← Becker’s Healthcare Podcast
Loading summary
Transcript28 lines
- [00:00]
A
Hello everyone. This is Erica Spicer Mason with Becker's Healthcare. Thank you so much for tuning into the Becker's Healthcare podcast series. So today we're going to talk about informatics and AI in cardiology. And joining me for this discussion today are two wonderful leaders in the space. We have with us Dr. Jagmeet Singh, professor of medicine at Harvard Medical School and a cardio electrophysiologist at Massachusetts General Hospital. And Dr. Sanjay Gandhi, an interventional cardiologist and the chief Medical officer for Philips Informatics business. Dr. Singh, Dr. Gandhi, thank you so much for being here today and welcome.
- [00:38]
B
Thank you, Erika. It's a pleasure to be here.
- [00:40]
C
Thank you, Erica. You know, it's really great to have Dr. Singh and you on the line and I'm looking forward to a great discussion.
- [00:47]
A
Likewise. Really looking forward to hearing the expertise that I know you both have in this field and where you see AI headed in the cardiology space. And, and before we get into that, I wanted to give you both just the opportunity to share a little bit more about yourselves, your background and your current role. Whatever you think would helpful for listeners to know going into the conversation. And Dr. Singh, would you mind kicking us off?
- [01:10]
C
Sure, sure.
- [01:12]
B
So I am, as you stated, a cardiologist and a subspecialist. So I'm a cardiac electrophysiologist. I'm one of those docs who puts in devices, implants pacemakers, defibrillators and does catheter ablations. I spend most of my time looking after patients, doing procedures as well as doing research. And much of those research is in the AI and senso. And I did serve as the clinical chief of cardiology at Mass General Hospital for a few years in the past. That really, I would say, enhanced my interest in trying to find AI and sensor based approaches to enhance clinical care pathways.
- [01:52]
A
Wonderful. And Dr. Gandhi would love to learn about you as well.
- [01:58]
C
I am an interventional cardiologist, just a little bit different specialty than Dr. Singh, who is electrician. I call myself a plumber. I've been in healthcare for about 20 years. Twenty years, and most of it has been in clinical practice in hospitals, teaching training next and next generation of future cardiologists, but also managing service line operations while opening up a blocked artery in somebody having a heart attack is very gratifying. I realized that I can only impact so many people over the period of my lifetime. So I switched to a role in Philips where I can drive innovation and empower clinicians across the globe to provide better care for more people.
- [02:44]
A
Wonderful. Dr. Gandhi and Dr. Singh, great to learn more about you both. And I'm loving this electrician versus plumber metaphor. Very insightful as the moderator and I'm sure for our listeners too, but you both clearly have such deep expertise in the cardiology space, whether that's your clinical practice, research, technology and so forth. From each of your perspectives, what are you seeing, seeing as the key challenges and opportunities in cardiovascular care right now? And why is it important for leaders to have a keen eye on these challenges and opportunities? Dr. Singh, if you wouldn't mind getting us started.
- [03:22]
C
For sure, for sure.
- [03:23]
B
No, I think that's a really insightful question. You know, during my term as clinical chief, I found, and it still exists, you know, that care is very fragmented. And not only that, I also recognize that we're very inefficient and very inequitable in the way we provide care. And beyond that also I think there's a level of opaqueness. There's this opaqueness in reimbursement, there's this opaqueness and practice variance. And much of the care that we deliver today is very reactive. So I think there are opportunities to really make this reactive care more proactive, make it more individualized, enhance access, really create it to be more equitable. And that's where I think the role of sensors with virtual care, along with AI that is integrated into sustainable workflow, translate into better clinical outcomes. With respect to leadership, I think it's so important that we recognize that this is a change in the culture of how we practice medicine in the future, and we'll talk about that hopefully in the next few minutes. But also that there's a change in the value proposition as we embrace these digital strategies that in turn embrace the life cycle of disease. I think we have to recognize that investing in individualized care approaches, using multidisciplinary care is so key to really making care more equitable and more, I would say, associated with better outcomes. But I think it's upon the leadership to not only inspire and enthuse folks, but also align incentives and really mold care into more of a data driven strategy that translates into better outcomes.
