Today: Nonclinical uses for AI
Episode 5520th March 2024 • This Week Health: Newsroom • This Week Health
00:00:00 00:09:09

Transcripts

Today in health, it, we take a look at some nonclinical uses for artificial intelligence. My name is bill Russell. I'm a former CIO for a 16 hospital system and creator of this week health. Instead of channel said advanced, dedicated to transform healthcare. One connection at a time. We want to thank our show sponsors who are investigating, developing the next generation of health leaders. Notable service now, enterprise health parlance. Certified health and Panda health.

Check them out at this week. health.com/today.

Hey, we want to thank our partners and everyone who participated in captain's campaign for cure. So many pictures, so many great. Things happened as a result of it. We were able to raise about $40,000. Over the last couple of months and we really appreciate you

everyone who participated, if you want to be a part of it. Go to our website this week health.com.

Click on the blue. Logo at the top of the page. One last thing, share this podcast with a friend or colleague, you said as a foundation for daily or weekly discussions on topics that are relevant to you. And the industry. Mentor someone that's going to be my big dry for this year. You're going to hear me say the word mentor in every show that we do.

Whoever you're mentoring, they can subscribe wherever you listen to podcasts. All right. Again, this article is up on our news site this week. health.com/news. And it's eight, nine clinical AI applications on which physicians are especially clean, are keen. According to the AMA. All right.

The American medical association lays out eight. In demand, AI use cases for which the organization says it has heard physicians express. Particular enthusiasm.

All right. So here are the eight nonclinical uses for AI. Number one access to care. So they talk about optimizing scheduling. And aligning things with the patient needs and the physician to improve the experience. Absolutely support prior authorization process, including completion of followup of prior authorization documentation. These are two areas under access to care.

Number two, administration and revenue cycle. This is an area that has been that has had AI for quite some time. They have three sub points here, identify appropriate billing and service codes based on medical notes. And we're seeing coding come up more and more. And I think we're seeing that in the EHR. That they're integrated, integrating it into coding. And I think we're going to see that continue. Predict likelihood of and identify opportunities to reduce claims, denials and supporting accurate coding in the context of risk, adjustment, and value based care payment programs.

There you go. Number three operations predict hospital staffing volumes and requisite staffing needs. Track inventory. And utilization patterns to forecast medical supply orders. And number three, monitor equipment availability and predict equipment failures. And, you'd have to show me the ROI on those.

I believe that those models do exist. So you'd have to show you the ROI on those for me to invest in them. Now, if it was part of a platform and I could just turn on those, that kind of functionality, I absolutely looked at it. A regulatory compliance and reporting, automate the tracking reporting and regulatory compliance measures.

Reduce administrative burden. Analyzed documentation and processes to ensure adherence, to evolving healthcare laws and policies. Absolutely. I think that's one of the things. Especially generally I can consume large volumes of information. I summarize it. And then you can ask questions against that volume of data.

So very interesting. Hey, is this in compliance and you could feed it. Obviously from some of your source systems. Number five patient experience and satisfaction analysis, analyze patient feedback and surveys to identify areas of improvement in patient experience. And predict patient satisfaction trends and identify drivers. Of patient trust again, if these are all in a platform or sort of guide how to absolutely tie into them either EHR or others I know that press Ganey. And others are working on these kinds of models, which are going to be a wealth of information. Quality improvement and management. Automatically track identified. Quality outcomes and generate reports, identify gaps.

In quality. Number seven, education monitor clinical interaction and model patient, and provide feedback to physician and trainee. That's a interesting use case based on the review of physician and trainees experience and skillsets. Identify possible learning needs and or recommendations. Number. Some 0.3 under seven, provide automated haptic feedback during robotic training. Interesting. And then number eight, research. And we're seeing seeing AI pop up in research all over the place, predict the structures of proteins from amino acid sequences, optimized research, subject outreach, and enrollment in clinical trials. Analyze electronic health records at scale to identify potential human research subjects.

So interesting. Ah, they closed the article with this elsewhere in the report, the authors. Reiterate that AMS interviews and physician surveys, underscore physicians interest. And ensuring that adopted AI tools have strong data privacy protections are safe and effective. Integrate well with existing technology solutions. And protect physicians from liability made by algorithmic error. Again, very interesting.

Dave Pearson wrote this summary in AI, in healthcare innovation. To transform. As I'm thinking about AI, I want to make sure we don't end up with too many point solutions. I want to make sure that if my partners are currently working. With AI. I understand where they're thinking about integrating it into our existing systems. How they plan on providing transparency. And how they are planning on adoption of those systems. Especially if we already have the systems and all of a sudden these things start to pop up. Do we have the ability to turn things on and off?

What about data privacy? Can we make sure that our systems aren't training a model or potentially are training a model? If we wanted to do that? What kind of controls do we have in those things? If I'm trying to avoid point systems, I'm looking for platforms wherever possible. And I know that's a buzzword these days. When I think about platforms, I think about something that can be implemented and then do many things like an AI platform. I could be applied to the administrative area can be applied to operations, can be applied to a quality. Education and other things. Rather than a specific one for education and a specific one for quality and a specific one for regulatory.

If you allow the organization to do this, they will come back to you with a shopping list of 20 different tools that do certain things, especially if they. Go to conferences. Like I just went to a, there's going to be a lot of point solutions. Try to step back and see if you can have solutions. Where the AI can be applied to many of your situations within the health system. That's how I'd be thinking about this right now, but I definitely get in front of it with our existing partners and make sure I understood where they're, where they are taking it.

How they're. Thinking about integrating it into the existing platforms. And And what that is going to do for me. And. By all means I'm not by all means. I need to implore you to think about the pricing. One of the things that's happening right now is all these organizations see this as an opportunity to integrate AI into the system and then uptick the price by X, Y, and Z. And if you do that over 10, 15, 20 systems. You're going to run out of budget money. Really important to consider and. From my perspective, I would be looking into what it takes to train these models.

There's a bunch of open. Models coming out. You have you have the Metta has their open model. Ex Elon Musk just announced they're open sourcing their AI model as well. I don't know where these things are at, to be honest with you. I've heard some decent things about the AI models and the open models and those kinds of things. But I would be looking at what it takes to train these things up or potentially is there a startup partner that I can partner with that could build some things that I'm not paying these significant licensing fees moving forward? Again, I'd be experimenting with it.

I'm not saying I'd build it out and it would compete with. With Chatsy PT and other things. I'm just saying I would be experimenting with my partners that I could. That could potentially help drive other costs. All right. That's all for today. Don't forget. Show this podcast with a friend or colleague. We want to thank our channel sponsors who are investigating our mission to develop the next generation of health leaders.

Notable service. Now interprise health parlance, certified health and Panda 📍 health. Check them out this week. health.com/today. Thanks for listening. That's all for now.

Chapters

Video

More from YouTube