EP196: Rethinking Medical Care: Dr. H Gilbert Welch on Overdiagnosis

Questioning Overdiagnosis: A Conversation with Dr. H Gilbert Welch

In this episode of The Girl Doc Survival Guide, Christine speaks with Dr. H Gilbert Welch, a general internist and senior researcher at the Center for Surgery in Public Health at Brigham and Women's Hospital. Dr. Welch is known for his critical views on the medical profession, particularly around the issue of overdiagnosis in cancer screening. He explains the concept of overdiagnosis, shares a compelling personal anecdote from his career, and discusses the complexities and consequences of early cancer detection. They also explore the differing patient perspectives on diagnosis and treatment, emphasizing the importance of balanced understanding between medical professionals and patients. The conversation sheds light on the unintended impacts of too much medical care and advocates for a more thoughtful approach to screening and diagnosis.

00:00 Introduction to Dr. H Gilbert Welch

01:31 Understanding Overdiagnosis

03:24 A Personal Anecdote on Overdiagnosis

07:48 The Complexity of Overdiagnosis

11:09 Challenges in Medical Screening

20:24 Final Thoughts and Reflections

Transcript

Christine Ko: [00:00:00] Welcome back to The Girl Doc Survival Guide. Today I am very pleased to be with Dr. H Gilbert Welch. Dr. Welch is a general internist who has worked for the US Indian Health Service, the Department of Veterans Affairs, and Dartmouth. Currently he is a Senior Researcher at the Center for Surgery in Public Health at Brigham and Women's Hospital in Boston. For over three decades, he has been asking hard questions about his profession. His arguments are frequently counterintuitive, even heretical, yet have regularly appeared in the country's most prestigious medical journals, for example, Annals of Internal Medicine, Journal of the American Medical Association, and the New England Journal of Medicine, as well as the Journal of the National Cancer Institute. He has also written op-eds in the Los Angeles Times and the New York Times, and his most recent book is titled, Less Medicine, More Health: Seven Assumptions that Drive Too Much Medical Care. Dr. Welch questions the assumption that more medical care is always better. His research has focused on the assumption as it relates to diagnosis, [00:01:00] that the best strategy to keep people healthy is early diagnosis and the earlier the better. He has delineated the side effects of this strategy. Physicians test too often, treat too aggressively, and tell too many people that they are sick. Much of his work has focused on overdiagnosis in cancer screening, and in particular screening for melanoma, thyroid, lung, breast, and prostate cancer.

Welcome to Gil.

H. Gilbert Welch: Thanks for having me, Christine. I'm glad to be here. 

Christine Ko: Can you first share a personal anecdote? 

H. Gilbert Welch: I wanna share an anecdote, but I think we should really start with what overdiagnosis is. Overdiagnosis is a abnormality that meets the pathologic definition of cancer but is not destined to cause symptoms or death.

Christine Ko: Overdiagnosis is not a misdiagnosis.

H. Gilbert Welch: No.

Christine Ko: It's a true cancer. I have also heard, sometimes people just say, they define it as an epidemiologic term where there's no death, but the symptoms [00:02:00] count too.

H. Gilbert Welch: Yes, absolutely. Symptoms count too. Overdiagnosis refers to the detection of a cancer, either through incidental detection or through screening, something that meets the pathologic definition of cancer, yet is not destined to ever cause symptoms or death. Making the diagnosis of cancer in asymptomatic patients. Lung cancer screening is probably where the term overdiagnosis originated. It's not me, I wanna be clear, nothing to do with me. It was really because of those studies done at the Mayo Clinic in the early eighties where there was an excess of cancer in the screen group that never seemed to disappear that the pulmonologist got very sensitized to the problem of overdiagnosis. And in the subsequent trials, they specifically said that for small lesions, the strategy isn't to [00:03:00] biopsy. The strategy is to repeat the film in three months, six months to establish whether it's growing or not. That's growth assessment protocols. And to me, that's a huge paradigm shift. That is really making use of the diagnostic value of time. Let's see whether this thing is growing.

