How Patients Use AI (Grace Vinton, Grace Cordovano)

[00:00:00] Dear listeners, welcome to Faces of Digital Health, a podcast about digital health and how healthcare systems around the world adopt technology with me, Tjasa Zajc.

[00:00:12] When talking about how generative AI revolutionized healthcare by making note-taking easier for clinicians, we rarely hear the impact that generative AI had on patients.

[00:00:27] Grace Cordovano is an expert healthcare navigating solutionist and an award-winning board-certified patient advocate.

[00:00:35] And Grace Vinton is an account director at the PR agency Amendola.

[00:00:41] She's a patient advocate and also the host of the Like a Girl Media's podcast, Hi-T with Grace.

[00:00:50] I spoke with both of them at the Digital Health and AI Summit in Boston.

[00:00:55] where we discussed the use of AI among patients, how it's improving the patient experience and how it can help patients better manage their care.

[00:01:08] In the short discussion that you are about to hear, I'm referring to Grace Cordovano as Grace 1 and Grace Vinton is Grace 2.

[00:01:20] Enjoy the show and if you haven't yet, check out our newsletter.

[00:01:24] You can find it at fodh.substack.com.

[00:01:28] That's fodh.substack.com.

[00:01:31] I'm going to link the edition about the insights from the Digital Health and AI Innovation Summit from Boston.

[00:01:40] And before we dive in, if you will enjoy the show, please leave a rating or a review wherever you listen to your podcast,

[00:01:48] because this really helps new listeners find the show as well.

[00:01:52] Thank you.

[00:01:53] Now let's dive in.

[00:01:57] We are in Boston at the Digital Health and AI Innovation Summit.

[00:02:13] And we talked a lot today about the innovation and the use of AI in healthcare.

[00:02:20] And Grace, too, I want to start with you.

[00:02:24] What were some of the things that stood out to you in the discussions that we've heard today?

[00:02:31] A lot of the discussions today were around typical use cases for AI, one, and the future of AI, two.

[00:02:39] And then the day ended with our great patient panel talking about how patients are viewing all of those things.

[00:02:46] And I think what people are excited about is the potential that AI can have.

[00:02:51] But I do think that some of the conversations that weren't happening really today were around cybersecurity and concerns that I think a lot of patients are having and caregivers and care partners are having.

[00:03:02] I really like the fact that a lot of emphasis was put on real-world use cases and challenges that we see.

[00:03:09] So I loved the fact that we are still often talking about the potential of the technology.

[00:03:15] And the key question is, it's not just about technology, it's about the outcomes of the patient.

[00:03:21] So if there's no impact, tech is just tech and many other things.

[00:03:27] But Grace, as a patient advocate, you mentioned that there are several use cases in the way that patients use AI.

[00:03:37] Can you name three?

[00:03:40] Sure.

[00:03:40] So one of the prominent ones is really reducing patient administrative burden, helping us work smarter, work faster.

[00:03:47] Because at the end of the day, many of us are working.

[00:03:50] We have other responsibilities in our life.

[00:03:52] We have to be a mom.

[00:03:53] We may be caring for aging parents.

[00:03:54] We may have a career.

[00:03:56] We may be a single mom or a single parent with children and aging parents and maybe a new pet.

[00:04:02] And life throws you this wrench and you have to learn about your diagnosis, digest all the information, coordinate your care.

[00:04:09] And it's all paper-based.

[00:04:11] So it's very manual.

[00:04:13] It's primarily manual labor.

[00:04:15] You almost need a wheelbarrow to carry all the paper.

[00:04:17] So I'd say patient administrative burden and using chat PPT and notebook LM.

[00:04:23] There's a number of advocates that are really deep in the woods on how do you work faster, taking your medical record and digesting it and then maybe translating it into a native language.

[00:04:33] I think other ways is really to help us prepare and operationalize.

[00:04:37] So many patients, when they're diagnosed, they want to prepare for that first encounter.

[00:04:41] And one of the most common questions I get is, what should I be asking?

[00:04:44] I don't even know what to ask.

[00:04:46] And you can easily prompt and come up with a almost business strategy for that initial new patient appointment or maybe a follow-up appointment to concerning results or questionable results.

