#14: Is AI the Ultimate Key to Unlocking Better Clinical Outcomes in Healthcare with Anish Patankar

Speaker A: You have to make sure that you have these best tools. You have to make sure they fit in with the incentives that every stakeholder has. That's going to be the key in boosting the adoption of AI, uh, for clinical outcomes and preferences.

Speaker B: Director of GE Health Cloud and a

Speaker A: visionary in digital health, Anish Patankar brings deep expertise in cloud, AI and data driven healthcare. He's passionate about turning complex tech into simple solutions that improve lives.

Speaker B: Healthcare industry is a very regulated industry. What can AI provide in terms, um, of uh, what will work for a patient?

Speaker A: Getting best access to all the recommendations, looking at patient data across the spectrum. I think that's what AI really allows us in the healthcare industry to do that. So that's very, very powerful. Even in markets in underdeveloped areas, SaaS and Cloud has made such a big difference where, you know, all you need is.

Speaker C: Hello, welcome to the Priyanka Shinde Podcast where we dive into the art and science of strategy, execution and leadership. I'm your host, Priyanka Shinde, author of the Art of Strategic Execution and a consultant for tech leaders. In this podcast, I bring you candid conversations with top executives, startup founders and industry experts who are driving innovation, delivering results and staying ahead in today's competitive landscape. We will break down what it takes to tap into founder mode and turn vision into action. If you are serious about elevating your leadership and execution skills, make sure to subscribe on your favorite podcast platform or YouTube. It helps the podcast and ensures you never miss an episode.

Speaker B: Hi, welcome to the Priyanka Shinde Podcast. Today my guest is doing something very special with technology and AI. Welcome Anish Patankar. Anish is the senior Vice President and General Manager for oncology software at Elekta. He's a health tech expert specializing in cloud, Internet of things, big data and artificial intelligence. From building leading edge digital transformation solutions to delivering revenue growth, Anish has experience spanning pivot digital health, tech, medical devices, education and security software industries. His work is powered by the belief that even seemingly infinite and unstructured data sets have patterns hidden within. And harnessing that data into a form that impacts the human experience is where he does his his best work. Anish, welcome to the show and I'm really excited for this conversation.

Speaker A: Thank you, Priyanka. Very happy to be here and look forward to chatting more with you.

Speaker B: Yeah. So tell us a little bit more about Elekta and your role.

Speaker A: Sure. So Elekta is a Swedish company and we've been around for more than 50 years. Um, the company Started with a very flagship product called Gamma Knife, uh, which was a pioneer in neurosurgery, essentially treating, uh, brain metastasis, which is brain cancer, which is spread to different parts, uh, in your head and neck area. Right. But since then, Elekta has been a leader in what we call radiation therapy, meaning treating cancer cells using, um, radiation or using a linear accelerator that blasts high intensity radiation to kill cancer cells. So we have been at the forefront of this frontier in medical physics for a very long time. And personally, I am a software person. Uh, I'm a technologist at heart. So I lead the oncology informatics division at Elekta. And by oncology informatics, what I mean is software that powers all of these devices, uh, software that is used for planning treatment using radiation therapy, as well as oncology electronic medical or electronic health record software. So think of it as a specialty EMR or a specialty ehr that, that we use to deliver to clinicians. So for those of you who you go to a doctor, you get your chart and you look at your patient records, think of it as a specialty oncology electronic health record. So we do quite a bit in the healthcare software space as well. And I lead the software division at

Speaker B: Elektao, and thank you so much for sharing that. And this is truly where I feel that technology plays such a pivotal role, is saving human lives or fundamentally impacting how we as humans live on this Earth and how technology can really help us either make our life better or prolong our life. And I think what you are doing and what Elekta is doing is really making so much difference in the lives of so many people. And we know that cancer is, you know, something that is so difficult to, uh, beat. And all of the work that you're doing through technology, through all of, all of the different ways is truly, truly commendable. And thank you so much for sharing that. And I'm sure, like you, like you say, your technologist at heart, and there are so many directions that as a technologist you could have gone to into what drew you into healthcare. Uh, uh, at the beginning.

Speaker A: Yeah, I mean, that's a very interesting question. You know, I come from a family of physicians. My, my dad is a, uh, is a dermatologist and a pulmonologist. My grandfather was a doctor as well. Of course, my family always wanted me to be a physician myself. But, you know, when we were growing up, as you know, computer science was a big deal and I was drawn to the lure of, of computer technologies. Right. But when I graduated and this was purely by circumstance that I got my first internship in a medical imaging startup in New York. And since then, I have not looked back. You know, I have been in healthcare tech since then. And then I've worked in different therapeutic areas like cardiology, radiology, oncology, and so on. So now, you know, I've, I. I always tell my family that I have the perfect combination of what I have wanted to do, being a technologist in the healthcare space, which is what they wanted me to do. So I have not looked back since then.

