Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes
A tech founder used one pediatrician’s public advice to train her own AI. Sonal Patel decided doctors should own the tools built on their knowledge instead. Patel is a pediatrician and neonatologist who built and controls her own AI for the fourth trimester. This episode is based on her article “Physicians must shape AI in medicine, not watch it,” published on KevinMD. You will hear why she believes the doctor who uses AI, not AI itself, is the real threat to the one who ignores it. She explains how insurers already use AI to deny claims, and how she keeps patients from being sold products they do not need. You will learn why a paid subscription model matters before you experiment, why physician critical thinking cannot be automated away, and why the human connection at the center of medicine is the one thing a machine can never replace. Press play to hear why the doctor who ignores AI is more at risk than the one who uses it, and where to start.
Partner with me on the KevinMD platform. With over three million monthly readers and half a million social media followers, I give you direct access to the doctors and patients who matter most. Whether you need a sponsored article, email campaign, video interview, or a spot right here on the podcast, I offer the trusted space your brand deserves to be heard. Let’s work together to tell your story.
PARTNER WITH KEVINMD → https://kevinmd.com/influencer
SUBSCRIBE TO THE PODCAST → https://www.kevinmd.com/podcast
Transcript
Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast. Today we welcome Sonal Patel. She’s a pediatrician and neonatologist. Today’s KevinMD article is “Physicians must shape AI in medicine, not watch it.” Sonal, welcome to the show.
Sonal Patel: Well, thank you for having me. I’m so excited for this conversation.
Kevin Pho: All right, so let’s start by briefly sharing your story, and then we’ll talk about why you decided to write this article on KevinMD.
Sonal Patel: Yes. My story actually begins a little bit rewound. I’m married to my college sweetheart, which is very integral to this conversation because he’s also a physician. We’re raising four kids, and I actually started in the postpartum world, where I created a home health clinic about seven years ago, where we bring medical care to the house.
Fast-forward to where this article came to be. I was on a podcast, and one of the things I think physicians are always battling is the juxtaposition of the wellness culture being a billion-dollar industry versus the health care industry. Whenever the wellness culture fractures, we are the recipients of that consequence and tragedy.
I do a lot of podcasts myself, trying to change medical fourth trimester care. I was on a podcast and found out that the individual sitting in Silicon Valley, a mom of two, just doing a podcast, had used my expertise in shaping her AI. It was just like, “Wait, what are we doing here?”
She was curating that AI with a knowledge base from her podcast, where she brings on expertise, and that was very eye-opening. Again, we are coming into this period of time. I didn’t realize the power of social media, I didn’t realize influencers, TikTok, all of this stuff.
We’ve always been late to the game. But here we have this tool in front of us, we’ve learned a lot, and we can shape it, and we are the experts. If we don’t step up into this light of it, it’s going to cause a lot of headache for us.
So what happened was that I created my own AI. We curated it. We did it. It’s called Naya. It goes with my branding, and it is only for the fourth trimester, because the other thing about my AI is that I always say I’m not solving all of pediatrics, because that’s impossible. For any physician who practices, there is the actual medicine part of it, and then there’s the art of medicine. That’s what we are training in, and we just know it’s not sitting in that algorithm. It’s just not. An AI can never capture the art of medicine. It just can’t, because it’s built on algorithms.
So we were like, “No, we’re going to go ahead and do our AI.” We launched our AI back in September. We had OB/GYNs, perinatal mental health therapists, all the people that touch the fourth trimester, come and test it out, and we saw results that were phenomenal. Up to this point, in the conversations that the AI has had, we have saved families $880 in medical care because of the right information that we have provided.
And I said, “Well, listen, if I’m an expert and I get to control it, and I do get to control it, and we can dive into how we can control it ourselves, why can’t my other physician friends do it? Why is it that they are going to build this off of our resources?” We can’t fight Epic. We can’t fight them. That’s not what we’re trying to do. But 80 percent of medicine is education, and that education is built on the backs of us, the leading experts in whatever aspect that we practice. So that’s where the article comes into play.
Kevin Pho: All right. So from what I’m hearing, in that story of the tech executive in Silicon Valley, she probably took what you said or what you published in the public domain and used it to create her own AI tool, and you decided enough of that. You took control of your own information, your own knowledge, and then created a customized AI.
So in terms of the fourth trimester, we’re going to certainly talk about AI, but just for those who aren’t familiar with the fourth trimester, give us a 30-second blurb about what exactly that is.
Sonal Patel: Yes. It’s actually recognizing medically that three months after baby is delivered, there are a lot of health issues that affect mom and baby both. And what we do is we separate the care versus treating them as a dyad. There are a lot of economic ramifications amid all that stuff too, which is another podcast in itself, but that’s exactly what the fourth trimester is.