- [05:17]
C
And I think I would actually also build a little bit on that. I agree. The care is fragmented, inefficient and probably inequitable in certain ways. And when I talk to clinicians, and I've seen in my own practice as well, there are three additional challenges that I would like to highlight. Growing staff shortages. I think, especially after the pandemic that has gotten even worse. The administrative burden on clinicians you spend more time charting things than actually spending time with the patients. I think the second aspect is the demand for healthcare, especially in cardiology, has gone up. We're seeing a lot of increase in chronic diseases, aging population that has now put additional burden on the health care systems and healthcare leaders to really meet that increased demand. And the costs are increasing. I think you referred to that as well. The reimbursements are falling, the cost of delivering care is increasing. So how do you truly bridge that gap? For health system to really not address this will make them irrelevant. I think health system need to embrace technology to help solve some of these access issues. And time is a critical resource, whether it is to see patients in office, what your schedule looks like. And I think for us to really leverage digital tools so we can remove those non essential tasks and essentially increase the capacity of workforce to deliver more care, I think it's going to be key. Also agree that we need to move from sick care to health care. I think a lot of time we focus on things when it is already too late. Health systems focus so much on taking care of patients who are sick than those who are actually healthy. And then lastly, I think patient expectations of healthcare systems are changing as well. If I can order something on Amazon at midnight and get it delivered next day, why can't I go see my clinician when I need to see them? We put patients on schedules. Well, they don't get sick on schedules. They need care when they need care. And I think that's a challenge that we can probably solve with some of the digital tools that exist there. So to me, I think there's a big opportunity ahead of us.
- [07:30]
A
Yeah. Dr. Gandhi, thank you so much for those additions. It's striking me in both of your responses you're touching on this need for highly individualized, convenient care that is satisfactory for patients, while at the same time provider organizations are dealing with growing staff shortages, administrative burden, higher demand and cost. There's a lot of tension between those two buckets. So I can see how technology is really coming in as this solution to make sure that all of these needs are being met. And we are certainly seeing a lot of buzz around. I know Dr. Singh mentioned them earlier, but wearables, sensors and AI in technology. So I'd love to learn a little bit more about the tools and technologies you're most excited about and the specific improvements that you think those tools will bring to cardiovascular care.
- [08:21]
B
So, you know, as an electrophysiologist, I implant devices and most of these devices already have sensor based Approaches that can measure heart rate, physical activity, transthoracic impedance, respiratory rate, a variety of sensor based approaches that can allow to, what Dr. Gandhi just said, provide continuous care because this data is remotely transmissible and is continually transmissible. So you can provide some element of that continuous care that Dr. Gandhi said that we really need to move from what is transactional to something that's more continuous. But in terms of just answering your question more directly, what are the tools I'm really excited about? So I'm wearing a ring, I'm wearing an Apple watch and an aura ring. I'm excited about these technologies because they help us transcend from wellness or prevent us from becoming unwell, towards chronic disease, towards getting more complex care. But I think the ones I'm really interested on are the patch monitors and variables that will integrate PPPG signals and ECG signals that will allow us to have hemodynamic surrogates that can provide much of our cardiovascular monitoring non invasively beyond the different measures that I really already mentioned to you. Now why is that important? I think that's important because that allows patients to be empowered with that knowledge of their data and allows for self management strategies to become common play in the future and also create closed loop systems where you have automated strategies that you can take care of disease and sort out the shortage problems that Dr. Gandhi just alluded to. So I'm excited about these technologies. I think there's a lot of time before they actually reach the point that they allow us to do what we just talked about. But I think we're moving in that direction and that direction is basically in preventing chronic diseases, early detection of diseases, and really preventing readmissions after patients develop diseases.
- [10:32]
C
You know, I agree with you doctor, we have a lot of exciting technology, whether it's variables or sensors that can really engage patients in their own healthcare journey. If I look at the other side of it from the health system standpoint, you know, all the data that has been generated by these variables and sensors. So how do we really address and use that data in a meaningful way? I think as a clinician, you know, having my patients bring in their Apple watch reading is great, but do we have the bandwidth to really address so much data that is coming at us? If I look at it from the other side of it, from the health care standpoint, I'm also excited about AI. We have over 1000 FDA approved algorithms that are now at least some of them are in actual practice and implementation. I'll maybe take a slightly different view on this from the health system standpoint. For example, in cardiac imaging, we can now automate positioning of patients and do image capture without necessarily requiring a tech to really position the patient in the right fashion. What that means for the patients is actually better exam, less radiation, less repeat examination. But it also allows for streamlined workflow and operations at the hospital level. If I look at generative AI in that space, generating reports for the radiologist or the cardiologist, I think it's a big value add for improving efficiency for them. But not only we improve efficiency of the reader, we can also generate reports that can be interpreted by referring physicians in the way they want it and maybe even for the patients so that they can understand what is being spoken in the medical language. I think ECG has been a great tool for us. We've always had digital tools to interpret ecg. You put an ECG and it spits out diagnosis on the top. But I think what I'm more excited about is how you can actually use ECG as a predictive tool as well. Not only read what it shows right now, but also use machine learning and AI algorithms to really predict whether somebody will have low ejection fraction when somebody has amyloidosis. What that will allow health systems to do is actually screen out patients and look at them early, but also patients who don't need additional testing. And maybe that also helps with the capacity issue that we were talking about. Lastly, as I was, you know, as we talked about, you know, there's a lot of data coming at us as clinicians. You know, you have electronic medical record, you have imaging record, your patient generated data. We just talked about how do you really, you know, while taking care of patients, keep up with all this data that is coming at you? The guidelines change, the clinical trials are new. And I think what I'm excited about is multimodal data processing that will allow us to integrate all that data, put that patient in the context so you can get clinical decision support at the point of care. So you're really not going back and looking at guidelines, but they're actually right at your fingertip when you need to make that critical decision for your patients.
