in Europe or somewhere else, so that's the other problem there like how do we make sure, I mean you have to bring these ideas into the discourse then you have to solve it so that's not happening at this point. Okay, thank you very interesting and obviously we're still in a very early phases I think there's a bit of a wild west approach everywhere when it comes to AI so it's not surprising that in the field of healthcare it is similar. Let me
bring in the perspective of private business and obviously Jan you are with a major corporation, major player and that field, GE Healthcare, you have a long track record, long experience in that field so how do you envisage the future let's say patient experience in a digitalised healthcare world, what are we, what can we expect there in the future? Matthias first of all thank you very much for having me, I hope you can hear me well. And look to your point and maybe just to add to what Jibu just said I mean from my perspective I believe these technologies right which are called artificial intelligence over the longer term will persistently and pervasively enhance all aspects of healthcare, so we will see it I think across the board everywhere impacting the lives of not just patients but healthcare professionals and everybody involved. And look I thought to start off with and I hope that's uh of interest, is to share a little bit or a couple of data points actually around what are the problems and changes we have in healthcare and and some of the fellow panelists mentioned some of them already, but also what are the opportunities here in the context of artificial intelligence? So look first of all what we see today is that it just takes about 73 days to double medical data right I mean just picture this, all global medical data is just doubled every 73 days right now. There are 6 000 medical journals out there publishing 900 000 articles a year. For instance, for a dermatologist there are 11 000 publications to read each and every year and to keep up with primary care literature a GP would need to read 21 hours each and every day. And look Jibu mentioned mammography there is more information in a mammogram than there is in the telephone book of New York actually I don't know if there is still like a physical telephone book, and it's not uncommon for a mammography image to be 1 GB in size speaking about mammography or radiology overall in a 12 hour shift a radiologist today is looking at around 50 000 images compared to just like 500 like 15 years ago so there's this explosion of data in all facets of our lives and so in healthcare and I think really with no end in sight and I believe this is something where only technology can play an important role and contribute to issues like access to healthcare, quality of healthcare, also cost of healthcare and so on. So another aspect, I think it was also briefly mentioned, is a medical errors, I mean
globally we see around 40 million medical errors each and every year and in Europe for instance it is estimated that up to 350 000 people die every year due to medical errors which is like the population of Córdoba or Venice in Italy or Toulouse in France; and for every 100 hospitalisations per year approximately 14 adverse events occur, so really things that when you see and hear those numbers really touching and I think this is something where we can make a difference with technology. Maybe one last point if I still have a minute Matthias on this. Right now I think this is a problem we face across the globe so that means also in like Germany and India which is the shortage of healthcare professionals which today I think globally is around 7 million, right healthcare professionals missing and also looking forward to 2035 there are estimations that in this year 14 million healthcare professionals will be missing due to I don't know retirements, not enough young people entering the space and so on and when you think about this you also need to know that the demand is going up, the complexity is increasing, care pathways become more and more complex so this is really things we have to tackle as I said access to healthcare, quality, efficiency, cost are really growing problems and on the flip side, right, now turning over to AI we have this massive amount of data which today very often is a huge challenge, we talked about the access to this data, get to a holistic view of the patient when data sits in different silos, we talked about the ethical challenges, data privacy and so on but I think also when we do it right and when we apply the right technologies data can be a huge, huge opportunity so maybe I stop here. And this is a good point and I think no one on this panel would dispute that technology, AI and all these things that we've been discussing have generally a positive impact on healthcare and the outcome for patients but I'm wondering is that universally true and I'm especially looking at healthcare professionals and we have a question here from our audience that fits in very nicely with that I'm going to read it to you: Is anything being done to address stakeholder education especially health professionals? So the question to the panel would be what is being done to get the healthcare professionals on board with this and maybe also to overcome some of the skepticisms that may exist along the lines of what Jan just exposed or explained. I would assume that probably not everyone but a lot of doctors, for example, would say that well you know it's impossible for an AI to be better at diagnosing a mammogram or anything like that, or any other form of radiological image; Jan maybe let's begin with you, are you working on that? Is that an issue? So Matthias thank you very much and look it's a fantastic question, first of all what we see in our work the last couple of years implementing AI into healthcare settings, trying to support healthcare professionals with those kind of technologies is first of all that it is not about AI replacing the healthcare professional or the medical doctor, what we really see is when healthcare professionals use AI right the outcomes are better versus healthcare professionals not using AI, it's really like joining forces between the human being, the expert human being and this technology that's going to make patient outcomes improve, that's definitely what we see. One big issue we have been struggling with