 I'm gonna stop there and tell you the anecdote. . There are a number of forces at work, but without a doubt, there was one patient early in my career in the early 1990s at the VA in White River Junction, Vermont; small Hospital, who I had already cared for three or four years. I'll call him Mr. Baker. He was a 60-year-old, and he called me because he was worried about hoarseness, and he was a smoker. It had been going on for six weeks. And I thought, Oh here's a guy who should be evaluated. And ENT was just down the hall. And I went and I said, Could [00:04:00] you see this patient for me? He's been hoarse for six weeks. Not sick. It's not getting any better. He came in about four days later, saw the ENT, did an indirect laryngoscopy, removed a small vocal cord tumor, and Mr. Baker's hoarseness resolved. And that should be the end of the story, right? That should be the end of the story, except someone along the way had ordered a chest x-ray. I gotta be honest, that someone could have been me. I don't know. But anyway, someone ordered a chest x-ray. And the chest x-ray showed a widened mediastinum, and all the docs in the audience know what happened next. The radiologist recommended a chest CT. And it turned out that the mediastinal finding was spurious. It was a little bit of rotation or confluence of shadows. The chest was fine, but the chest CT of course, gets a little bit of abdomen. And there on Mr. Baker's right kidney was a five [00:05:00] centimeter renal mass, and it sure looked like renal cell carcinoma. And the minute everyone looked at that, it was clear what should be done next, it should be taken out. And the urologist was very firm on this. Oh yes, this has gotta come out, and so on and so forth. Now I wasn't quite so sure and I was also at the time working in Medicare data at Dartmouth, and I knew the 30 day mortality following a total nephrectomy, which was the operation done at the time, was around 3%. So that's not trivial. It was not a small operation. I talked to the patient, and he was apoplectic. He was going, Are you serious? I come in with hoarseness. You fix my hoarseness, I feel better. And now you want to take my kidney out. Doc, there's gotta be another way. I really need to give him credit. It was his pushing on this that got me to push it. And I said, Look, [00:06:00] let's see what happens. Let's repeat the scan in two months. And we did. And we did the measurements and maybe it had grown a little bit, but let's repeat it two months later, do the measurement. Oh, maybe it shrunk a little bit. And then we kept doing these serial CT scans. Ultimately, we followed this thing for about a decade before Mr. Baker died from something else.

Now he knew the issues involved, and I had asked, I would like an autopsy. And he agreed to an autopsy, and that was one autopsy I had to go to. And indeed there was about a five centimeter renal cell carcinoma, but nowhere else in his body was there any evidence of metastasis. We stumbled on a tumor that was never relevant to him. He was overdiagnosed, and it was a very important [00:07:00] observation for me to see that personally, and I recognize the problem with finding things that don't bother people is you're not sure what's gonna happen in the future. 

Christine Ko: Yeah. I think aside from the overdiagnosis that's in that story is he trusted you, and you listened to him. You even said he pushed you on it.

H. Gilbert Welch: He definitely did. He deserves a lot of credit in terms of getting me a little bit beyond the standard medical mindset. I wanna be clear. As a primary care practitioner. Oh man. It would've been much easier for me just to turf him off to urology and wash my hands. But I was curious too. He wanted to watch it. I wanted to watch it.

Christine Ko: Yeah. One thing that I've been pondering when I think about overdiagnosis from the doctor side, but also as a patient, who's had a cancer diagnosis, is the [00:08:00] question of really what to do about over-diagnosis. If you had the crystal ball and you knew that this was never gonna harm you, in order to really talk about that to a high level and to an appropriate level, you really have to have the right physician patient relationship.

H. Gilbert Welch: Yes, you do. And I wanna be clear, there are really two broad categories of overdiagnosis. One is tumors that are simply eliminated by the immune system. Some tumors are very susceptible to immune elimination, particularly small, early things. The second has to do with the issue of competing risks, where you may have a tumor that is in fact is slow growing, and it is gonna grow, and it would cause symptoms if given enough time, but the patient dies from competing causes of death. So that's particularly important in the elderly.

Prostate cancer [00:09:00] screening. Just about all older men could be said to have prostate cancer. I'm sure I have prostate cancer. The question is, do they have a prostate cancer that's going to create a problem for them before they die from something else?

Christine Ko: Yeah. 

H. Gilbert Welch: So two very different things. One, immune elimination of indolent, relatively non-aggressive lesions, or, things that grow so slowly that people die of something else before they cause symptoms. 

Christine Ko: Yeah. And we don't really know, right? If the immune system or the body has some other mechanism of eliminating an existing cancer, we don't really know the percentage of that 'cause we can't follow that.

H. Gilbert Welch: The percentage of that is a function of two things. How much cancer we find. As we look for more and more subtle lesions, we're bound to be dealing with lesions that are more likely to be eliminated by the immune system.