[00:04:57] And then I think from a navigation standpoint, your common acute encounters are more of that patient experience or that journey, more predictable.

[00:05:10] We heard some really powerful stories on stage today about what about when it's an odyssey and you don't know what that final destination is or your final destination is, like Kate said, I have a terminal diagnosis and I need access to information and we need to move faster.

[00:05:31] Taking this whole concept of what a patient is and recognizing that there's different categories, the proactive wellness seeker, someone who has an acute encounter, and then you cross over into multiple comorbidities, life-altering, life-limiting conditions, terminal illness, and then even end-of-life care and active death.

[00:05:50] Patients and families still need information.

[00:05:53] And I think the audience was very receptive to the way that Kate ended the conversation saying, not all of us have the luxury of time.

[00:06:03] A lot of the conversation about what's coming in the future, we have to recognize that some people's future may be on the scale of days, weeks, and months.

[00:06:10] And what would be your advice on how exactly to use AI?

[00:06:15] I know that you mentioned that it can be used to prepare for medical visits, but promising the AI engines is a little bit of a, not an art, but definitely a skill.

[00:06:27] So you need to try several times.

[00:06:29] So if you would try to give advice on step one, step two, step three, what would it be?

[00:06:34] And also keeping in mind that you don't necessarily want to share your concrete medical results is chat GBT.

[00:06:41] That's such a great question because something that my colleagues and I are noticing, we were just talking about this before our panel.

[00:06:47] There's no library of responsible prompts.

[00:06:50] Where do you start with prompting, whether it's trying to analyze maybe new proposed rulemaking that has come out that requires public comment?

[00:06:58] Maybe it's how do you prepare for an appointment and how do you ask these questions?

[00:07:01] A number of us and this hashtag patients use AI are starting to share examples of what our prompts look like so people can see.

[00:07:09] And I'm still learning and I'm fascinated with prompting.

[00:07:12] And I liken it, this refinement almost to when I write, I use a thesaurus because some of the words I see that I'm using redundantly.

[00:07:20] I can't get enough of looking at prompts, not just in healthcare, but just the wording, the sentiment, the voice, taking on a role of whether it's a character or in a prompt that I wrote today,

[00:07:32] you're a expert attorney in the state of New Jersey specializing in Medicare disability benefits.

[00:07:40] So really honing and tweaking, build that character.

[00:07:45] What is the information that you have, whether it's an appointment at your oncologist or cardiologist?

[00:07:51] What are you looking for?

[00:07:53] And tell me what that output is.

[00:07:55] For me, today's output was an appeal letter.

[00:07:57] For someone else, it may be, give me a list of 15 questions.

[00:08:01] And then there's follow-up prompts and refinements where it could be, that wasn't good enough.

[00:08:05] Drill down harder or you can do better than that.

[00:08:08] You're more of a creative marketer.

[00:08:10] Refine it this way.

[00:08:11] So keeping adjusting those lenses.

[00:08:13] I was recently at the eye doctor and I don't know if you've been to the eye doctor, but they always change the lenses and which one's clearer or blurrier.

[00:08:22] Lens one or lens two.

[00:08:23] And I'm almost likening that as an analogy for prompting because there's so many versions of the prompts that you have to revise.

[00:08:30] But it takes someone to almost hold your hands to know what you're doing.

[00:08:34] And I know Hugo's doing that too.

[00:08:36] Absolutely.

[00:08:36] I would say if patients are interested in starting to use AI to hack their own health, the best way to start, like she said, is with questions.

[00:08:44] Asking Chachi BT or whatever you'd like.

[00:08:46] I'm about to go to the doctor and this is what I think I might have or these are some of my symptoms.

[00:08:53] Can you give me a list of questions that I should be asking to be able to take advantage of the time that I have with this specialist?

[00:09:00] It's taking so long to get even access to care, to even get in front of specialists.

[00:09:05] And when you have an issue, you need to really take advantage of that time.

[00:09:09] So I always tell folks when I'm doing patient advocacy, let's start by just asking them to help us ask the right questions.

[00:09:16] Right.

[00:09:17] And so that's a great way to start because with hallucinations and stuff, you may not be able to trust a lot of the answers they give you.