really the last couple of years is the fact that implementing healthcare AI and translating it from like research into real clinical settings is challenging due to the fact that healthcare professionals, and actually not just doctors but also nurses, radiology technologists and so on, they don't trust it because of these things like black box behaviour and so on yeah so they don't know how this technology works so they are they're not trusting it and and therefore they are a bit conservative; so what we as an industry player but I mean many other companies do as well is focus a lot on basic education, so we for instance run basic education programs for healthcare professionals, one of them is called HelloAI Professionals, you can just Google this, where we spend time, invite healthcare professionals from all sorts to just join us, it's an online program where they really get to understand the really the fundamentals of artificial intelligence - what is this technology, what is deep learning, what is this black box and we believe with some sort of fundamental understanding you start building trust which is needed to really leverage and utilise those technologies in your day-to-day workflows. Sandra in your experience working in rural areas is it more difficult to overcome skepticism from healthcare professionals or from patients? Or do you not experience any issues like that? Yeah thanks, this is certainly a very important question and topic I think for because we first need to think about how to translate those AI-based technologies into the usage actually so we can develop all of these fancy tools but they need to be used and I think Jan said very importantly that we need to do a lot of education so you know providing this understanding for healthcare professionals who come from a completely different area of education and have not been exposed to kind of this education focusing more on the understanding of actual AI- based models, deep learning and all these kind of things that are creeping into their daily lives now. But I think we also need to look at the patients as well because those are the ones being treated with AI-based models let's say and we have done a systematic review looking at how AI is being used in Africa already so what kind of AI-based technologies have been implemented in healthcare and what models are they using and what are kind of shortcomings and advantages? And certainly one shortcoming is also in this healthcare setting between the doctor and the patient that you know when you sit in the doctor's office and the doctor has let's say a tablet in his hands or in her hand, let's imagine this doctor to be female, so she's holding the tablet you know and and then she's going through the diagnostic process with the patient and looking a lot in the tablet right kind of maybe using a decision-based tool kind of helping to navigate her like Jan pointed out like those huge pile of information which does decision based model is kind of digesting, but in this patient- doctor interaction the doctor is spending a lot of time on the tablet and maybe if you imagine now some older lady who hasn't been exposed to a lot of technology and maybe she's distrusting this doctor like thinking okay "why is this doctor looking so much at this tablet, shouldn't he know how to move forward you know by himself, like what is happening here, I'm not trusting this doctor?" because we are changing so I think there is also a lot of education that needs to go into patients and kind of awareness of why this is happening and that it's a good thing so it needs to be both sides it's not only the doctors but also the the patients and at the same time the doctor you know may be faced with this kind of loss in in prestige let's say because you know he's not the person who knows it all anymore but who has to rely on a tablet so I think there's a lot of complexities involved in just you know handling this process and we need to address this because otherwise AI or any digital technologies will fail to be used on the basis that we want it to be used to have an effect on healthcare in general right so yeah. That reminds us that technological change is not just about the possibilities but it's also about human acceptance factor and and psychology. Very much so. We see this a lot in Africa. And also I don't know I guess the same is true in India maybe? But in Africa we have a lot of organisations who bring a lot of technologies and who basically just throw them at people without like this proper education awareness training and there's like, I mean we have to be careful to kind of avoid this technology fatigue and so that people think oh my God there's the next technology that I need to get used to I just want to treat this patient and now I have to spend like one hour to kind of understand and navigate this technology, oh no and when it's getting hectic we also see this happening in Africa and I think Jibu said this before as well that the healthcare records they're not really maintained very well because maybe after technology fatigue because people then tend to use paper you know something they have used, they know it's working and then they switch to old models back again so we really need to be very careful on how we introduce new technologies, and I think Jibu wanted to add something because your hand was up. Sandra made some important points and I think regarding this healthcare approach right so for example that's one big problem, see you have this advanced AI and language models but you need the right data. For example I have friends who are in medical