Speaker B: That's. Yeah, that's a good way to sort of. Yeah, bridge the gap. Right. And make them happy and make yourself happy. I mean, just like with a lot of different fields, every field has. Technology has become so much more integrated into every field, even every doctor now, every physician needs to understand how technology plays a role today than it did before. Even surgeons probably leverage it so much more. So that's really great that you could do that. And then how did, how did, once you got into this particular field with medical imaging, how did it, it intersect with data cloud and now AI?

Speaker A: No, uh, I think, you know, this industry, traditionally healthcare industry, is slow for adoption of technology, right? I mean, typically the disruption happens in finance and other sectors and so on, but with data and AI, I think healthcare is poised to have so much improvements. I mean, I won't use the word disruption because that's not a word we like to use. We want the doctors to give the best treatment and patients to get the best outcomes. But, but just imagine, right, A patient walks into a clinic with certain ail, right? And there are so many things that a physician has to consider. They have to look at the patient's past history, what the patient has, the patient's genotype, what is the genetic makeup of the patient, and that comes from, you know, being of a particular ethnicity, having certain conditions and so on, uh, the patient's, uh, phenotype, things that are really presented by the patient, right? Meaning what environment they come from. Then social determinants of health, you know, where do they stay? What access to healthcare do they have? Are they staying in an area where there's quick access to healthy food supply, access, uh, to exercise, access to open spaces? And then, of course, then there are other things, um, like what symptoms the patient is manifesting. So if you look at the data parameters that a physician takes into account, and they know all of this just from experience, that they look at all these parameters and then they combine and say, in a patient, you do these tests, you look at your lab tests, your genome tests and so on. And the physician is doing so much intuitive calculation of combining all these factors and then saying, I think you should go for this treatment, or I think you should go for this treatment. Now, uh, imagine what I can do with all this. So if you have an intelligent AI models that can take into account patients history, genotype, phenotype, lab results, testing, social determinants of health, where do they stay, what access to health care they have, and now in the US Especially, also what insurance they have, Are they on Medicare? Do they have private insurance? What access to care can they get as a result of what insurance they have? Putting all this together and then determining which treatment can lead to the best outcome for this one patient, that's the power AI and combining this data can give to the physician as a recommendation. And I think that's immense power that is at the disposal of the physicians today, which previously was completely manual. So I think there's so much possibilities that we are starting to unlock in health care, and I think the potential is just tremendously, uh, in front of us.

Speaker B: You're so right. Like there's physicians can do so much, but now there's so much more data today than there was before. And something that you have mentioned is that you really like looking into the patterns within the data itself, right? Like all the hidden patterns, and your job is to harness that. So tell us more about that and how you're doing that and what's an example example of that, Right?

Speaker A: So I'll give you, uh, maybe a couple examples, but I think, uh, AI can really help with these patterns, right? I mean, previously, and I'll give you a couple of examples, right? Uh, if you are a physician, let's say who's in the Bay Area, in San Francisco Bay Area, and you've historically been treating patients of a certain demographic. And now, you know, as you know, this area is, is ripe for immigration. Now you start getting a lot of patients who are of a different genotype, of a different ethnicity, and the symptoms of the same disease may present themselves differently. But this one physician may not have had that experience in the past, right? So the pattern recognition that this physician is used to is based on different data sets. They have been trained on a different set of patients and so on. Now, of course, the clinical literature takes all these things into account, but with AI, this information is just more readily available that this kind of patient with their history, with their genotype, can present in a certain way, and then the outcomes are the kind of Treatments that these patients respond to can be different. Right? Uh, the other example I can give you here is what we call, you know, how do we elevate patient engagement with AI? So for example, when patients are on treatment in cancer care, right. In oncology, patients also have preferences because, uh, depending on what the age of the patient is, what is their tolerance and threshold to pain, what kind of side effects they are willing to live with, what is the compromise in the quality of life they're willing to have, their risk profile will be different, their tolerance for risk will be different. And accounting for these patient preferences, not only in the beginning, before you start treatment, but while they're on treatment, like in radiation therapy or in chemotherapy. Chemotherapy meaning, uh, drugs are delivered to patients, the patient using infusion, right between, between two sessions of treatment, the patient is at home, and the patient really doesn't have any contact with the care team unless they call and, you know, tell, tell the doctor's office saying, I'm not feeling well or I'm having this side effect right now with patient engagement, they can real time report what the symptoms are of the side effects and the care team can sort of look at them in the patient's chart. But more importantly, with AI, we can already flag these to the care team saying this symptom is a normal symptom. We can even tell the patient saying, this is expected, so don't worry. Or this is something that's looking a little bit out of tolerance. So not only wait for the care team to call, you just pick up the phone and call them right away. And on the other side, we can alert the care team saying this patient's symptoms are out of tolerance and we need to address it right away. Call the patient in first thing tomorrow morning. So making a real difference by looking at these passions and uh, these patterns and thresholds that uh, you make a difference in the quality of life for the patient, determine better outcomes for the patient and reduce the wait time before they're seen next by the clinical care team to make a difference in adjusting the treatment going forward. So really looking at patterns, looking at data historically and currently, AI can make a big, big difference in sort of how patients are being treated and leading to better outcomes in the long run.

