Name: Brainstorm AI 2024: The Role of AI In Human-Powered Health
Uploaded: 2024-12-10
Duration: 17 min 19 s
Description: Patricia Brennan, Vice President, Science Technology, Chan Zuckerberg Initiative, Tilak Mandadi, Executive Vice President, Ventures and Chief Digital, Data, Analytics, and Technology Officer, CVS Health, Shiv Rao, CEO and Founder, Abridge Moderator: Verne Kopytoff, Senior Editor, Tech, Fortune, Co-chair, Fortune Brainstorm AI

Patricia Brennan, Vice President, Science Technology, Chan Zuckerberg Initiative, Tilak Mandadi, Executive Vice President, Ventures and Chief Digital, Data, Analytics, and Technology Officer, CVS Health, Shiv Rao, CEO and Founder, Abridge Moderator: Verne Kopytoff, Senior Editor, Tech, Fortune, Co-chair, Fortune Brainstorm AI

Brainstorm AI 2024: The Role of AI In Human-Powered Health 

Conferences

Patricia Brennan, Chan Zuckerberg Initiative, Tilak Mandadi, CVS Health, Shiv Rao, Abridge, Verne Kopytoff, Senior Editor, Fortune

Hello, everyone. Thank you to our panelists for, for coming and talking to us today about a topic. It's very important that our lives depend on it. Uh, first off, can we perhaps start, start with you to? Um, you know, CBS is a, is a huge company doing a lot of things in A I, what is your favorite? What's your, who's your favorite child? Favorite use case? Are you sure, go ahead. Alright. Um, look, I think the, the um, anything we are doing with A I that makes our care and clinical professionals job easy is my favorite. So we have care managers that call people for next best actions and so on and so forth. We implemented, uh A I based uh, case prep that saves them 4040 percent of their time is now saved so they could spend more time with the, uh, with the patients so that those are the kind of things I like most and then Patricia to you. Uh, I, I can guess what your favorite thing is today. Yes, she has a bit of news. Um, today, I'm very excited that, uh, at CD I, we're announcing and launching um a platform to make uh kind of a one stop shop environment for biologists and ML developers to access models um of biology across uh cell data sic data, as well as imaging data. And these models are really designed to help uh biological research go faster and save time. So we can get uh to understanding disease uh quicker and, and get the cures um in a greater pace of time. Uh Sure. Could you tell us what, what a bridge does? Yeah, absolutely. So we're really focused on unburden, unburdening clinicians from clerical work so that we can bring the most important people in health care professionals and their patients closer together. So when you think about it, I'm a practicing cardiologist. When I see patients after every single encounter, I'm always mindful that I'm really serving multiple stakeholders. I have to write a clinical note for example, to give other folks on the care team. My colleagues an idea of why I fought through this differential diagnosis, why I prescribed a certain diagnostic or therapeutic. I also have to think about revenue cycle because in this country, we're not compensated for the care that we delivered. We're compensated for the care that we documented that we delivered. So these, these documents have real implications on all things related to what keeps the keeps the lights on in the hospital. And then there's the third constituent, the third stakeholder who's the most important, there's the patient and the family member. And if there's a term like transcatheter aortic boal plasty in my note, then I'm understandably gonna get epic inbox messages or phone calls asking what was that term? You never said it. I googled. It sounds scary. And so what we're focused on doing is taking from one conversation from one conversation, all the information that we need to do to do to do a kind of style transfer where we can serve all of those masters such that the clinician can just focus on the patient in front of them. And so like your, your company is a partner of, of a bridge. Uh uh how, how has it turned out? And I mean, have you encountered any wrinkles with the, with service? No, we, we work with uh Shiv and the team uh for the ambience uh uh in our uh in our clinics. It's working really well. Um It allows the doctor to pay more attention to the patient as opposed to turn around and start typing stuff, which is quite annoying. Um There, there's also um uh a I that we use in our call centers that has the same similar impact. So when you call somebody, when you call the insurance company, what have you, the most annoying thing that can happen is let me put you on hold while I pull your stuff together and 10 minutes will pass. Uh That's just that, that should not happen. So the way we are using A I in that context is Dutch stuff would come together in real time, so the agent continues to speak. So really good, good, good, positive impacts. And Patricia, I mean, health care is a very regulated field when it comes to data and privacy. I mean, uh how much harder does it make to innovate in a field in the field and, and make progress because of it? So I, you know, with uh the research that, that we do and the uh products that we're focused on, uh it's really at the sort of very early stage in the basic research. So um one of the things that uh we've really been able to do over the last several years is build up one of the largest um integrated uh data sets of single cell