Short Wave - Why you'll probably be misdiagnosed in your lifetime

Episode Date: September 15, 2026

Nearly every person will experience at least one diagnostic error in their lifetime. That means either a delayed diagnosis, a wrong diagnosis – or no diagnosis at all. That’s according to a report by the National Academies of Sciences, Engineering and Medicine from 2015 on diagnostic error. Today on the show we talk to a physician who was personally affected by the crisis, and New York Times editor and author of the book The Elusive Body: Patients, Doctors and the Diagnosis Crisis.If you liked this episode, check out these other episodes:- How science is taking tripping mainstream- What If You Took The "Trip" Out Of Ketamine?- Could Psychedelics Become Tripless?- The Reality Of OCDInterested in more medical science? Email us your question at [e-mail] public media with NPR+ and enjoy perks for over 25 podcasts like this one. This show’s perks include sponsor-free listening. Learn more at plus.npr.org.See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.NPR Privacy Policy

You're listening to Shortwave from NPR. Hey, shortwavers, Emily Kwong, back in the host chair for a day. I missed you all. Hi. So at school, it's all about learning the answers. So you can one day diagnose your patients. But what happens if the medical mystery is you? I kind of felt this hard lump beneath my skin on the right side of my neck just tucked up under my jaw.

This is Diana Sejas. As a med student in the early 2000s, she noticed this hard lump on her neck was growing. Until at some point it got big enough that it was there all the time and I started to just be very aware of it. She was surrounded by doctors and her personal care team wasn't worried. They told her this was probably a swollen lymph node. Diana wasn't so sure.

In her classes, she was learning about the characteristics of abnormalities, including tumors, And she thought this lump had a lot of red flags. It was hard fixed in place, meaning I couldn't really move it around very well. And it just kept kind of growing slowly over time. So she sought out more opinions. But each time... It just was always, oh, don't worry about it, or it's nothing,

or it's just a lymph node, it's just an infection, or most of the time they're like, it's just anxiety or just anxious because medical students get everything. Until years later, she got the lump scanned. And I remember while I was in the scanner, looking at the tech through the window where they can kind of see into the room. And just based on the face that he was making, I was like, oh, I know something's going on. Diana had not even made it up the stairs from that appointment when her doctor called and said, yeah, this is a tumor. By that point, it had infiltrated her lymph nodes and was wrapped around an artery.

And in surgery, her doctor realized it was leaking. So usually when that happens, he'll just kind of put a stitch in it and close it up and it's fine. He went to put a stitch in the artery and he said that the artery fell apart. So the wall was just destroyed. And so, of course, I started bleeding. There was what I understand, a lot of activity in the OR at the time. You don't want to see that happen at all, especially for that major blood vessel to rupture in such a way. Diana learned all of this later. At the time she was under anesthesia and when she woke up,

I was incredibly angry and I couldn't figure out why because everybody was being so nice to me. And things felt very cloudy. At some point, I remember being in the room in the bed by myself. And I looked down at my lap and I was like, why is my grandmother's arm in my lap? That doesn't make sense. My grandma is still in North Carolina. I'm here and I was in New Orleans at the time. Why is my grandmother's arm in my lap? Diana was hallucinating because she was having a stroke. She lost feeling in her left arm and still experienced the side effects to this day, all because of a tumor that was diagnosed far too late. So today on the show, the diagnosis crisis, how diagnostic error affects patients and what the doctors who are the best at diagnosis do differently. You're listening to Shortwave the science podcast from

NPR. Okay, so to delve more into the diagnosis crisis, Diana and I are going to be joined by Alexandra Sifrelin. She's an editor with The New York Times who just wrote a whole book about this issue called The Illusive Body, Patients, Doctors, and the Diagnosis Crisis. So, Alexandra, let's talk about the scope of the problem. After reporting this book, what did you learn about how widespread misdiagnosis is? So one of the things that I found very shocking when I was doing the reporting for my book was just how great the scope is. The National Academies of Sciences, Engineering, and Medicine put

together a big report where they tried to wrap their arms around diagnostic error and how common it was. And they ended up concluding that nearly every person will experience at least one diagnostic error in their lifetime, sometimes with devastating consequences. Wait, every person? Every person. Wow. Yes. And they defined diagnostic error as a incorrect diagnosis, a delayed diagnosis, or a missed diagnosis. So person doesn't get a diagnosis at all. Is it a medical education problem? Is it that doctors don't have enough time with patients? What are the systemic reasons why this is happening? So I think it's multifactorial. And one of the interesting things that came through to me and my reporting was that both patients and physicians would complain about the very structure of the health care system and how it felt like it was organized in this way that wasn't ideal for diagnoses, especially something that wasn't immediately obvious. So patients would say, I have been having this health complaint for a really long time. It took me months to get an appointment with this doctor.

I've gone from doctor to doctor. You know, this is the third person I'm talking to. Now that I'm in the room and I feel like they have no time with me. And then when I would talk to physicians, they would often share the same frustrations. So they would say, I have to see 20 patients in a day. I have at most 15 minutes, maybe less, with each person. I don't have enough time to literally sit and talk with them, or I don't have time to look more thoroughly through their record in a way that I would like to, or I don't have time to even think. You know, you mentioned Alexandra in your book a particularly gifted diagnostician. His name is Gupreit Daliwal. He's a professor of medicine at UC San Francisco. I'm so intrigued by him. What is he doing differently?