space and they had told me how difficult it is to get a doctor in India to move from a pen and paper to a keyboard, it's almost impossible right I mean it's maybe, as Sandra pointed out, could be this fatigue, technological fatigue or something like that at the same time there is this problem for example I think there's a company in India called Wadhwani AI, they're a research organisation focusing on AI and they were doing their research and they found out that in a particular hospital in India, in Maharashtra, every baby born has the same weight, every newborn baby has the same weight and that's happening for years right and then it's become a fascinating question and then they realise that they're not actually measuring the baby they just put the same number every time right? So this is the problem when it comes to you know the emerging countries or the developing world is the data collection of course now how did we solve this problem using AI is the interesting part they created a smartphone application with AI which can then take the picture of this newborn baby and and you know analyse its weight so that's the exciting part. Similar, another challenge, I think Jan put it clearly, is that the there are interesting works happening in India in research world but they are all limited to the lab setting right so you had when COVID came out there were so many applications that came out that said you know we can identify COVID's symptoms using chest x-rays right you don't need the you know expensive lab test right now the problem was all these models were trained on this perfectly tech chest x-rays they had for the training purpose but if you come and see an expert in an Indian government hospital, I think Resham will agree with me, you have to have a special skill of your own to see where to stop and where is down, so that's another point I just want to add that point to what Sandra and Jan was mentioning. I guess that's true you need to convince people of the importance also of collecting data using data and otherwise you're gonna have something like the example with the baby weight on the other hand I found it fascinating that apparently you can take a photo and then AI can guess or estimate the weight so I think that will probably also make data generation more easily accessible and probably also more widely available. Resham I wanted to get to you and it's interesting to hear that also in India it's difficult to manage that transformation from pen and paper to digital devices because at least from the German perspective we often look at India with amazement about how quickly this transformation works you know with payments and all the kind of things whereas here in Germany we sometimes seem to be lagging quite a bit behind and I think healthcare actually is no exception if I go to see doctor yes they may be using computers but to what extent that data is being used widely is a different matter partially because I think technological acceptance isn't there partially also because at least in Germany we're very very careful when it comes to data sharing, data usage especially in the healthcare sector; Resham is that an issue as well and or how do you see that topic, data sharing when it comes to healthcare? I mean the benefits are obvious in terms of knowledge generation and insight that we can get from that but there are obviously also risks for the individual if they don't know what's going to happen with their sometimes very personal data. So, two points and both I'd like to address with a story so the first one is on digital transformation and skills so this is the positive one, positive story, so I had gone to Nagaland, Nagaland is a northeastern state in India, it's more than 90 percent of the population is tribal.
So I've gone there and we were conducting a landscaping study in some of the most remote areas in Nagaland and I was surprised to see healthcare workers with a tablet, that's the first time in their life that they have ever seen a tablet and what they were doing is they were collecting data, health data, whether on non-communicable diseases, ASHA workers or community health officers, they were going house to house and collecting data on the tablet, so that is digital transformations that been seen in India. Now the challenge there is that poor lady was getting so anxious because she was not digitally literate enough to collect that data, so even though certain departments are providing digital literacy skills or programs to these health care workers it is important that we doing it at a recurring basis. So the importance is not just collecting data but the importance lies in educating those workers who are collecting data because if they are not literate enough to collect the data then it's going to be a bias or it's going to be falsified data right. So while we're talking of digital transformation and laptop getting our laptop, tablets getting to the last mile especially in tribal areas in India, we also have to keep in mind that we need to literate these women.