data, which now represents over 100 million cells. Um And those data, um we've been able to, you know, make sure that you can see, you know, where this data has come from with the studies, they link back to the original papers and whatnot. So it's really about making sure there's transparency uh to the data that underlines the models. And I hear that uh we have a new admin administration coming in and they have some ideas about A I and, and health care. What uh what have you, what are you expecting? Tla um Look, I my view is, is pretty straightforward. I think there is an arc of how A I is going to transform health care. It's an inevitable arc administrations may uh change the timing a bit, may, may go here and there some side branches. But the overall arc in my opinion is going to be in the direction of A I is going to augment clinical health care, never automate. Um A I is going to take a lot of administrative work off of the plate. A I is going to make human errors less and minimized. I think all those and, and that is going to help clinicians to get the information they need in real time. Uh That's very difficult to get otherwise. So I, I think that's gonna continue, that's going to be fine. Administrations are not. And, and so do you think that, uh you know, the Trump administration will be any better or, or worse than, than the previous one for start up? Well, I think the administration is absolutely gonna promote responsible innovation. You know, you think about it with health care in this country right now, two out of five doctors don't want to be doctors in the next 2 to 3 years, 27% of nurses don't want to be nurses in the next 12 months. So we have a public health emergency. We have to find a way to, to catch up to this emergency and technology has to be a part of the solution. And I think there's a recognition of that at every single level where I think it gets more complex is that not every use case. Um Can we, can we sprint right into? And I think that there are the lower stakes use cases like clerical work, for example, where we're simply trying to take off people's plates, the work that nobody wanted to do in the first place. And then I think there are higher stakes workflows or use cases around, for example, clinical decision support where we necessarily need to have some friction, some good friction in those workflows in order to make sure that we can deploy those technologies responsibly. So I think at this moment in time and what we've seen over the last two years is that those, those solutions, those products, those technologies that are on the safer end of that spectrum they've scaled. And I think that they'll continue to scale probably even more quickly over these coming years. And at the same time, I think we also as an industry have to start thinking about what those next innings look like even around the clinical decision support and how we can responsibly set the scaffolding for us to go after them. And, and um insurance claim denials have have reached the headlines following the murder of the United Health Care CEO um does a and I should say CBS owns uh an insurance company, does A I play a role in the insurance claim process and is it possible that A I could help uh with any the error rate of denials. First of all, it was a horrific thing that happened and sincere condolences to the family and United Health Care. Um Look what do people want from, from an insurance company. They want transparency if it, why did that decision happen? And an explanation that is in real time, not later. Um transparency to benefits. What is the coverage and transparency to cost? What what is my out of pocket cost and um an easier way to uh present a case, right? These are these are fundamentals that, that people want in a in a in a experience, healthcare experience and insurance companies uh work very hard to provide that A I is going to help immensely in all those things. Um speed up, you know, why can't we have real time preauthorizations at the time of care? The doctor is sitting there with the patient, they should be able to do a preauthorizations uh at at that time of care as an example, right? Um exposing the insurance contracts especially in the commercial space tend to be pretty complicated. So A I can help navigate that very quickly in real time to show what is covered, what's not covered and so on and so forth. So yes, I think it's an ongoing process. You know, we've been working on this for a for a while and A I is going to help even more to address those, those things, the transparency, the explainability, uh the ability to have uh a presentment of case and so on and so forth, make it easier. Uh In a few minutes, we'll go to audience questions. So please uh come up with a great one for us, please. Uh Patricia, um you know, everyone knows that health costs are, are soaring. Um Do you think that A I uh is a solution to at least some of that? So I think that um A I can help accelerate how basic research is done, for example, um or save time in uh across the entire sort of life cycle of, you know, sort of a binge to bed as it were or a binge to care application. Um And so I think there are opportunities to um you know, save the time of a researcher um by bringing more computational capabilities, more tools to bear on the work that they're doing and get the research out faster, get it to the clinician uh or the application health