Speaker B: I think something you mentioned is like having more of a proactive approach to managing the health of the patient or either the diagnosis or the prognosis, uh, for that patient.

Speaker A: The important part is that, you know, proactive approach, but also determining that intervention only when needed, not otherwise. Right. Because otherwise I Mean, the clinicians are just bombarded. They are overworked, too much stress. There's a real, uh, challenge we have in the physician community that physicians are burnt out, right? So giving them the tools to say that you only need to intervene now, they don't have to look at every single patient symptom coming in, but of course, the care team will look at that. But elevating, sort of saying, look at this now, look at this later, and so on and so forth. So really making sure that the patterns are making sense and determining the level and the timeliness of the intervention. I think that's really the key here.

Speaker B: So it's almost like it's for everyone. It's not just the patient, and it's. It's every person in that sort of healthcare workflow, if you will, that is impacted by the advancement of data and tech in this situation. And I was just reading today that there is, I think it was some hospital or something, I don't remember the name, but with the use of AI, they were reduced, like burnout of clinicians by 70% just because AI was taking notes. This is very mundane. We talk about us taking notes with AI, but this was just clinicians just because they were there taking notes, uh, during, uh, either patient conversations or something like that.

Speaker A: No, but believe it or not, this is one of the most leading innovation with AI in healthcare. So sometimes, you know, we always think of healthcare and AI as something that will help with better clinical outcomes and better clinical efficiencies. But when you talk about the physician burnout and stress, AI is helping tremendously on the operational efficiencies. So previously, if you remember, if you go to the doctor's office, half the time the doctor is busy charting everything. So everything you say, they have to record. Because the next time you come in, they need to know what happened. Ask them. And then that's the. Previously they would write it down, now they have to type it. But now you have what you call ambient listening, so voice recognition. And it's not just, you know, just recording what you say. So they listen to the patient and they translate it beautifully. The AI translates it beautifully, mapping it to patients, you know, existing clinical history, what symptoms they have, and then also sort of holistically saying what goes into the patient notes, what are the patient results? And the. And the physician really doesn't have to click 200 times to say, this is an allergy symptom. Mark this as a side effect. Mark this as a side effect of drug A combining with drug B. So they of course, have to make sure that the AI is doing the right thing. So they just have to look at it and say, just change this or this is all good and approve it. So it makes a tremendous difference in allowing the physician to focus on the patient rather than focusing on the tools and that. That's a big deal.

Speaker B: That's true. And I think it makes the experience better. Because who here has not felt like, you know, the doctor doesn't really, like, listen to me or support me or like, you know, is not really, uh, I don't know, like they just think I'm a 15 minutes and then I'm done or something like that. So it does improve, I guess, the doctor patient relationship.

Speaker A: Absolutely, yeah.

Speaker B: Something you said earlier, which also brings me to this point that you were just making, is that usually the healthcare industry is a little bit slow in adopting technology and there are very much valid reasons for it. That being said, how have doctors and patients both adopted this technology? So, for example, the ambient listening you were sharing, how have doctors or hospitals, um, physician offices taken on this technology itself? And how have you seen, what are the challenges in this AI adoption, uh, itself.