Another very important point in India is that we have a very strong network of ASHA workers; these are local community health workers or midwives, who are the backbone of the country's health system, who are working at the grassroots. So when I was interacting with them the one important thing that they highlighted was "Can we have courses or programs on our phones that will help us build our skill set? Like we want to serve the people but we don't know how to do it using these digital technologies?" During COVID Nagaland government, and of course different state governments in India, as well as the national government, we had a lot of digital services, mental health, or online telemedicine, telehealth services, but the problem was the connectivity from the local communities to the doctor with the people in the middle and the people in the middle were the health workers so along with digitally literating them, digitally educating them it is also important that we have enough connectivity for digital transformation in these areas. Now the second point is the con, during COVID vaccinations were taking place and there we made an online platform in Punjab, there was COVA, centrally it was COVIN so it was all online now the problem is everything is great, but people in the rural areas don't even know how to use a phone, they may not even have a smartphone so suddenly everything is becoming online and and there is a huge number of people who are not even knowing how to use WhatsApp and then they're going to the Health Centre they want to get themselves vaccinated but they have to register online because the health workers telling them to register online and they don't know how to do that and also they may not have a phone because there are a lot of people who have a regular phone, it may not be a smartphone, so that's another segment that needs to be addressed in India and that still exists. Now in terms of
data, sorry you said data privacy or? Yeah, yes. Okay so health is, health data is very sensitive so we really need to come up with strong data governance systems, I think EU laws, data protection laws, that exist of course European Union has a very strong data, GDPR guidelines and Canada also has very strong guidelines, India is coming up with one which may be even more stringent than the GDPR so I think we're working towards that but what happens in this is during in between the time from where from now until the law comes into place there are a lot of organisations who are using this gap in a not very good way so we have to be careful because I may be educated in giving the data or not giving the data and having the choice to exercise whether I want the other person to access my data but most of the people may not be that educated so we have to be really careful and while building trust we need to have very strong laws to protect the sensitivity of this matter. I guess it's difficult to find the right sweet spot there between enough protection on the one hand and not too much in order to not lose out on some of the benefits of the data analysis that we can have with the use of AI for example. We do have another question from the audience and that question goes like this: "Other industries are ahead of healthcare in AI. Are there things they have learned or things that we in healthcare
can skip as this rolls out in this sector?" Maybe Jan would you want to address that question? Yes thank you very much. Look I think pretty much every industry is ahead to be honest. When I think about not just AI but digital transformation overall in the majority of countries, I mean people some people say healthcare is like 10 years behind, right when I think about finance, air traffic and other industries, banking so digital and AI technologies are used there right it's kind of standard and in healthcare we're just talking about the right introduction of artificial intelligence broadly so I think pretty much every other industry is behind that's just my opinion I don't know if you guys see it differently but yeah. Okay but that also can also be an advantage right sometimes being late and I think this is what this question also makes reference to gives you the opportunity to avoid some of the mistakes or the pitfalls that the front runners have fallen into. I mean look one of the reasons and it's a very valid one why we are late in healthcare is that, as we all know, it's a very complex and highly regulated industry which is good right, which is for the benefit of the patient all technologies we bring into healthcare need to be tested, validated, we need to make sure they work well, they don't harm patients at the end and actually do the opposite and therefore everything what we do here and especially getting AI to the market into healthcare settings takes time and is actually also very costly but there's a reason for this and it's a good reason. We do have more questions, at least one more and that's from Krish Dasgupta, Krish asks: "How do we ascertain equitable treatments in underserved communities where data is minimal.." and