I had a medical student say, and she actually was a fifth-year surgical resident, “I know I’ve gone to med school, but no one taught us about the fourth trimester.” And I go, “I know. Don’t worry about it.”
Kevin Pho: Yeah, there’s not a lot that’s been taught. Now, you took that next step, of course, and you created your own AI tool. And I talk to a lot of physicians on this podcast. They’re very on the vanguard and innovative when it comes to integrating AI and medicine. But as you know, not all physicians are as tech-savvy or tech-forward as you are. So what exactly made you take that next step, and how did you find out how to take that next step?
Sonal Patel: So again, it’s my IP. You’re my intellectual property. You’re going to benefit off of me. I’ve worked so hard for it, and I was just tired of it. The other thing is that, particularly in the fourth trimester, my families get targeted a lot, because these families are searching for answers. They are shoved materialistic things that do not solve anything. They’re spending God’s amount of money, and half of the battle is, “No, this is actually normal,” versus, “No, this is actually abnormal, and you have the right to have all these feelings, but I don’t need to give you a product all the time.”
So that was that part of it, that anger that now a tech company might use this AI. My biggest competitor right now is Amazon Baby AI. Amazon has come out and created it, and you do not think that they’re not going to be pushing products on these families. So there’s this anger that my families are being taken advantage of.
I also was really afraid of how to use this as a tool, but it’s actually quite easy. And one thing that I really want to stress is this: Come on, guys, we’ve gone through med school. We’ve gone through residency. We’ve gone through all this training. If we view this as another tool, like TikTok, website development, blogging, and we approach it as another tool, that fear-mongering is already taken away, and we approach it with these intellectual eyes that we all have, because we’ve gone through so much already. We want to learn stuff. We’re constantly learning, and that was the thing. It’s like, OK, well, let’s figure this out.
And in figuring it out, the tools are so in front of us, and they’re so easy to manipulate, that if that’s the one thing I want to tell my fellow physicians, that barrier of entry is not that big. It really isn’t. We did go create a company called PulseMed, and this is where this article actually highlights that as well, because we want to show physicians, and we’ve already created some AI chatbots, how much control they actually have.
My developer, who is actually a medical student, which is kind of, as I always say, physicians. My 11-year-old went to him, and initially it was a three-hour training that my developer, Colton, gave. And he initially asked my 11-year-old, “Hey, what do you think about AI?” And he goes, “Oh, it’s going to rule the world.” After this three-hour training, he comes back, and I go, “What do you think about AI?” He goes, “Oh, I’m going to control AI.”
Kevin Pho: Now, when you talk to your other physician colleagues, how are they interacting with AI most commonly? And you said there are tools available. What are some easy ways for them to, I guess, metaphorically trademark what they’ve learned, their experience, their wisdom, and their knowledge?
Sonal Patel: Well, first of all, always do a subscription model, because then your own intellectual property is underneath that firewall. Please don’t do it on ChatGPT that’s open or anything like that. There are plenty of platforms, Claude. I personally use Magi. I pay $20 a month for it. It’s not that impeding financially.
And literally go on there and explore it. But first, that’s a first step. If you want to start exploring it, please do it with a subscription model, because then everything that you are doing is behind a paid part of it and it doesn’t leak out.
Have fun with it. Play with it if you’d like. Give it prompts. Do explore it, and don’t be afraid to explore it. That’s the second piece of it. First paid, second explore.
And then if you’re really like, “OK, I really would like to create something,” I’m going to plug my company. Come. Come to us, because the way that we also differ is that there are other companies that are targeting physicians, because they know. They know that we have the intellectual property. We’re saying, “No, we’re not going to make money off of your intellectual property,” because again, just as a physician, I have colleagues that are still paying student debt. How ridiculous is that? We’re still paying student debt. And so the way that we built it is that we’re going to build it, give us a flat fee, give us a maintenance fee for it, and then whatever you want to do with it, you know?
So if we can just change it and turn it on its head and say, “No, this is how it’s supposed to be built,” then basically in two years, with how fast it’s going, it’s like building a website. I mean, AI can build a website. I actually had to build a website. I stopped hiring people to do it.
And that’s the part of it. The other part of it is, if you really want to explore it, we do have a CME to give you all for it, because for the chatbots that we have already created, we’re offering CMEs on it. So come, explore it that way. Get your feet wet with it.
If you work in a hospital system, I’m sorry, but Epic is huge. They’re going to use AI. You’ve already signed the contract. That’s not going to go away. They’re learning off of what you are doing. There’s whole stuff with all of this. But your education, going back to medicine, the core of medicine, 80 percent is education, and that’s what you have as that physician-patient relationship. And we should really start investing in that portion of it.