Speaker A: Right. So I think this is the most important question, Priyanka, that you've asked. Right. I think the technology is certainly improving. We see that in every day, in every facet of our life. We are using it on our phones, at our fingertips to do so many things right. When it comes to healthcare, it is ultimately a mission critical process. It's a mission critical workflow. So the adoption will come when there are enough proof points that it is actually working. And there are many facets to this. Right. One is, I think there are a certain, there are always early adopters in any industry, in any domain. So there are many physicians who are super excited about this and are the early adopters who set the bar. And then like in any scientific field, they are the ones who will then publish saying, this is something that's really helping and this is the right direction. The accuracy levels are so on and so forth. The adoption on the operational side is even higher, like the ambient listening and then sort of helping with patient charting and so on. Right there it's just a matter of saying, let's sort of introduce these tools and technologies in front of the physician, let them use it and let them see the benefit of helping them get better listening and talking time with the patient, reducing all these clicks and typing that they have to do. So I think the adoption there is really much higher. The challenge to adoption is more on the clinical recommendations because you know, and it's not the tech, because it's also in the science on the healthcare side saying that if someone is a cancer patient and they go to an oncologist A in clinic A versus oncologist B in clinic B, and I'll just make it sound very simple, but starting with chemotherapy and then doing radiation versus start with the surgery and then follow chemotherapy, how do you know that what treatment A is going to be better than treatment B? And the science is not there yet to just say that treatment is better than treatment B. So then the software or the AI tool recommending that go to go with this particular treatment regimen versus another treatment regimen. It's not proven for the physician. And I think the path to adoption there is to say why is the software recommending treatment A over treatment B? And then showing that it's referencing the following clinical literature, it's following sort of what data set has it been trained on. So showing the makeup of the recommendation rather than saying this is my recommendation, that's going to be the key in boosting the adoption of AI for clinical outcomes and preferences. That's going to take a little bit of time. Physicians getting comfortable with the recommendation, physicians knowing why and what's making up the recommendation. It has to be really solid clinical literature, clinical guidelines backed data. And then over time sort of showing that, you know, for a particular patient case in a particular setting, this outcome has led to better. This uh, treatment has led to better outcomes over time. So I think that's going to be the challenge that we have to deal with here.

Speaker B: So similar to how there are clinical tests that are run and you see results and based on those results is then you say, okay, now yes, we have proof and based on this proof we will go forward. There is proof that is needed in order to adopt certain aspects of the technology for certain uh, directions.

Speaker A: Right. And you know what AI does here? You know, traditionally you go in healthcare, you go through randomized clinical trials. So you sort of, you know, to put it simply, if you're a pharma company, you have a pharma drug trial going on and you compare the outcome of patients taking drugs with patients taking a placebo. And then it's randomized, it's a clinical trial, very controlled. And then you look at data statistically to say does it make a difference to what extent? And so on. What AI allows you to do is just monitor patients who undergoing different treatments with whatever they have. And if you collect this data over a much bigger Size between different clinics and so on. This allows us to really hardness what we call the real world evidence. Meaning this is not a controlled randomized clinical trial, but you're looking at which type of patients are getting which type of treatments and which treatment is leading to better outcomes for which patient. So finishing that loop and then really sort of coming up with recommendations and evidence based recommendations, what we call real world evidence. And I think AI really allows us to do this at scale. Trying to do this in a controlled setting in one hospital is different than trying to look at data across hundreds of thousands of patients. And I think that's what AI, uh, really allows us uh, in the healthcare industry to do that. So that's very, very powerful.

Speaker B: Definitely. I think yeah, there's the whole clinical trials and that aspect is probably just uh, ripe for, you know, the scale at which it can be taken off. Uh, with AI, I think definitely something that we can probably get into. And that can be a whole other discussion, I'm sure. But I did want to bring it back to yes, there is this aspect of like we need a lot of um, evidence and proof in terms of also what can AI provide in terms of um, proof of what, what will work for a patient or not or things like that for the physicians to be able to use recommendations. And at the same time healthcare industry is a very regulated industry and there is a lot of concerns around privacy of data and of the patients themselves. So with that itself, how do you see that impacting the adoption of AI?

Speaker A: Right, so I think the, I mean maybe two parts to that Priyanka. One is that uh, the FDA itself, the Food and Drug Administration is looking actively at adopting AI to help with sort of, you know, all the innovation happening in healthcare, be it drug discovery, be it launching new techniques, being looking at new medical devices. So that's very encouraging that um, the FDA is also looking into this and promoting the use of AI. And the second part of your question is around patient privacy. Now that's super important thing and very critical that we maintain patient privacy in all of this. And there I think the, I don't believe AI is going to cause a disruption one way or another because there are existing robust practices, uh, in every country, in every region that govern how clinics and vendors deal with patient data and patient privacy. So in the US we have the law called hipaa, you have GDPR in Europe, uh, that uh, really streamlines how you take patients, uh, consent to use your data for a particular, be it for research, be it for any other purpose. So I think those safeguards are in place. So that does increase a little bit of the burden in sort of how you gather the data to be used. But that is already in place and I believe AI will only make things easier, right, in that aspect. It's not going to sort of complicate things, uh, any further.

Speaker B: I guess that's a good thing because like you said, there's so much data, uh, and AI can really help to harness the insights from that data at scale, which would really help us as a humanity, uh, itself, uh, in terms of discovering what all can be done. Maybe it's finding cures and things like that moving forward. As you are in this space working in the healthcare industry, you obviously are in conversations with other healthcare experts. What do you see or what might be some common, say, mistakes or pitfalls that health tech companies should be, um, wary of, especially when thinking about AI?