[00:09:24] But you can trust the types of questions they could give you to ask and get the answers from the medical professionals.

[00:09:30] Right.

[00:09:30] So I think Chachi BT is just one example.

[00:09:34] And I think another panelist said today, you know, there are other parts, of course, types of AI that patients are excited about and are using to different types of AI that are checking your records to make sure that there aren't any errors or duplications, which is such a huge issue.

[00:09:49] And patients care about it.

[00:09:50] And I think that's something that the industry needs to know and take away from what the conversation was today from patients, caregivers and care partners about AI is that we care and we do have a say.

[00:10:02] And we are using AI to hack our health and help the medical professionals get to the final diagnosis and care plan.

[00:10:10] Which reminded me of the first session that we saw today, which was all about how can we put the medical discharge letters or just the medical notes into understandable language?

[00:10:26] Because as patients, we often don't understand them.

[00:10:28] Amir Al-Haw finished the conference by saying that companies shouldn't talk to patients because it's a nice thing to do and because it's becoming a thing that should be on the checklist of the things that you do, but because it's the right thing to do.

[00:10:45] So what would you add to that in terms of patient involvement, patient engagement?

[00:10:51] Where can companies maybe find out more about how patients think?

[00:10:57] You mentioned a specific hashtag, which I forgot at the moment.

[00:11:01] So maybe you can repeat that.

[00:11:02] Hashtag patients use AI.

[00:11:05] Okay.

[00:11:06] Where would you say that patients and the solution makers can learn more?

[00:11:11] I think starting with some references would be the patients included charter.

[00:11:16] This has been a foundational document.

[00:11:19] It's online.

[00:11:20] Check out patients included.

[00:11:21] And it gives you the reasons why.

[00:11:23] And the reasons why is it's solid business strategy.

[00:11:26] Whether you're putting on an event, whether you're in pharma, whether you're designing trials, whether you're a startup, having a diverse group of patient voices that can partner with you throughout your journey.

[00:11:39] Not just for that one survey and say that you did it.

[00:11:42] But longitudinally to give that feedback, whether it's good or bad, is important.

[00:11:47] And how you do that, you have to have a budget for it.

[00:11:49] And we talked about, I love that you emphasized that it was so important to compensate patients.

[00:11:54] I want to take that a step further.

[00:11:56] And the risk of not compensating patients.

[00:11:59] If you don't compensate patients for their time, only the ones that have the privilege and the luxury of being able to afford to come, pay for the hotel and pay for the travel and pay for the sitter and have paid time off.

[00:12:11] Are going through a time.

[00:12:12] So you're already baking in bias and failure into your business strategy.

[00:12:17] So the way to do it is to look at it, set a line item aside for your new year, create a budget.

[00:12:24] National Health Council has a great resource, the Fair Market Evaluator.

[00:12:29] And you can take a pass through.

[00:12:30] There's already a little calculator and it'll give you a range of what is the Fair Market for a patient participant, whatever project level and whatever commitment level.

[00:12:38] And that's a great starting point.

[00:12:40] Talk to your CFO or talk to your town and people find a way and it will bring you maximum ROI and also save you on your marketing because something that works well.

[00:12:51] Patients are hungry for knowledge and we heard people, they're going to want to celebrate it and they're going to sing your praises to the communities that they work with and online.

[00:12:59] And there's no better marketing.

[00:13:02] Patients are one of the most underused resources in health care is also one of the thoughts that we heard today.

[00:13:09] Grace, too.

[00:13:11] I was going to say also there's organizations that can help you source patients that are diverse and in their own subsets.

[00:13:17] Savvy Coop is one that will help you work through what types of patients you should have for your clinical trial or for your digital health innovation.

[00:13:25] And they will help connect you with patients and caregivers and care partners and surviving caregivers that can help give you that insight.

[00:13:32] And they'll give you all of the information you need so you can have just patients in a box, you know, that situation we working with a Savvy Coop.

[00:13:39] And if you want to do it on your own, just make sure you take a look at that patients included charter.

[00:13:44] Make sure it's a diverse subset of patients and make sure that you include a feedback loop.

[00:13:50] So if something a patient or caregiver or care partner said changed something, tell them.

[00:13:55] And that's funny because the industry doesn't do that right now.