Christine Ko: Yeah, it relates to what I'll do as a dermatologist sometimes, especially if someone is [00:10:00] covered in sun damage. Maybe there's 10 things that I'm a little bit concerned about, but I'm not a hundred percent sure they're cancer, and I can just keep following them. Okay, let's take off these three that look the worst. And then you come back in a month or two months and we'll see. Or, if something really takes off in the next week, then let me know. So that's the same idea 'cause we can see the skin.

H. Gilbert Welch: Yeah, that's the diagnostic value of time. That's the value of having a longitudinal relationship. 

Christine Ko: My experience with patients though, and you have experience with 'em too, so I'll see what you think: Every patient's different, and even though the patient you started off with, he has this five centimeter tumor and he's upset in a way. I started off with hoarseness. The hoarseness is gone. Now you're telling me that I need this surgery.

H. Gilbert Welch: Yeah. That's what he was. Seriously?

Christine Ko: I'm okay with watching and waiting. I feel fine. Whereas someone else might say, Oh my gosh.

H. Gilbert Welch: Yeah. 

Christine Ko: Dr. Welch, you've saved me. You found this five centimeter tumor. Yeah. Cut it out [00:11:00] asap. Get rid of it for me. So there's no one size fits all.

H. Gilbert Welch: No. Of course, you're right. Patients have different expectations. I think we should be clear. In many ways, we have set those expectations and we, I'm talking about the medical profession or the public health group and so forth, we have had a drumbeat that early cancer detection through screening is gonna save lives, and that it's a harmless intervention that can only help you. That's not true. It's a complex intervention that may have some benefit, but it has known harms. And so until we, begin to even balance the message, I'm afraid we're in a situation where we're gonna have more overdiagnosis. Patients need to understand the problem. It's not a simple problem to explain, but I'm never gonna blame the patients. I think what we've done is we've really overstated the importance [00:12:00] of early cancer detection. What really matters is whether we have treatment. And to me the most powerful example is in breast cancer, where, our treatments have really improved over the last two, three decades. Our understanding of the heterogeneity of the disease has really improved. We recognize different types of breast cancer and we have treatments tailored to the types.

Christine Ko: When I first went to my first oncology appointment, I was shocked by what she outlined for me. It was five different treatments, and that's the treatment plan that I did undergo. And it's partially 'cause it's Her2 positive. As you said, they're all these different subtypes of breast cancer now, and they are treated differently. And ultimately, actually, I would say I was happy in the end to have a Her2 positive tumor, even though it meant more aggressive treatment. But there's a targeted treatment.

H. Gilbert Welch: You had a targeted treatment.

Christine Ko: Yes. [00:13:00] A treatment that should directly get at the really bad cells. But I was really surprised because even though I am a doctor, I was like, Oh, this will be cut out. And I didn't really think about it more than that. And so I was really surprised by all the treatment. And the treatment had its costs. Like its toll on me. I'm not talking about financial, but just emotional and physical and all that. And I will say that there's never much handholding, in a way, of the downsides of treatment. It's more, You want your cancer to be gone, so you should just be grateful and accepting of what comes along with it.

I didn't really think about it, maybe, as carefully as your patient did, where he was like, No, I feel fine. I don't want all of that. For me, I actually, I found this lump myself, so I did have symptoms, this lump.

H. Gilbert Welch: Absolutely. 

Christine Ko: But I was just thinking, Oh, cut the lump out. The symptom I found.

H. Gilbert Welch: You didn't think it was gonna start all these [00:14:00] drugs whose name you couldn't pronounce. 

Christine Ko: No. 

H. Gilbert Welch: Oh yeah. I agree that the toll of treatment's real. Totally real. You're symptomatic, felt a breast lump. You get it get the breast tumor typed, and you get a treatment that's targeted to that. That's good medicine. But still has a toll of treatment, no question. And that's why I'm worried about overdiagnosis, 'cause it expands the number of people who get treated. Some of them go through it for nothing. 

Christine Ko: So what do you think that we should do about overdiagnosis, either from the patient side or the doctor's side? 

H. Gilbert Welch: From the doctor's side, I think we need to first, recognize the problem and acknowledge it. We should definitely think about growth assessment protocols, making use of the diagnostic value of time. That's one way to limit it, but in general, I think our threshold to test has gotten too low. It's too easy to test. Testing [00:15:00] brings up issues. It has a way of unearthing problems you weren't even interested in. We need to recognize that simply doing tests can add a lot of burden. On the patient side, we need to share this story. Patients need to understand that our diagnostic technologies now have remarkable resolution. We're seeing things in the millimeters, and we can test at the molecular level. Whether it's data that makes your life better or not is an open question. We would do well to encourage patients to have a healthier relationship with medical care.