care um faster. Um There are some, you know, costs associated with scaling up and, and um you know, actually applying A I in, in the research process. But um I think ultimately it's about saving time. Uh Maybe we could go to questions here. Uh We have somebody uh with a paddle. Yeah, good sir. Go ahead. Um and state your name and affiliation and go ahead. Thanks by then. Um co founder and CEO of NANA, where we're a home based digital diagnostics platform. Uh sh if you touched upon this point that I thought was interesting, I think we can all agree in health care that A I will be great for workflow optimization, logistics, things like that. That's not as much as the next bar of diagnostics and therapeutics. What is the panel's opinion on some of the new aspects around that la that latter bar around validation liability. This concept of assurance labs versus FDA reduction of false positives and false negatives. Obviously, if we can corner and figure that part out, there's a real unlocking revolution in in health because curious on you guys' thoughts there, I can, I can kick things off uh maybe to, to to lock's point as well. The only currency that matters in health care is trust, the stakes can be so so high, especially in care delivery. And if trust is some combination of transparency, reliability and credibility, I think what a lot of us are starting are trying to get our arms around is like what does transparency but also reliability look like if if you're, if you're using a solution in the wild, how, how accurate is it? How reliable is it? Um And, and how do you measure that? So in our space, you know, we, we have technologies that under understand language and try to recognize the words that people are saying regardless of what language they're speaking in, even we have technologies that try to sort of a bridge the conversation, sort of distill the essence and structure that data as well. We have technologies that try to summarize as well for all the different audiences. And so for, for some of these challenges, it's pretty straightforward for us as, as, as um on our machine learning team, on our science team to sort of understand. OK. That's an F one score. We're gonna be measuring precision and recall or OK. For, for, for this, we're gonna be measuring medical term recall and um word error rate, for example, for speech recognition. But for some of these technologies, it's a little bit more challenging, for example, on summarization, how do we measure ourselves? How do we benchmark ourselves? And so 11 pursuit that we've invested a lot of time in cycles into is just being able to publish benchmarks around where we are today, being able to also give people a sense of how we assess quality ourselves within the company. And also trying to like expose these playbooks out to the broader industry so that we can all learn from each other. When I think about these assurance labs, I think to some extent, I see them sort of pulling inspiration from, from what a lot of different parties are publishing. And will we come up with one standardized sort of checklist on how to think through any given technology in any given use? Case, I think it's going to be a challenging task. And I think one of the reasons why is that the ground is moving underneath us so quickly, like these technologies are evolving so quickly that it's, it's hard to sort of take a snapshot of, of any given solution, much less problem and sort of call it done. You know, we take a lot of pride that, you know, bridge on Mondays is, should be great, but it should be even better on Tuesday, Wednesday, Thursday, Friday. We also take a lot of pride in sort of moving up the stack and expanding the scope of what we're building into other use cases. And so how do we, you know, create these standardized benchmarks for all these different technologies and all these different workflows. It's, it's tough, but I think we need some kind of like high level rubric rubric probably across the industry to just sort of guide us. My sense is that it's gonna come down to proxies for trust anyone else with a question. Go ahead. Do we have a paddle up here? Uh Lorraine Frazier at Columbia University and I happen to be the Dean of Nursing and I agree and disagree with you. I think the only currency is off of the dollar and trust hospitals, insurance companies. Let's be honest. How do you bring ethics into developing these models of A I? I would like to understand that you Patricia. Sure, I'll start with that, I mean, I think that uh ethics uh in all aspects of trust, you mentioned privacy already having transparency into uh the data actually. Um considering even before the the research uh into the model starts, what are some of the scenarios and use cases that are really trying to be enabled uh by them and bringing those to bear throughout the process. Uh benchmarks are a huge part of this. There's a I think real opportunity and gap in in the uh biomedical research field as it relates to uh sharing and uh bringing some consistency to how benchmarks um are uh made available across the research uh life cycle. Um So I think it, it really is about applying some of these core principles of uh privacy, data transparency. Um openness is another uh key aspect to this. So if you could probably take that offline, we, we're out of the time here. I wanna thank our panelists uh for a fabulous uh conversation. Thank you very.