we've heard about these areas both in India and in Africa and "...could synthetic data help in analysis for certain rare diseases?" Now I'm not 100 percent sure what synthetic data means in this context but maybe some of our panelists do so who would like to leap and address that question? Why don't I? Go ahead, please. So I'm not getting into synthetic data but I think it's an interesting question. How do we address this kind of underserved communities? So I want to give you an example of one clever trick that you so in India there's a tribal belt near the state of Kerala and this region is notorious for infant death so according to the number, despite India having good number, the infant depth in this region, this Attapadi region, is more than Somalia so that's not a good news. So what is happening is there was this this gentleman who created this particular tracker, you know a wearable tracker and you know it was a very beautiful design that he donated it, I think it's a startup called [inaudible] health or something like that so he gave this to the woman in the area but they were not willing to wear because they are tribal, follow the tribal culture and it's not in their line to wear this technology, these things. So what they did was they again went back to the drawing board and they made it look like a watch, now it's like a fancy watch, now they were getting data, they were all happy right they had kept these you know the broadcasting stations in the area so to collect the data wirelessly and everything was there but they were getting data but something was wrong with this data they keep checking and then they'll realise since it's looking this fancy, it's not worn by these women, it's their husbands, so their husbands took this thing "Why do you need a watch? You know I'm the man of the house." Right considering that there's a patriarchal culture here they started wearing this, they started wearing this and walking and they were getting the data from this husband; now then they went back to the drawing board and they created something called a Rudraksha Mala which is like a bead, which is a prayer beat in India it's like a rosary but bigger beads, so they created and they hide the chip in this this beads and they gave it to them and finally they found a solution. So that's one important thing right when you are
serving many of this underserved communities there's always a suspicion what they are trying to do with this technology, what, you know, what they're gonna take away from this technology so I think gaining their trust and finding this kind of a clever ways of utilising it is very important. Which leads us back to the human acceptance the importance of human acceptance. Sandra also something that you noticed in your work in Africa that people are not wearing the devices in the way you anticipated or maybe different people are wearing them for whatever reason? Indeed but it was only a few participants though and because we asked them also how they like the devices and you know how is everyday wearing going for them because you never know what other people may experience so we learned that one participant told us that it's like a fashionable item for them so she, you know, preferred wearing it to the market to kind of show it as kind of a piece of jewelry as far as we understood from her. For example in Kenya we had this problem with that we included in our first feasibility study also children, so below the age of 18, so we also had younger children like six years, and we found that one mother she was quite protective of the child so she started wearing the wearable device instead of the child and also for example in Kenya there's also regulations that school children that are not allowed to bring technology into the classroom so they are not allowed to bring smartphones and also watches, like wearables of course, so in Kenya we had to move our inclusion criteria to people older than 18 years because of that so certainly just this can be an issue. Just quickly coming back to the question I think we discussed already, like all the Resham told the stories about that it's very important to make sure that we have good data and it starts with the data collectors so educating people why it is important that we have accurate data because we want to build those systems and if the data is not you know accurate presentation of the reality that exists then our models will also be not good right they will not be a good representation of reality and therefore will not lead to the anticipated results that they could actually lead to and the same is true for synthetic data so we can only generate let's say just some synthetic data if the model underlying the data generation process is accurate of the actual model it represents right because the model is always an approximation of reality and this is based on how good the data is so I just wanna underline that we really need to make sure that we understand where the data is coming from, feeding those AI-based models because essentially they can only be as good as what goes into those models and we have found that to this regard currently AI-based technologies in Africa they're being drained with data from the US and which was a bit shocking but it's also obvious because we don't have good data coming from Africa and this is why and the same I I heard is a true for India and for the Asian context very much so we need to make sure that we have good data because otherwise those models will not work and it will only lead to frustrations and essentially in the healthcare sector even to death which can be avoided by having the data readily available and we also need to make sure that we actually know where the data is coming from, training those models because if we train it with data coming from a different context, particularly in healthcare, populations still differ right because each population has different diets, has different cultures, you know and therefore health is even individual I mean we are moving towards personalised medicine in some countries but in some we are lacking behind and not even having appropriate data to train the models with local data so we really need to be very aware of this fact and I think this is why also we need to formalise that into regulations and that we understand how those models have been trained, how they come about to make sure that those AI-based models are not biased for example.