Speaker A: Yeah, yeah, I think that's a good question. Right. I think sometimes, uh, there's always a technology hype, right? We see that in the Silicon Valley all the time. And sometimes we've seen that companies sort of bring the Kool Aid to quickly and try to attack the most difficult problem right away. And I'll give you an example with, you know, with IBM's Watson AI. If you remember a few years ago, there was a big sort of push from IBM to use Watson AI and everything. And they entered the healthcare space too, and they tried to attack the most difficult problem of giving physicians clinical recommendations on the best treatment courses right away. And this was a few years back, right, when AI was also still developing as a technology for, uh, healthcare. And I think the adoption just didn't happen. And part of that was because, like I said, uh, it's because you don't really know why Watson is recommending A over B. So the makeup of the recommendation not being there, rather than taking incremental steps towards that, sometimes incremental is not bad because you have to change the way of working of physicians in terms of having them trust the tools and secondly, giving transparency. I think this is the most important aspect, right? That if even if the physicians want to trust the recommendation, you need to give them verification saying why A is better than B in this case, because in another patient case B may be better than A as the treatment outcome. Right? So showing sort of what is making up the recommendation, right? So I think those two things we saw in one example there that tech companies really sort of going all in without understanding the nuances of what makes up sort of the decision making for a physician and then helping them in that direction. Now let's say in the consumer world when you search on Google saying oh how do I do something or what's the best way to do something? You look at the Google recommendation or you know, using any AI search engine, you're not really always looking so much as to why Google is saying A is better than B. And now even in fact with Gen AI, even Gemini and so on, they started showing the sources of that recommendation. So they're doing that in the consumer world as well. But doing that in the healthcare space is tremendously useful. And that's where I think uh, the health vendors, health tech vendors have to be really keeping the emphasis on that. The, the, the recommendation they give is based on clinical literature, peer reviewed articles and so on and not just read it forums where you know, you have to make sure your recommendation is coming based on solid clinical evidence that published, peer reviewed and so on as well as from industry bodies like in the case of oncology, it's National Cancer Institute and so on, not just something else. Right. So that's, and that physician needs to know that your recommendation is grounded in scientific evidence and literature, not just something that it's reading off the Internet.

Speaker B: Right.

Speaker A: So that's critical.

Speaker B: Yeah, I think um, it's, it's a very important point that you call out because we always look for evidence like why and it's as, as humans we want to know the reason behind something also because AI can be very non deterministic at times and this is where it's required until it gets better. Till. Because especially with AI, ah, there's so much hallucination still happening that if you don't provide that I think proof behind the scenes of the recommendation and I think that is why it is hard as of today. But hopefully that's what gets better. And this is where like you said, companies need to be very careful in not sort of following the trend.

Speaker A: Yeah, yeah. And then m. Maybe just one more point to add that you know, um, even if AI is right almost every single time, if they give a wrong recommendation, even one time and that leads to certain outcomes that are not desirable, that's gonna cause the trust to go away rapidly. So I think. And that's where the mission criticality comes in. Right. You have to be right almost every single time. It's not being right 90% of the time or 95% of the time. Right. So for the physicians to develop trust, you have to get it right or you have to not give a recommendation. Rather than hallucinating. Right. I think that's the danger we have here with AI hallucination, that if it doesn't know where to go, it will make things up. You don't want that. And of course, there are safeguards that the vendors are putting in, in place for all these. Now these are being done by the tech companies, but then health tech vendors who use, uh, the tech, they have to put additional safeguards in the healthcare space on top of that. So that would be another area that we have to watch out for.

Speaker B: Yeah. It's so fascinating to see all this. Right. Because it's a matter of life and death. You can't like, lose trust. Like one, one offense is not okay. I mean, if you're just like doing some random, uh, content generation, it's okay if it's hallucinating or something like that. But in this situation, like if, if a doctor breaks your trust, you're not going to ever go to them again. So it's, it's. Yeah. Uh, this is such an important point I wanted to also touch upon as we, I know we talk about trust and we talk about all of this tech advancements, but there's something that you had mentioned, uh, in your work is this gap, uh, in healthcare, especially where tech is concerned, in urban areas versus rural areas. Right. Like there is a healthcare access in rural areas. We have a gap there. And then when it comes to tech, that gap sort of widens. What's your take on that? And how do we continue this advancement without widening that gap?