[00:13:57] And they'll ask patients, caregivers, what do you think?

[00:14:00] And then they won't tell them if anything that they said actually changed what they decided to do.

[00:14:04] Keep those folks in a feedback loop of understanding how their insights impacted what you are developing.

[00:14:11] We discussed advice.

[00:14:13] What about warnings?

[00:14:15] What kind of warnings would you give to patients around the use of AI?

[00:14:21] And I can give you a personal example.

[00:14:24] So last year I had a lot of chest pain, which is super scary during the summer.

[00:14:29] And before I got to the medical appointment with a gastroenterologist, I obviously checked a lot of information online.

[00:14:36] I tried to figure out what it was.

[00:14:38] My GP had an idea.

[00:14:40] I suffered from confirmation bias when I was searching online.

[00:14:43] And then I came to the gastroenterologist and I said, I have this.

[00:14:46] And he said, says who?

[00:14:49] And I was like, oh my God, I totally over-relied on the internet and on changing the team.

[00:14:55] So what should we be mindful of apart from hallucinations?

[00:15:01] Which means if anybody doesn't know when AI makes up things in an answer and the facts are not really facts.

[00:15:10] Being very honest and saying, this is not a silver bullet.

[00:15:14] This is a new tool.

[00:15:16] It's an up-and-coming tool.

[00:15:17] And every tool has great uses and then uses that you can use a tool for.

[00:15:24] So that precursor and disclaimer has to be there.

[00:15:27] And I think that there is a general awareness.

[00:15:30] But I think each health system and cancer center and physician's office could do a much better job of addressing it on their website.

[00:15:37] Maybe there's something in your patient portal that gives just a quick blurb and best practices.

[00:15:43] I want to see more of that because I sit in a lot of exam rooms, waiting rooms, emergency rooms.

[00:15:48] There is nothing on the walls or on the TV or on the coffee table about or a QR code that you could scan or a commercial about AI and how you could use it.

[00:15:59] So we're really far behind in life.

[00:16:01] A lot of opportunity.

[00:16:02] You basically opened up this whole new question of digital literacy and who should be responsible for it, who should prepare the materials for it.

[00:16:09] If you're in healthcare, you're responsible for digital literacy.

[00:16:13] No matter who you are, where you are, what you are, it's not somebody else's responsibility.

[00:16:18] It's yours.

[00:16:18] I think that's very critical.

[00:16:20] I think another thing that patients should really be concerned about is security of their own data, right?

[00:16:25] Right now, patients don't really care.

[00:16:27] If you're using AI to better my outcomes, like, go do that, right?

[00:16:30] But folks like individuals at the Light Collective are really taking social security concerns at heart and are really advocating for patient rights as it relates to the security of their own data.

[00:16:42] So I recommend you going online to the Light Collective.

[00:16:45] Go online, check out Savvy Coop.

[00:16:47] I was just doing those because those are great organizations that are really standing up for patients and caregivers and care partners.

[00:16:52] And it's because we don't really know about it.

[00:16:55] There's not a lot of education about what patients should care about as it relates to AI and their own care.

[00:17:01] Be careful about what information you trust, whether it's through ChatGPT or any other website.

[00:17:06] Be careful about what information you give away.

[00:17:08] Make sure you really read what people are making you sign at the doctor's office.

[00:17:13] I think those are all really critical things patients need to keep in mind.

[00:17:16] Grace one, grace two.

[00:17:17] If you have any last thoughts, let me know.

[00:17:20] Otherwise, thank you so much for taking the time after this long day that started very early.

[00:17:25] And we...

[00:17:26] Wow.

[00:17:26] I'm just going to add this reference in for anyone that's listening.

[00:17:30] You mentioned the Light Collective.

[00:17:32] They just did an amazing project on the AI rights for patients.

[00:17:36] And I highly encourage everyone, no matter what stakeholder you are, take a look at it.

[00:17:41] It's a fabulous document.

[00:17:42] It was created and co-created by a diverse group of patients.

[00:17:45] And there's so many guardrails and frameworks that are being published on fair, trusted, safe, equitable AI.

[00:17:52] This one definitely should be a leading, trusted resource for everyone.

[00:17:55] Thank you.