But I understand this has, as far as I know, been an issue in in medicine for a long time also for example along gender lines right that a lot of the the medical research tended at least in the past, as far as I understand, you focus on male patients so they may not accurately reflect different, you know, biological realities of women in the treatment. I have Jan who raised his hand would like to contribute. Thank you very much I just wanted to maybe explain a little bit this term 'synthetic data' as there was a question around this. I mean look simply this
is like artificial data one of the struggles we have in healthcare right and this is different to other industries and maybe this is also the reason why we are slower than other industries in terms of AI adoption, is the fact that we don't have those massive data sets that are required to train those models right whereas I don't know in finance you have very structured data and you get easily access to this kinds of data. Here in healthcare we work with any sorts of data structured, unstructured and it's and ownership is often right not something difficult to get to, even we as GE Healthcare, we are operating in 160 countries, a global company but also we have of course limited access to this data that's acquired with our machines across the globe because it's not owned by us right, so the patient data is not owned by often by those AI vendors and therefore getting access to it is difficult and this is why synthetic data right, artificially generated data, becomes more and more relevant to train a powerful model. Fascinating topic I think we'll have to have another session on that, we're almost at the end I would like to make one final round among our panelists and my question to you is if there's one area when you look at digitalisation in the healthcare sector, one area that you believe that progress needs to be made urgently what would it be? So one point, please be brief and we'll start with Jibu. So I think the progress I will say not just from technological perspective because technology has been advanced right we have very smart AI models right now I think it has a lot to do with mindset and the infrastructure we need to develop it. So I think from a country like that, first of all there should be a framework, first of all there should be framework and guidelines to ensure that all these privacy, security, personal data, requirements are maintained so at the same time we need to work on building the public trust and the mindset, right the mindset is the most important part.
Okay thank you very much. Resham what would be your point? I'll agree with Jibu definitely and also health systems strengthening is very important with a solid, grounding [inaudible] and the reason why I'm saying that because it's important that we have the now evidence-based, data-driven decisions, data driven programmes and data driven policies in this space for having a sustainable impact and serving the last mile in India as well as other countries. Thank you Resham. Sandra what would you choose as the one crucial area?
Certainly I can agree with Jibu and Resham it's infrastructure so technological infrastructure to be able to use the systems and at the same time and just change management and openness towards technologies and being open to learn how to use technology so which requires to have an understanding so those two I think are key. Thank you and final panellist Jan, what would you say what is the most? Thank you I agree to what you said of course but maybe I want to bring in an additional point and maybe also as a call to action to the audience here if you know any, I don't know great software engineers, data scientists, developers that work in gaming, banking, any other industries I mean ask them what's their purpose to work in such an industry and explain them about the massive needs in healthcare because one thing that we see as well is that we do not have, right, in the healthcare industry overall, do not have enough of those fantastic technology experts to move the digital transformation forward so again if you have friends in those I don't know areas, in other industries talk about healthcare, talk about the importance, the purpose of what we do here and convince them they should have a look across the fans and join us in healthcare. Excellent that is a wonderful call to action and I agree that of course healthcare is a noble sector so if there are any I.T experts out there who consider switching sectors I think Jan
you made a very powerful plea and I guess the other panelists would also agree with that. This was a wonderful, very interesting session once again and once again time was too short and it just flew by but that's, I guess, also a good sign. What I found interesting among the many things that we discussed is that actually we the least amount of time we discussed technology per se so we've spent a lot of time talking about the human factor acceptance how do we make people work with patients or record data in the appropriate way understand the benefits and so on and so forth, so I think maybe that's also a positive sign for all those people who think that AI and technology will make us humans obsolete at the end of the day we spend a lot of time discussing how it serves us as humans and that obviously is also how it should be. I would like to thank all four of my panelists very, very much for this engaging, very interesting discussion and of course to all those listening and viewing this session for being good viewers and asking questions. This was our session for today and next month we will be back with
another installment of the Indo-German Business Talk, thank you very much and have a great day.
2023-06-18