There's been an idea promulgated that the path to health is through medical care. That's not true. Medical care is a great way to deal with 

symptoms, suffering. I think that's a real distinction that we [00:16:00] ought to be a little bit more clear about with patients that, by all means, come and see us when you're sick, but think a little hard about how much to be evaluated when you're well. 

The old joke is that the only healthy patient is the one that hasn't been, adequately worked up.

Christine Ko: Yeah. But you must still believe to some degree in medicine being able to be preventative? 

H. Gilbert Welch: Prevention is such a charged word. Like of course prevention must be good. We ought to be really clear on the different parts of prevention. One is health promotion. I like to think of that as what your grandmother might have told you. Eat your fruits and vegetables. Go play outside and by all means, don't start smoking. That stuff is really important. Whether it's best mediated through the medical care system, that's slightly a different question, but health promotion, finding real meaning, [00:17:00] real food, regular movement. Those are all really important. Finding purpose, socially engaged, that's all very important to health, no question.

But then there's this other side when prevention gets medicalized, and I'm gonna blame Richard Nixon on this. And of course in someone in my generation Richard Nixon is a convenient punching bag. But when he instituted the war on cancer, the answer was testing. Early detection/ screening, that's a very medicalized approach, and it's really not prevention anyway. On the early detection side, and by the way, I wanna be clear, this is a question of how early when patients have new symptoms, they ought to have 'em evaluated. New breast lump, get a diagnostic mammogram. That's not screening, that's a diagnostic. That's good. Patients should present promptly when they feel something's wrong. The question of whether to invite healthy people in [00:18:00] for us to look hard to see if we can find something wrong is a totally different kettle of fish; I would argue doesn't prevent things. It finds things. 

Christine Ko: Yeah. There's also the standpoint of when you said yes, if you have a new symptom, like you feel a lump, then get in as soon as you can to have the diagnostic testing. My own experience: at first, they weren't gonna give me an appointment. 

H. Gilbert Welch: This is when I get angry. You want me to get angry? Should I get angry on your show? Could anything be more crazy than it is easier and cheaper for a woman to get a screening mammogram than a woman who has a new breast lump who needs a referral and will almost certainly be charged for a diagnostic mammogram. Let's be clear. That is crazy. The higher risk woman faces more barriers. That is wrong. 

Christine Ko: Yeah. I find the lump, I go to ob gyn. They're like, Okay, call this number to get your diagnostic mammogram When I [00:19:00] called, they're like, Oh, you're not due for your screening mammogram. I said, No, I, I need one. I have a lump. And the scheduler just sees what's open in front of her. She was like, Maybe in a month or something. I was like, I can't wait for a month. I was like, I don't wanna wait for a month.

H. Gilbert Welch: That story I want you to write up. That's just an incredible story to me though. That's a story you should communicate because it gets to the heart of just how crazy things have gotten with breast cancer screening. It's now become an impediment to women with symptoms. That's the kind of thing that I really blame screening for. It's set up a system where it's extremely easy to get screened. It's harder for the woman with a new breast lump to get a mammogram.

Christine Ko: Yeah. This touches all cancers, also in dermatology, where you have something you're concerned about, and you can't get in to see the dermatologist.

H. Gilbert Welch: Exactly. And I think this is a more general problem in [00:20:00] medicine that we have increasingly directed our attention to the well, partly 'cause it's a good business. It's well remunerated. The patients tend to be well insured. They tend to be pretty easy to deal with in the sense they don't have other comorbidities and so forth. It makes It harder for symptomatic patients to get seen. That's not right. 

Christine Ko: Yeah, that's not right. I appreciate your passion and your dedication to this. Do you have any final thoughts? 

H. Gilbert Welch: No. I'm just happy that you asked me to come on and I'm very happy you shared your story, and I really hope you at least make some effort to communicate it, whether write it up or you just said it on your podcast, but that's an important flaw in our system now. I think that we've gotten so wedded to screening that in fact it's become a impediment for people who are sick and suffering or have new symptoms. 

Christine Ko: Definitely something I'll [00:21:00] have to ponder. Thank you so much. 

H. Gilbert Welch: Thanks for having me on.