Speaker A: I think this is something unfortunately that we see. Right. It's urban and rural, it's developed markets and underdeveloped markets also. For example, if you look at the space, uh, I work in radiation therapy, Almost more than 75% of access to radiation therapy care is only with 25% of the world's population. And that's why even, uh, as Elekta, we really have this goal of increasing access to radiation therapy worldwide. And we have this program which we really promoted and worked on for the last few years. And even if you see even within a developed country like the United States. Right. The access to care between urban and rural areas, the gap in care is very high. And technology, unfortunately, sometimes instead of helping us bridge that gap, can widen it. Right. In unintended ways. This is the law of unintended consequences. Right. And in many ways I think tech is now making it better. But just as an example, and when I say tech, it's not just tech in terms of, of, uh, access to computer, uh, tech and so on. It's sort of how you deliver technology, how you deliver these cutting edge AI solutions and so on. Right. So as we put in a lot of uh, R and D into AI, these AI based solutions, just as one example, are expensive. And buying prowess of clinics who are in more established urban centers is much higher and they will have these technology resources at their disposal to help the clinicians. And if you go into rural counties where you have community hospitals, smaller clinics, they cannot afford these kind of, uh, investments in these expensive technologies. So now tech is actually increasing the uh, best access to care gap. Right? Because you not only have the super specialist in the urban centers, you also have these super specialists with the best technology tools only in these centers and not so much in the rural setting. Right. So how do you bridge the gap? Right, because one is of course an economic problem in terms of, of uh, the investment and the pricing of these tools, uh, and second is also where the population centers are, where the super specialists tend to be and so on. So I think one thing for us to look at is to say how do we sort of bring, uh, consistency in terms of the AI tools that we roll out, uh, at a price point that's affordable. How do we make sure? Because AI, if you imagine that AI works very well in operational efficiencies, clinical efficiencies, and assuming it's deployed evenly across community hospitals, small clinics and super specialty clinics, then I think it actually bridges the gap. Because then a physician sitting in rural Montana can give the same recommendation of treatment that you may get in a highly super specialist clinic in an urban setting. And that's really bridging the gap. Right. Of care. That's where the promise is. We just have to work as an industry to make sure everybody has access to these tools, tools and that we have to make sure they're affordable, that they can be deployed in all these settings. They can be deployed in a setting where they're offered, um, a software as a service over the cloud so that the burden of deploying them is also lowered. So I think this is where most of the industry is now focused on. And it's very, sometimes very heartening to see that even in markets, uh, in underdeveloped areas, SaaS and Cloud has made such a big difference where all you need is Internet connectivity to connect to the systems, to get yourself started, right. In terms of the software piece of it, right. And then you can sort of really put more innovation, uh, continuously, uh, and give them access. So it's, it's very promising. The cost is always a challenge. Right. But at least from us on the software side, I think things are uh, there's tremendous potential to improve and bridge this gap.

Speaker B: Yeah, there's always, I think technology can do so much more if you can think about how to operate in the constraints that are there. From the environment perspective. It's sort of like I got reminders like a lot of these rural places now obviously smartphones and cell phones, uh, have become very common and sometimes we can't log all of these huge devices somewhere. But can you do something with this small device like a phone and collect all of that data? And I think that's what you're saying is if we can just get something through SaaS or cloud and just get that information through, then you can deliver a lot of that health care to these rural areas that you could not have otherwise, say a few years ago.

Speaker A: No, absolutely. I think we cannot put heavy duty medical devices there. That still has to be done, but at least decision making, getting best access to all the recommendations, looking at uh, patient data across the spectrum, I think software has tremendous uh, promise to help in these areas.

Speaker B: You have, I mean you have been in this healthcare world for so long now, uh, and seen a lot of different technology evolve over uh, so many years. What is a, uh, belief that you held early on in your career about healthcare or healthcare tech that you have changed your mind about over the years?

Speaker A: I think when you're young and when you come as a, uh, as a technologist to any field, you know, you believe technology can solve all problems right? In any industry and you see that mindset even today sometimes. But I think having worked in healthcare for more than 20 years now, I think technology can solve access problems, can automate things, can make workflows better and so on. But ultimately I think the incentives in the healthcare system is not something technology alone can solve. And what I mean there is uh, the payment models, how are uh, all the stakeholders, the providers, the payers in the healthcare system, uh, what are their incentives? That drives a lot of the workflow. Right. And that's not something technology can change overnight. Where technology can help here is to provide more transparency for patients in decision making, provide more transparency for clinicians in their decision making. So I think these things are making differences now, but I think uh, you have to work in this industry to understand all these competing incentives at play, so many stakeholders and how they operate and what limitations technology has to face when trying to change all these things in the system. So that's one thing I've learned over time, and that's also very useful now for me in this industry, because you know that it's not just about rolling out a set of tools and expecting adoption and expecting workflows to change. You have to make sure that you have these best tools. You have to make sure they fit in with the incentives that every stakeholder has in that system. And ultimately, how do you drive towards making outcomes better for patients? Because ultimately, that's the end goal for everybody, is having better outcomes for the patients. But it's not a linear path that you just put something and immediately you start have patients deriving the benefits. So I think it's a hard lesson for some people who are new to healthcare to adopt and learn.