Kevin Pho: Talk to us about some of the dangers if physicians stay passive in this AI era. And in your article, you did give us a corollary back in 2023 about UnitedHealth and Medicare Advantage denials. So for those who didn’t read your article, just tell us about that warning story.
Sonal Patel: Yeah, the warning story is basically, I think we know our health care system is broken. Insurance is there, and insurance is literally using AI for denial. I’ve had a physician who has created an AI to literally write notes to approve claims, because at the end of the day, all this burnout, all of these things, this system is being built off of our backs.
And that is a cautionary tale. If you think that you’re just going to passively fly by it, it’s going to overtake you. And it’s not really these other administrators and stuff like that. It’s actually going to be physicians who have adopted this new tool. They’re the ones who are going to overtake you.
The other thing is there’s a lot of fear-mongering that happens here in the States about AI. Speaking to a lot of thought leaders in this who actually are physicians and who go to different health systems, they don’t see it in those health systems as much as we are so suppressed with the fear-mongering.
So that’s the danger. It’s the physicians who don’t adopt it in a way where they’re like, “Oh, this is just another tool.” Just think of it as learning the stethoscope. And I’m really saying it that way, and I don’t want to minimize it, but honestly, it really is that simple.
Kevin Pho: Now, physicians are always worried that AI may take their jobs, but you’re saying it’s not really the AI themselves, but the physicians who utilize AI, they’re going to be the ones who potentially can replace those physicians who don’t, right?
Sonal Patel: Definitely, because AI needs that critical thinking analysis. OpenAI has come out and said that you have to have that critical thinking in order to use AI, because you can’t blindly trust it. And in that piece of it, we are so trained to go beyond the algorithms, that we are trained as physicians to recognize.
Like, I teach NRP to medical students. NRP is not training for the 90 percent. You should know the 90 percent in your sleep. It’s training for that 10 percent, and I know my math doesn’t add up, but it’s training for that 1 percent. That’s how our training has been done, for that 1 percent. It’s for that inkling in the back of your head.
So that critical thinking that we physicians are ingrained in will not be replaced with AI. It just cannot be replaced with AI, because human beings as a species are so different. We know this as physicians. If you just look at one disease, it does not present in the same way. It follows a system, but when the system fails that disease process, that’s when we step in.
So that’s a part of it that I really want physicians to know too. Our training has been built for this moment, and this critical thinking piece AI will always need. And companies are really asking us to come there, but it’s 200 bucks for an hour? Come on, people. It’s just that they’re trying to take advantage of us again.
Kevin Pho: So as of now, you’re saying that AI cannot replace what we bring, the wisdom, the critical thinking that human physicians can give. But what about in the future? Because AI, as you know, is iterating so quickly, and I’m sure that you’re familiar with that recent paper in Nature that shows that the frontier models like Claude outperform the curated models, right? Is there going to come a point, potentially very soon, where AI can do a lot of the critical thinking that physicians can do? Even not now, but in the near future perhaps?
Sonal Patel: I think there’s one thing that we keep missing in this conversation. We’re human beings. We saw it in COVID, we’ve seen it in major disasters. This human-to-human interaction can never be captured with an AI tool, in a sense. And so if we use it to really ingrain that patient-physician relationship that has been taking a beating, it’s been taking a beating with policies, with insurance companies squeezing us down to five or 10-minute appointments.
That human touch is just so critical. Sure, fine, let AI do all the things it can do, but at the end of the day, that human touch is what’s going to really make sure that we can control the AI aspect of it.
So, for example, my AI keeps learning, for my app. But again, they know I’m behind it. The patients know, the people who use it know that they can touch me every single time. We have consults, we do this. It’s like, “Oh, but I’m trusting Dr. Patel. I’m not trusting the AI.” And that’s the change that I would love my physicians to adopt.
And sure, it can improve, it can work on itself and everything like that, but I truly still believe, and I think COVID showed us over and over again, that human touch, that human community. And maybe this is how it’s going to be used for us too. Now it allows me to spend more time with my patients. That human connection a machine can never replace.
Kevin Pho: We’re talking to Sonal Patel. She’s a pediatrician and neonatologist. Today’s KevinMD article is “Physicians must shape AI in medicine, not watch it.” Sonal, let’s end with some of your take-home messages that you want to leave with the KevinMD audience.
Sonal Patel: Physicians, please do not be afraid of it. Number two, if you’re going to dive into it, please do it under a subscribed model so they don’t use your IP knowledge for it. And then I’m going to plug my company, but if you want to get CME, come. Come check us out, because we’re building models for physicians by physicians, and I think it’s a really amazing frontier to be in, where we can sit here and say, “No, this is our knowledge and we’re going to use this as a tool.”
Kevin Pho: Sonal, thank you so much for sharing your perspective and insight, and thanks again for coming on the show.
Sonal Patel: Thank you so much.
Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes
