Latest / Full Tech Ahead / The Role of AI in Healthcare

Transcript

The Role of AI in Healthcare · Full Tech Ahead

• 0:19
  SPEAKER_00: Hello and welcome again to Full Tech Ahead.
• 0:22
  SPEAKER_00: I'm your host, Amanda Razzani.
• 0:24
  SPEAKER_00: And today I'm excited to be here today with John Edwards of Sidious Tech.
• 0:29
  SPEAKER_00: How are you doing?
• 0:30
  SPEAKER_01: I'm doing great.
• 0:31
  SPEAKER_01: Thanks for having me today, Amanda.
• 0:33
  SPEAKER_00: Can you share a little bit about Sidious Tech and what services do you provide?
• 0:39
  SPEAKER_01: You know, Sidious Tech has been around for over 20 years, and all we've ever performed is healthcare consulting.
• 0:45
  SPEAKER_01: So we do engineering support as well as management consulting for the healthcare market, payer provider, health tech, and life sciences.
• 0:52
  SPEAKER_01: We have decided to focus because we think by focusing on healthcare, we can help our clients with some of their biggest problems and solve it with people that really understand healthcare rather than just people that understand technology.
• 1:06
  SPEAKER_01: And so while we do a lot of engineering work lately, a lot of data and AI work has been dominating what we're selling because that's what people are buying.
• 1:13
  SPEAKER_01: We field it with teams that know and understand the nuances of healthcare.
• 1:19
  SPEAKER_00: Okay.
• 1:20
  SPEAKER_00: Well, that goes into our topic today, which is AI in the healthcare business and the impact.
• 1:26
  SPEAKER_00: So, first off, healthcare AI adoption, it's accelerating rapidly.
• 1:32
  SPEAKER_00: What are some of the biggest changes you've seen in the market?
• 1:35
  SPEAKER_01: You know, we have seen in the market a desire to move away from just doing proof of concepts to show that you understand AI and how to build something to really wanting full-scale solutions.
• 1:46
  SPEAKER_01: And so that's really exciting because the elusive return on investment only really occurs when you adopt AI.
• 1:53
  SPEAKER_01: AI has made it easier to build things or to find partial results, but actually rethinking your workflow, your process, and using AI requires you to think differently than just experimenting.
• 2:08
  SPEAKER_01: And so it's great to see uh clients now ready to take that leap.
• 2:11
  SPEAKER_01: When it first came out, I think there was so much concern about can I trust it?
• 2:15
  SPEAKER_01: Will it hallucinate?
• 2:16
  SPEAKER_01: Will I make decisions that will be damaging to me?
• 2:20
  SPEAKER_01: And so there was a lot of experimentation that occurred.
• 2:23
  SPEAKER_01: Some companies are still stuck in the experimentation, but we're seeing more and more people really wanting to get to that scalable solution and uh find that return on investment from the work they're doing in AI.
• 2:35
  SPEAKER_00: Wonderful.
• 2:36
  SPEAKER_00: Well, AI has enormous promise in healthcare, but there is a lot of complexity to it.
• 2:41
  SPEAKER_00: So, what are some of the biggest barriers preventing healthcare organizations from fully operationalizing it?
• 2:49
  SPEAKER_01: I think that one of the challenges that makes sense when you think about it, but people sometimes overlook is that healthcare information comes in through time.
• 2:58
  SPEAKER_01: The tests that I had for my doctor a year ago may not still be relevant for the physical that I'm about to get and the tests that I need now.
• 3:07
  SPEAKER_01: And so there's a time sensitivity to some of the information that is part of the context that a doctor would use in making a decision.
• 3:16
  SPEAKER_01: AI has to be taught how to treat that information and which information is valuable over a long period of time, and which information has uh less value and has to be considered um differently.
• 3:30
  SPEAKER_01: And so all that context of how a decision is made by a physician or a nurse or a clinician has to be trained in AI.
• 3:39
  SPEAKER_01: And the challenging thing is those decisions are often being made in people's heads in their education, and there may not be, you know, documentation about how they used all this information.
• 3:52
  SPEAKER_01: And so training something for scalability requires sufficient information, a lot of information, so you can better get AI to make decisions that are reliable, like a human would make.
• 4:04
  SPEAKER_01: And so it is complex, it does require thinking about quality and how you are going to build your solution.
• 4:14
  SPEAKER_01: And it also requires you to think about where you're going to keep a person in the loop.
• 4:19
  SPEAKER_01: Because it's healthcare, even with all the quality that you can build, there has to be a way for supervision of the model.
• 4:27
  SPEAKER_01: And so coming up with the metrics that matter is an important part of healthcare AI, because there will always be some voice inside of your organization or criticizing you from the outside around you're using AI.
• 4:41
  SPEAKER_01: Can you trust it?
• 4:42
  SPEAKER_01: And so there can't be black boxes.
• 4:45
  SPEAKER_01: There has to be clarity about how the decision is made, and there has to be metrics about the performance of the model that gets built, not just the model itself in its use.
• 4:56
  SPEAKER_01: So I'd say measures that matter is the first, you know, obstacle that has to be overcome.
• 5:02
  SPEAKER_01: The second obstacle that has to be overcome is inertia.
• 5:05
  SPEAKER_01: People used to listen to a heart by putting their ear to a chest.
• 5:09
  SPEAKER_01: The stethoscope came out and it gives you a much better sound of a heart.
• 5:13
  SPEAKER_01: And you can't imagine going to a clinician today without them listening to your heart with that device.
• 5:18
  SPEAKER_01: But it took decades of that facility, that capability being available for clinicians to change the way they'd always done it.
• 5:27
  SPEAKER_01: They had a practice, they had a pattern, and even though there was technology available, there was inertia because the way that I do things, I know, I trust, I trust what I hear.
• 5:36
  SPEAKER_01: I don't know if I'm going to trust this other device or be willing to adapt my practice.
• 5:42
  SPEAKER_01: AI is like that because AI is going to change the way some people do their work if it's built correctly.
• 5:50
  SPEAKER_01: And so overcoming that inertia of unwillingness to change, doing it the way I've always done it, is a significant obstacle.
• 5:58
  SPEAKER_01: And it's all about adoption.
• 6:00
  SPEAKER_01: You know, in a typical IT project, adoption might be 15 to 20 percent of your budget about getting a technology adopted.
• 6:09
  SPEAKER_01: There are some that say you should be spending$2 for every one on adoption if you want AI to work.
• 6:17
  SPEAKER_01: That's significantly different spending and attention because AI won't bring value unless people trust it and adopt it.
• 6:28
  SPEAKER_01: So you have to have the metrics that matter, and then you have to overcome inertia of wanting to change to be able to get to the value of healthcare AI.
• 6:37
  SPEAKER_00: Absolutely.
• 6:38
  SPEAKER_00: That makes a lot of sense.
• 6:40
  SPEAKER_00: Well, you recently created a new, like a 2.0 version of a tool.
• 6:45
  SPEAKER_00: Can you share that tool and what problems were you trying to solve with it?
• 6:50
  SPEAKER_01: Yeah, we we built a tool called, we called it Neuron, K-N-E-W-R-O-N.
• 6:55
  SPEAKER_01: And what it is, is a tool that allows you to more easily build agentic AI.
• 7:00
  SPEAKER_01: It has the context of healthcare built into it, it has the data quality measures and capabilities to track and measure the model, but it also can integrate into existing workflows or use algorithms from other areas and be an orchestrator of the agentic AI capability that's desired.
• 7:21
  SPEAKER_01: We felt like healthcare is so complex that if you used a generic tool in AI, you'd have to build all this context every time to make sure that it was interpreting healthcare information correctly.
• 7:34
  SPEAKER_01: Some of our clients have us build that way within the case tools that they've chosen.
• 7:39
  SPEAKER_01: We build a tool, though, to allow us to go faster and to be able to expedite the use of AI and healthcare through Neuron.
• 7:48
  SPEAKER_01: And so we have uh clients that are using it today and experiencing a more rapid um deployment of their solution because of what's already pre-built and the capabilities.
• 8:00
  SPEAKER_01: And so that was really the goal was to offer healthcare an option around building within a healthcare-specific tool set.
• 8:07
  SPEAKER_01: Again, we we build outside of that tool as well because we're consultants.