Speaker B: When you talk about incentives and payments models, I think, I think I know what you're referring to, but can you kind of share it a little bit more explicitly?

Speaker A: Yeah, sure thing. And you know, it's not controversial as such, right. I think everybody, uh, who is in the healthcare space knows this, right? It's sort of what drives payments. Now, when you talk about, when you go, there's the providers, which is the hospitals, the doctors and so on, then there's the payers, the insurance companies and so on, right? So everybody wants to make sure that the money is spent in the best way possible. So as an example, right now in oncology, immunotherapy is picking up, has tremendous promise. Now, the insurance company will always say that they will allow immunotherapy only after you've exhausted and tried the conventional therapies, because immunotherapy is tremendously expensive, even if the outcomes are tremendously better. But on the flip side, immunotherapy only works on a certain set of patients, not on everybody. And then proving which patients will respond better to that treatment, sometimes there are biomarkers. By biomarkers, I mean some certain genetic mutations that the physician can say, okay, if this mutation is present, we are pretty positive that the immunotherapy will work. But if not, then the science is not fully ready to say, okay, will the patient benefit from it or not? So even if the physician wants to put a patient on this immunotherapy, in the absence of a biomarker, for example, the payer or the insurance company may not be ready for it right away. So the incentives may not be fully aligned yet. Right? So the patient may have to prove that they've gone through conventional therapy first, that did not work. And then you go to immunotherapy which takes time, can have other consequences and so on. So that's just one example of what incentives are in the system you are, and this is in a certain market, in another market, maybe in another country, uh, the payment models are different and maybe access to immunotherapy, maybe it's easier or maybe even more difficult as a result of that. So the stakeholders that are a little bit different in a socialized medicine concept versus in a private payer model concept. Right. And in another setting, if it's all patient funded, then it's wholly different. And it's like okay, just directly a conversation between the uh, clinician, between the physician and the patient saying what can you afford? If uh, you go to India, that's very much the case where the physician will say you can do A, B or C, it will cost you X, Y or Z. Here it's saying okay, if you. In the US it's more to say, okay, your insurance covers this or doesn't cover this or will cover A only if you do B first and B doesn't work if you go to uk in a uh, socialized medicine concept, everybody's in a queue and that just takes a matter of time. And then, okay, they only have three therapies that are sanctioned and everybody, you only have the choice of that. Right. So that's what I mean about sort of deploying the tools in these, just the three examples I gave you three very different social settings, very different stakeholders and very different incentives at work world. And the adoption of your technology is going to be different in all of these three cases as an example. Right. So understanding that, learning that and then sort of, you know, uh, the adoption curve is rapidly, uh, dramatically could be different in these three settings as an example. Right.

Speaker B: It's a lot about like when you talk about when you come in and you think about tech and you are very much enthusiastic, it's going to like solve all of the world's problems. But then you kind of like slowly realize what the truth is. And, and uh, thank you for sort of expanding on that sort of payment models and incentives because yeah, like every country is so different. So what might work in one setting is not necessarily going to work in another. And we hope, looking forward, like I do hope that even AI or like some of this tech can help solve where say these payment providers or insurance, like they understand like here's how AI can actually tell us what might be actually better or something like that. Right. Or it can actually solve these problems that as technologists we very Worse, because I think technologists are also much more optimistic. If you think about AI and how AI took off after 2022 and all of these startups this came up and they thought, oh, AI is going to solve every single thing. And I think now the reality has set in a little bit.

Speaker A: Yeah, nobody. We are very optimistic and as technologists, we have to be optimistic that things will get better, things will get standardized and the adoption curve will take off. Right. Sometimes it's as there's more of the hype cycle and then I think we'll come to the plateau of reality and adoption and we are getting there much quicker with this technology than in, um, previous cycles. Right. So I think the optimism is still there. I think the reality check will come in and I think then the adoption will just take off and we are very, very optimistic and confident that will happen.

Speaker B: On that note, what excites you most about the next, say, five years of AI in oncology or just healthcare more broadly?