Yes. What really interested me about Grapri Daliwal is he is called this master diagnostician, even though he kind of rejects the term. He will say, in his view, being a master at diagnosis means recognizing that you will never be good enough, that you should always be working to improve. That's a real posture of humility. That seems actually essential for being a good detective. in this way? Absolutely, because, you know, one of the things that he and many other physicians who I interviewed in this book explained to me was that you actually have to realize that you can improve in the diagnosis in order to really do the work that goes into becoming a better diagnostician. And in so many cases, physicians don't end up finding out whether or not they got a diagnosis right. And nobody is telling you later, you know that patient, like you got it wrong. Sometimes that happens, but that's not the norm.

There's not like a formal mechanism for that kind of feedback loop. I've never even thought about that. Yeah. Because my work performance, I get the feedback later on how it went. Right. Like whether an episode did well. But doctors are not getting that natural feedback when it comes to diagnosis. Not necessarily. So Dr. Daliwal works in the ER. And so he's often seeing patients,

briefly, and sort of, you know, you see them, you treat them, and then you discharge. And so what he realized was that he wasn't getting that kind of follow up. He occasionally will go back and look at the last week or two weeks' worth of cases and try to go through, did I get it right? Did I get it wrong? And if I got it wrong, why was that? Right. It's a skill that can be taught. Yeah. Diana, I want to bring you back into this conversation. We heard what happened to you at the beginning of this episode. Now you are a pediatric neurologist at UNC School of Medicine.

You are a doctor. When you are diagnosing a patient, what is going through your mind? What are you doing? Yeah, I think it is a lot of things that are happening at once. I mean, you're collecting your information from your patient, so you're taking your patient history, meaning they are telling you everything that is going on with them. You're taking what you're seeing on your physical exam. Trying to see what the body is trying to tell you or show you physically. Trying to put all that together with what I have seen in other patients or what I have learned while I was in training or through literature or through some of these other kinds of learning networks that we have and take all of that.

try to distill it to, okay, this is what I think is going on. And if there are other things that are on that list that are kind of vying for that number one spot, what other kind of workup I need to be able to do to figure out if that is the case. To diagnose properly requires, I'm hearing you both say, deep listening and making space for all of that, being aware of your cognitive biases, and then maybe having some mechanism for accountability and feedback and colleagues who can like come in and say, well, did you think about this? Can AI do this too? I'm biased. I mean, I'm a physician, so I'm going to be biased here. But I don't think that AI can do everything. I have had patients who will say, you know, they put

something into chat GPT and that's how they came up with questions to ask me. And that's absolutely fine. I see that as you are interested enough to engage in your own health care in a way that you know that you need a little bit more support and trying to figure out what questions to ask in this interaction. I also hesitate with AI in some ways because AI is something that is developed by humans, which means it's going to have some of the same biases incorporated as any of other humans. So if we're talking about an algorithm that has intrinsic biases against certain cultures or people of certain backgrounds and they're the ones developing it, well, is that not just using a more technologically advanced tool to do the same things that some of us are unfortunately already doing in these

interactions. I think it has its place for sure, but I definitely don't think that it will be able to replace everything that is important in that diagnostic process in the clinic. Alexandra, do you have anything you want to add on? Yeah, I mean, I basically agree. I think there's some promising areas when it comes to very rare diseases and that, you know, because, you know, of the sheer amount of data, perhaps some AI tools could help with surfacing rare diagnoses options earlier in the process that I'm hopeful about. But to Dr. Sejas' point, like, I also think the human element of medicine can't really be replaced. And it really is the humanity that feels like it's missing from the healthcare system that patients really want

and desire. Yeah. I mean, it sounds like this experience really shaped you as a doctor and how you approach patients in your own practice. Absolutely. I would say I like to think that I approach most of my patients, almost all probably. I'm trying to come from like a trauma-informed care practice model and try to also understand that, I mean, I've had my own negative experience with doctors. I know how this goes. I'm thinking of one patient that I had a few months ago where there was a really negative interaction that the parent had with someone in the healthcare system and I had to just say, okay, let me stop for a second. Let me acknowledge this so that we can figure out how we're going to move forward together in our, you know, doctor-patient relationship. And sometimes that also means

like thinking about different ways that I'll need to communicate. Maybe it means I need to have, you know, grandma in the room or I need to have a support person that I'm working with. And however that happens is however I need it to happen, it's whatever's best for that patient. Diana Sejas and Alexandra Sifrelin, thank you both so much for coming on Shorewave. Thank you. Thank you. If you liked this episode, share it with a friend. We have all kinds of content about medicine and how it's changing.

You'll probably like our episodes on psychedelics and OCD treatment. Check them out at the link in our show notes. This episode was produced by Rachel Carlson and Burley. McCoy. It was edited by our showrunner, Rebecca Ramirez, and fact-checked by Tyler Jones. Jimmy Keely was the audio engineer. I'm Emily Kwong. Thank you for listening to Shortwave from NPR.