• 8:11
  SPEAKER_01: We we follow the architecture principles that have been established at a given company.
• 8:16
  SPEAKER_01: But um, Neuron is an option that they can consider.
• 8:19
  SPEAKER_01: And uh some of our clients are deciding to go down that path for their AI builds.
• 8:24
  SPEAKER_00: Well, when we look at all the AI tools out there, it does seem a lot of companies are still struggling with uh implementing these AI tools uh across the company.
• 8:34
  SPEAKER_00: What are some of those roadblocks?
• 8:36
  SPEAKER_00: Uh, why do you think they exist?
• 8:37
  SPEAKER_00: And what advice do you have for integrating these tools more easily?
• 8:43
  SPEAKER_01: You know, Amanda, I think that having specificity in your goals is important because there are going to be different people with perspectives of what AI can do, should do, could do, whether the CFO, CEO, COO, or on down in the organization.
• 8:58
  SPEAKER_01: But if you aren't clear about why you're doing AI, then it'll be difficult to be able to get the inertia overcome within the organization to be able to make the change.
• 9:08
  SPEAKER_01: I think the organizations that are doing it best are those that have set up some type of governance.
• 9:14
  SPEAKER_01: Sounds burdensome and overhead, but that governance process usually drives forward a business case and clarity about why we're doing something.
• 9:25
  SPEAKER_01: So it moves it off of the data scientists' experiment land into a managed implementation with clear goals and objectives and buy-in from the areas that are going to be impacted by the change.
• 9:39
  SPEAKER_01: And so that's a big shift.
• 9:41
  SPEAKER_01: And I think organizations that have that type of governance process and get that buy-in have a better chance of seeing the implementation occur.
• 9:50
  SPEAKER_01: You know, we're working with uh multiple clients that have open lists of AI priorities.
• 9:58
  SPEAKER_01: And the challenge is everyone that submitted one of those ideas really, really wants it.
• 10:03
  SPEAKER_01: And often, a client I was at recently had, I think, 49 submissions by their senior leadership team, 49 different projects they wanted.
• 10:11
  SPEAKER_01: They thought they had budget to do two.
• 10:13
  SPEAKER_01: That means 47 ideas aren't going to be fulfilled that somebody was passionate about.
• 10:19
  SPEAKER_01: And so you have to have a process to be able to choose from a lot of great ideas that come in from creative people to get specific things that you can do.
• 10:30
  SPEAKER_01: And then there will be people upset because their idea wasn't processed, they're not getting the good stuff, you know, that AI capability and AI help that they desired.
• 10:39
  SPEAKER_01: But with discipline and with a clear process by which you're chasing those business outcomes, you know, AI can pay for itself within a year easily and get a return on investment because of the possibilities it brings.
• 10:52
  SPEAKER_01: And so often, if you can put that business case story together, you can show a path towards self-funding and find the money within your operations.
• 11:03
  SPEAKER_01: But that money may come from having different ways of doing work.
• 11:10
  SPEAKER_01: And if you're not willing to change what you're working on, you may not be able to justify the investment.
• 11:15
  SPEAKER_01: The biggest mistake I see people making is automating a bad process.
• 11:20
  SPEAKER_01: The processes we all use were created when AI wasn't around, when people had to do things, when when you relied on the human brain and the person to be able to make every little decision, and then they were supported by an app that you logged some of your results.
• 11:34
  SPEAKER_01: But most of the thinking and decisioning was in your head and in your work with others.
• 11:39
  SPEAKER_01: With AI, you you want to pull some of that thinking into the AI machine if you're gonna create agentic AI capabilities.
• 11:47
  SPEAKER_01: And so that causes the person that used to have all those ideas in their head to feel threatened.
• 11:52
  SPEAKER_01: Because what I do and the value I'm creating, they want to get it from me.
• 11:57
  SPEAKER_01: And so overcoming the governance, you know, of this is the right decision and we decide we're gonna do it is one part of the inertia.
• 12:06
  SPEAKER_01: But then at the ground level and the process that's being used, the knowledge workers need to still feel appreciated, still need to understand how their job is going to be more interesting and better and different.
• 12:19
  SPEAKER_01: Because if you don't have them involved in the process, it's gonna be very difficult to get them to adopt it later or to make the changes that are needed.
• 12:26
  SPEAKER_01: And so I think that um as we move into this scalable enterprise AI solutions, it's gonna be more about the people than about the technology.
• 12:37
  SPEAKER_01: More about changing how we do things than generating something with an AI chat window because the real value comes from automating some of these processes.
• 12:50
  SPEAKER_01: You know, I use AI every day now in my daily work.
• 12:56
  SPEAKER_01: It makes me more productive, doesn't it?
• 12:57
  SPEAKER_01: You as well, Amanda.
• 12:58
  SPEAKER_01: I mean, I I can't imagine not having these tools.
• 13:02
  SPEAKER_01: My children use AI, and you know, my 12-year-old is always telling me things.
• 13:06
  SPEAKER_01: I said, Where'd you find that?
• 13:06
  SPEAKER_01: She goes, He goes, Oh, ChatGPT told me.
• 13:08
  SPEAKER_01: He doesn't go to Google anymore to ask questions.
• 13:10
  SPEAKER_01: He goes to Gemini or ChatGPT, and he's 12.
• 13:14
  SPEAKER_01: You know, the next generation of workers that are coming up will have used AI in college and in high school and in their daily life, and adoption is going to be easier.
• 13:24
  SPEAKER_01: But all of us that came from before, that came from a different world, have to find a way to embrace the change, or you could be an obstacle towards the success of the company.
• 13:36
  SPEAKER_01: You know, AI doesn't replace people, AI unlocks the knowledge that people bring and allows you to do things you were always able to do, just do them faster.
• 13:45
  SPEAKER_01: And it gives you the possibility of doing other things too, which is exciting.
• 13:50
  SPEAKER_01: But it's scary, it's scary to change.
• 13:53
  SPEAKER_01: And I think that if people don't think about the human side, not just the engineering side, that they're gonna miss the opportunity to be an early adopter in AI because they won't overcome that inertia, they won't get the buy-in from the people that need to change, and they might just automate a bad process.
• 14:10
  SPEAKER_00: Right.
• 14:10
  SPEAKER_00: So it sounds like yeah, it sounds like the most successful people are gonna be those who are willing to be adaptable and flexible.
• 14:18
  SPEAKER_01: That that's it.
• 14:19
  SPEAKER_01: Yeah, and you know, it is part of it's a big change, and it can create momentum inside of the company, but there's a lot of blocks to it, and they're not technical.
• 14:31
  SPEAKER_01: The the technology is easier now.
• 14:33
  SPEAKER_01: The AI tools that are out there are amazing.
• 14:35
  SPEAKER_01: There's so many options of how to build things and how to create the scalable solution, but the solution has to have knowledge and context, knowledge of healthcare practices and context of how decisions are made.
• 14:48
  SPEAKER_01: And creating knowledge graphs and context graphs as part of the training of healthcare has to be governed by that overall quality process.
• 14:55
  SPEAKER_01: Then you can have reliable solutions, but you still have to convince people to use them.
• 14:59
  SPEAKER_01: You have to help people understand how this works better.
• 15:04
  SPEAKER_01: I remember when we trained, uh, this is going to show my age, people to use mouses rather than green screens with shortcut keys.
• 15:12
  SPEAKER_01: And I remember nurses that would say, but that's going to slow me down.
• 15:16
  SPEAKER_01: I know the shortcut keys.
• 15:17
  SPEAKER_01: I don't want to point and click.
• 15:19
  SPEAKER_01: Well, it probably did slow them down a little bit, but you didn't have to memorize a bunch of things, right?
• 15:24
  SPEAKER_01: And get lost in green screens.
• 15:26
  SPEAKER_01: Now AI has the possibility of changing that again.
• 15:29
  SPEAKER_01: We can speak to the application more readily rather than click and type.
• 15:33
  SPEAKER_01: We can ask questions and review content and provide workflow without hands on keyboards if we design the systems that way.
• 15:43
  SPEAKER_01: It creates the possibility of faster results.
• 15:48
  SPEAKER_01: And finding the places in your business process where those faster results will be valuable and valued is part of that art of prioritization and creating capability that alters the company and its ability to fulfill its mission.
• 16:03
  SPEAKER_01: And so healthcare is tricky, but AI is very possible within healthcare if built correctly.
• 16:09
  SPEAKER_01: And, you know, getting people to adopt it and understand it can solve real friction points in the workforce as well.
• 16:15
  SPEAKER_01: So many doctors complain about the pajama time, the work they're doing at home after they've left the office, starting to finish up their documentation.
• 16:23
  SPEAKER_01: And I remember uh in a doctor's visit I had, I had a resident that was there being trained, and he said, he goes, You're in a in uh AI, right?
• 16:32
  SPEAKER_01: In consulting.
• 16:33
  SPEAKER_01: I said, Yeah.
• 16:33
  SPEAKER_01: He goes, promise me that sometime in my career I will be able to quit typing in the EHR.
• 16:38
  SPEAKER_01: He was using Epic, but he didn't care which one.
• 16:40
  SPEAKER_01: He didn't want to type anymore.
• 16:41
  SPEAKER_01: He goes, It's such a waste of my time.
• 16:43
  SPEAKER_01: I said, It's just around the corner.
• 16:45
  SPEAKER_01: Interesting, a few months later, when I was back at that same doctor's office, the doctor asked me, Do you mind if my digital assistant attends this meeting too?
• 16:53
  SPEAKER_01: They had started ambient listening.
• 16:55
  SPEAKER_01: Ambient listening will turn into not just listening, eventually they'll turn into suggestions and workflow and creating a partnership with the doctor about changing how they practice medicine.
• 17:07
  SPEAKER_01: But it starts with being willing to be recorded.
• 17:10
  SPEAKER_01: I remember when we first talked about recording, like doctors won't want to be recorded.
• 17:13
  SPEAKER_01: They'll think it's just a place to get sued.
• 17:15
  SPEAKER_01: You know, they didn't they'll resist.
• 17:16
  SPEAKER_01: They won't want, no, if you can solve the friction of pajama time of extra work that I shouldn't be doing, I'm more than happy to be recorded.
• 17:25
  SPEAKER_00: Absolutely.
• 17:26
  SPEAKER_00: It seems like there are limitless possibilities for the future.
• 17:30
  SPEAKER_00: Well, if there was one key takeaway that you could leave our audience today with, what would that be?
• 17:35
  SPEAKER_01: If you're pursuing a healthcare AI project, make sure you have people on your team that understand healthcare, not just the AI side.
• 17:44
  SPEAKER_01: The AI side is interesting and complicated and gotten easier.
• 17:49
  SPEAKER_01: But if you don't have knowledge and context in your solution about how healthcare really works, you will be unlikely to have the same impact that you should have.
• 17:59
  SPEAKER_01: So as you're thinking about your team, don't leave out the BAs, don't leave out the change management people that help with thinking about adoption.
• 18:07
  SPEAKER_01: Because a perfect mousetrap that's never used won't catch any mice.
• 18:12
  SPEAKER_01: You need to be able to get the human side of it engaged and excited and helping deform the future of the business process.
• 18:23
  SPEAKER_01: Technology needs to support all that, and it needs to be open to that business input in a way that's stronger than it's ever been before because it's becoming personal.
• 18:34
  SPEAKER_01: It's about my job, my career, my decision that my name's with.
• 18:38
  SPEAKER_01: And the business has to feel comfortable or they are not going to use your solution.
• 18:43
  SPEAKER_01: So involve the business, involve BAs, involve change management is my strong advice if you're pursuing AI and healthcare, probably more important than any other industry around having that business voice, because the decisions we make affect us all.
• 18:57
  SPEAKER_01: Our health, our wellness, our happiness.
• 19:00
  SPEAKER_01: So thank you for your time, Amit.
• 19:03
  SPEAKER_01: And thanks for having me today.
• 19:04
  SPEAKER_01: This was great.
• 19:04
  SPEAKER_00: Yes.
• 19:05
  SPEAKER_00: Thank you so much for sharing all your information with our audience today.
• 19:08
  SPEAKER_00: It was a great discussion.
• 19:10
  SPEAKER_00: And thank you to our audience.
• 19:12
  SPEAKER_00: If you have any questions or comments, share those and I'll make sure to get back to you as soon as possible.
• 19:17
  SPEAKER_00: Have a wonderful day.