Speaker A: I think what excites me, or maybe before I answer that specifically, the challenge in oncology always has been managing personalization of care for a patient. Everything we said with the patients, uh, all the patterns, the examples we spoke about, what's the best treatment for this patient in front of the physician, that's one really the holy grail in any healthcare setting. On the other hand, there's always been this challenge that as you go into community hospitals and so on, on, it's about productivity. How can you treat more patients with less, uh, healthcare expenses are through the roof. As you know, we are spending close to 20% of our GDP in the United States on healthcare. It's trillions of dollars now. Right. So how do you achieve more productivity? And productivity by definition comes with standardization, comes with consistent workflows and so on, which sometimes can be in conflict or at odds with personalization. So, so there's always been this tussle between if you dial up personalization, your cost goes up in terms of customizing treatments and so on, versus if you dial up the productivity, then it's like, okay, same workflows and same standard of care. So it's always a little bit in conflict. And with AI, if I look into what excites me into the future of AI in oncology, it's just that I think we can get to a more better harmony and balance between personalization of treatment and productivity for the providers. I think that that's the holy grail is achieving. It's not dialing one up and one down. I think we can dial both up at the same time. And that I think AI can really help us to say, you can have standardization of care, you can have consistency, you can improve productivity using these ambient listening tools, interpretation of data, bringing in all the patterns together, recommending treatments, predicting what outcomes can be, while giving insights into what can work for patient A versus patient B, and giving personalization that is super exciting for us in the future.

Speaker B: I've been hearing about personalized healthcare so much more in the last few months. And definitely, uh, I think that is an area that is, uh, exciting to me as well. Because you're right. And even some of the things you were sharing earlier about, uh, the genome or, you know, where do you come from, what's your genetic type, what's your environment like? There is so much that depends, uh, and is very specific to the person that the treatment actually has to be very much specific, specific to them. It doesn't matter, you know, that if this works a, works for somebody or not. And yeah, that's, that's really exciting. And uh, I'm sure, like, as you're looking forward and you're excited about a lot of things, you know, sometimes you feel like, oh, if I could just solve this one thing, it could be so much better. So if you could snap your fingers and solve like one structural barrier or like, you know, innovation in healthcare, what would it be?

Speaker A: You know, uh, just reflecting on everything we discussed, I think it's just getting everybody access to this AI, regardless of the financial incentives that we spoke about, regardless of urban setting or the rural setting, regardless of how expensive it is. If we could just snap our fingers and say everybody gets access to the latest technology in any modern setting, that would be one thing I will do with the snap of a finger, if possible. Yeah, if it was that easy. Yeah, yeah.

Speaker B: I think that, yeah, the equity, uh, in this, uh, equity in AI, ah, would be really powerful because then everybody is on the same playing field and you're on.

Speaker A: Yeah, everybody gets the best access. And then I think it's not dependent on where you stay, which doctor you see, which clinic you have access to, uh, what's your insurance payment system. I think if we can circumnavigate or cut through all of that and get the best access, that just would be tremendous.

Speaker B: Well, to Rabab, I know we didn't talk a lot about sort of know you. We talk a lot about the AI and healthcare, not a lot about you leading sort of, you know, product group, uh, in your work and leaders. And so as you Think about for product leaders building in this complex industry and healthcare, what do you think is like the one mindset that they need to succeed in this industry?

Speaker A: Mhm. I think maybe two things I would say, not just one. And they go hand in hand. One is being relentless. It's not about having the brightest idea or being the first to market and you know, the uh, winner takes all. It's not that it's about being relentless and being second. That goes hand in hand with that is being persistent and patient in pushing through like I think we discussed, all these different payment models, different incentives, driving through all that takes a lot of um, persistence I think being be persistent, keep your eye fixed, uh, on the bigger goal, but be flexible with the tactics and details. Uh, that is one thing I will say that is really required to be successful in the healthcare industry.

Speaker B: That's great, great uh, to know and I think great advice. Last question. Any other final words of wisdom, um, to anybody trying to build AI products for healthcare?

Speaker A: Like I said, I will just repeat saying be persistent, be relentless and have some patience.

Speaker B: Well, thank you so much. This was a great conversation. Really appreciate it and really appreciate all the great work that is really truly meaningful that you are doing with Alecta.

Speaker A: And thank you so much Priyanka. I enjoyed being on the show and good luck to you with the podcast.

Speaker B: Thank you. Bye.

Speaker C: Thanks for tuning into the Priyanka Shinde podcast. If today's conversation sparked new ideas or challenged your perspective, I'd love to hear from you. I work with tech leaders and founders to tackle complex challenges, whether it's scaling teams, making high stakes decisions, or turning ambitious ideas into reality. If that sounds like something you are navigating, let's connect. Reach out via my website, thepriyankashinde.com or find me on LinkedIn. Until next time. Keep building, keep leading and keep making an impact.