- [13:40]
A
Yeah. Dr. Gandhi, Dr. Singh, I appreciate these robust examples so much. You know, we've covered everything from the multimodal use cases, predictive use cases, radiology example that you highlighted. Dr. Gandhi and then Dr. Singh, you really highlighted the importance of self monitoring tools for continuous care and how that empowers patients. You know, so many exciting developments in the space and a lot of it sounds like hope that you both have for how these tools can truly drive those improvements that you're looking at. But I imagine at the same time, adoption is something that's pretty top of mind for most clinicians and patients, as well as they're being introduced to some of these new tools. So, Dr. Singh, I'd like to kick this one over to you first. You know, as hospitals increasingly lean into AI, what are those top risks that you think they should be most mindful of or that you have a close eye on? And what really needs to happen to increase adoption across clinical practice?
- [14:40]
B
That's a great question, Erica. Actually, I was thinking about that as. As Sanjay was alluding to the different use cases. I think it's so important to recognize that the data sets that we use to really generate these algorithms really have to be unbiased. They have to be heterogeneous. They have to be representative of the patient population that we're examining that algorithm in. They need to be suitably validated because, you know, there's always the risk of having false positives, and. And that can alarm not just the patients, but also the clinicians. So. So I think validation bias are issues. I think as you get into these streams of continuous data, some of the other additional risks that one would encounter are risks related to privacy and security and data governance, something that was already alluded to, and I think the whole black box effect. Why do people not necessarily adopt some of these newer technologies? Because it's unexplainable. And I think the solution lies in the fact that we not only have to make this generalizable, we have to make it explainable. We have to make sure that they are validated and they are integrated into the workflow so that it is seamless. And most importantly, we have to ensure that it translates into better clinical outcomes for acceptance. Because, you know, clinicians don't change their practice on a dime. They need data to really help drive their change in practice that they've been used to for so many years.
- [16:17]
C
I think I would agree to all the points you just mentioned. I think it's not so much about the tools and technology, but also making sure that we have the clinical expertise to guide those tools and also use the tools to empower people, not replace them. I think one of the big factors in AI implementation is trust. Whether it is the clinician trusting the data and the output from a tool, or whether it is patients actually trusting that they will not lose time with their physicians if they use those tools. Does it take away from them, does it really impact that physician patient relationship? And I think that's one aspect that we need to address as well. I think the other is the integration of this in the workflow. I think you alluded to that it is a key that these applications integrate into the workflow. But think about from a second, from a hospital perspective, we have 10, 20 different vendors actually approaching us with algorithms that can do one or two things. How do you really validate and test it at that hospital level, whether those applications really make a difference to your patients? And I think we need something like a platform approach where you can very easily integrate these applications in both current and future workflows. I think one other thing that we haven't so far also addressed is the technical literacy for our clinical workforce. You know, we train physicians, we train clinicians who are focused in taking care of patients in a clinical manner. How do we really train our clinicians to work with AI, to work with digital tools? And I think that is something that we also need to address to really bring adoption on a wider scale.
- [18:00]
A
Yeah. Such important considerations from you both. Thank you for giving our listeners some practical takeaways in the trust and the adoption realm to really think on here. And I know we're winding down on our time together, so I wanted to be sure before we close that I kind of got the future projections from you both on what the next few years are really going to look like and how you're seeing cardiology evolving in the age of AI. So if you could each kind of share how you're seeing this landscape change in the next few years and one step that you'd advise leaders to take now as they're navigating AI and cardiology.
- [18:39]
B
I'll go first. So I think care is a continuum. And I don't say I can clearly say what's going to happen in the next two to three years, but I can say over the ensuing years, within the next five to six, for sure, as Sanjay alluded to very clearly, is that that care will evolve from becoming what is transactional to something that's more continuous. And the only way to actually be able to implement that is to really have the appropriate platforms that are integrated into the workflow. And some of these are remote monitoring platforms. These are remote monitoring centers that will embrace the life cycle of disease. Whether you're an outpatient, inpatient, you're in the emergency, or you're in your. Your office, it'll be able to continuously get data that can allow individualized care. Some of that care is going to move upstream. So that means what is, you know, chronic disease today will hopefully be nipped in the bud in the wellness stage while we move from. So when I look at care, if I may digress a little bit, I look at it at the apex of the pyramid is complex care. The big large midstream is chronic disease care and the large mainstream care is wellness care. I think they're all interrelated to each other and I think much of our care will be focused on the wellness and the midstream chronic disease care. And much of that care will eventually also be handed over to the patients armed with their information so that they can be self management approaches. And a lot of this will also be outcome focused reimbursement in the future. And I think the payers will reward institutions that have better outcomes and reduce readmission. So I think all those things are gonna eventually morph the way we actually look after our patients.
- [20:34]
C
Yeah. And I think in the short term we may also see shift of that care from hospitals to out of hospital setting. Right now there's a lot of emphasis at hospital, at home providing some of that wellness care and empowering patients. So I see patients getting more empowered. I see shift of care from traditional brick and mortar hospitals to actually shift of that care in the community, whether it's at home or community health centers. And I think second, at least in the short term, I see AI minimizing a lot of redundant administrative tasks. I think, I think the AI is already here that can do some of those things. How do we really balance the cost of those applications to the time value that they provide? I think it's something that needs to be worked out in the long term or maybe in the near long term. Like you said, the care will become more personalized. I see us leveraging data for that personalized and predictive care. So that pyramid that you described where the wellness then becomes a bigger chunk of that pyramid and we reduce the sick care and the complex care at the very top to very little. So I agree that we're going to see that shift of emphasis on preventing and personalized care.
- [21:51]
A
Yeah, yeah. So many through lines in both of your responses about empowering patients more the care shifting from brick and mortar to the community to patients homes opportunities with AI reducing admin burden and really, really just this larger shift toward wellness care versus reactive and sick care. This has been such a great conversation. Dr. Gandhi and Dr. Singh. I want to thank you for your time. Any other final brief thoughts from from you both before we close Today, if.
- [22:20]
B
I can just take on from what Sanjay just mentioned about AI reducing administrative jobs, I think it's going to affect all of our jobs. I think all of us clinicians are going to be practicing medicine in a different light, in a different vein, in a different way. I think what MDs do now will probably be done by apps with algorithms. What apps do now will probably be done by nurses with algorithms. What nurses do now will probably be done by healthcare navigators with algorithms. I think there is going to be an evolution of our job descriptions through AI and sensor based approaches and a lot of it is also going to be patients integrated into their own care. So. So I think we need to watch this space very closely because a lot is going to morph over the coming years.
- [23:09]
C
Yeah, I would maybe say it a little bit differently, Jag. I don't think we'll lose jobs right away. I think it will change how we deliver care. To your point, it's not so much about as a cardiologist worrying about if AI will open up any blocked arteries in the middle of night, but it is so much about how can I use AI to really make that life simple for me and maybe in the place first, first place prevent that patients from coming to the cath lab. I think AI is already here. It's making a difference. But I think we want to be careful that it's not so much just a hype, but I think we really take things that we can implement in practice and I think it needs to be understood and trusted. So I would emphasize that truly we need to embrace AI, but we also want to make sure that It's a responsible AI that can make a difference for our patients.
- [23:57]
A
Dr. Gandhi, Dr. Singh, thank you both again. This has been a fantastic conversation and I appreciate you sharing your expertise with our audience of leaders and I think there's a lot for them to think about and reflect on in listening to this session. So thank you both again and thank you Philips for sponsoring the session today.
- [24:17]
C
Thank you. And I think thank you so much for a really interesting discussion.
- [24:22]
B
Pleasure is entirely mine. Thank you all. Really appreciate it being here.
- [24:26]
A
Yeah. Thank you both again and listeners. Be sure to tune into more podcasts from Becker's Healthcare by visiting our podcast page@beckershospitalreview.com.