A few weeks ago, some new research made a lot of noise.
Clip 1: A new study is raising some questions about colonoscopy screenings and how much they actually reduce deaths from colon cancer.
Clip 2: In one of the largest studies ever, European researchers found colonoscopy screenings cut cancer risk by 18 percent, and made no difference in death rates
For two decades, colonoscopy has been a rite of passage for Americans over 50, though the starting age has recently been lowered to 45. Journalist Katie Couric even televised hers after her husband died of colorectal cancer, to emphasize the importance of the procedure. Physicians and public health experts have long urged people to get screened, because the best available evidence suggested not only could colonoscopy find cancer — it could also prevent it. Was everybody wrong? Some doctors challenged the recent study’s findings, insisting that, above all else:
Clip 3: Colonoscopy saves lives.
From the Freakonomics Radio Network, this is Freakonomics, M.D. I’m Bapu Jena. Today, I’ve got a cold, so please forgive my voice. It’ll sound a little different when I talk to you than when you hear me in conversation with our guests. First, Dr. Michael Bretthauer, the lead author of this new colonoscopy study, will tell us what we should really take away from his group’s research on colorectal cancer screening.
BRETTHAUER: So you could just say, well, chapter closed. There’s no difference. This doesn’t work. I don’t think it’s that simple.
But first, Dr. Aasma Shaukat will explain why we rely so much on colonoscopy in the U.S., compared to other countries — and how that can lead us astray.
SHAUKAT: Colonoscopy, as effective as it can be, is heavily, heavily operator dependent.
* * *
SHAUKAT: Colon cancer is the third most common cancer for both men and women in the U.S., accounts for about 150,000 new cancer cases in the U.S. alone, and responsible for about 10 percent of all cancer-related deaths. So it truly is a big deal. It’s common and it’s lethal.
Dr. Aasma Shaukat is a gastroenterologist and a Professor at New York University School of Medicine.
SHAUKAT: How I got into gastroenterology was I wanted to do something public health related. And I thought, “What are some of the large public health questions that we have some good evidence for, but there’s a lot of missing pieces where we could truly make a big difference?”
Aasma has tried to make a difference by focusing on colorectal cancer screening. She’s published nearly 200 articles on this topic and other related ones, and she was lead author of the current version of the American College of Gastroenterology’s Colorectal Cancer Screening Guidelines.
SHAUKAT: In writing guidelines, it’s very important to look at the evidence. We start with very specific questions. For instance, one of the questions might be: At what age should we start screening? When does the benefit start to accrue? A second question might be: Which modality should we be using? A third one might be: How often we should be screening and what benefit can we expect? The idea is then you take all the evidence and kind of weigh it to answer your questions and come up with a recommendation.
The new guidelines — the ones Aasma co-authored — say that most people between 45 and 75 years old should get checked for colorectal cancer with a colonoscopy every ten years, or a stool-sample analysis once a year. There’s a good reason colorectal cancer screening is so strongly recommended.
SHAUKAT: A lot of colon cancers arise in precursor lesions called polyps. So the whole idea is essentially to, one, find cancers at early stages before people are symptomatic, because when detected at early stages, the prognosis is excellent. In fact, it’s one of the cancers where we can actually use the word “cure.” People can have a normal life expectancy if the cancer is found and resected early. And the second goal is to detect these precursor lesions so that by removing them, we can derail that cancer. So we can actually talk about cancer prevention in that context.
The kind of screening Aasma describes, where precancerous polyps are deftly found and removed, is colonoscopy. She says colorectal cancer stands alone across all of oncology in terms of how precisely we can look for it.
SHAUKAT: Colon cancer is unique compared to other cancers where we can actually detect it in these pre-neoplastic conditions. That makes it very different from other cancers where we actually look for the cancer itself. So in that regard, it’s one of the more optimistic cancers to screen for.
Colonoscopy has been recommended to screen for colorectal cancer since the mid-1990s. Uptake struggled initially; in 2000, only around 20 percent of adults over age 50 — then the recommended starting age — said they’d undergone the procedure. But by 2020, that number was closer to 70 percent.
Despite its popularity among gastroenterologists and its perceived value in terms of finding and even preventing cancer, it turns out there are still a lot of remaining questions about colonoscopy. Starting with: how beneficial is it compared to other, less invasive screening methods — like, say, the fecal immunochemical test, or FIT, which analyzes a stool sample?
SHAUKAT: That is a million-dollar question because we don’t know the answer. The two modalities have never been compared head-to-head in what we consider a randomized clinical trial. Having said that, there are two trials ongoing. And one of them is actually in the U.S. I happen to be a part of it. It’s all across the Veterans Affairs hospitals across the country. We’ve enrolled 50,000 veterans and randomized them to yearly FIT, or the stool test, versus a colonoscopy every 10 years. And, as you know, these studies take a long time to get done, and then the outcome is going to be risk of dying from colon cancer. And we’re looking to see if, say, colonoscopy reduces it more than FIT screening. We have to wait 10 years before the results are available. The last patient was recruited in 2017. So in about five years, we’ll hopefully have some results of which test is best.
JENA: So you’ve been involved in devising guidelines for colorectal-cancer screening. What’s the evidence base for them? Are we talking about randomized, controlled trials? Are we talking about observational studies?
SHAUKAT: What we look for is randomized controlled clinical trials because those are considered gold standard of research studies. And in the realm of colon cancer screening, fortunately, there have been several since the 1970s, the largest being one in the U.S. that I work with closely called the Minnesota Fecal Occult Blood Trial, which truly put screening on the map. And then there have been trials in Europe with stool testing. Because it was the original and the first modality, a lot of evidence has gathered around stool testing.
JENA: How do guidelines for colon-cancer screening differ between the U.S. and Europe?
SHAUKAT: The U.S. tends to be an outlier for pretty much the entire world in that we use colonoscopy preferentially as our colon-cancer screening modality. There are about 15 million colonoscopies done in the U.S. every year. Everywhere else, Europe, Canada, and other places that have organized screening programs such as Australia, parts of South America, the predominant colon-cancer testing modality is the stool test, or the newer version of that called FIT. And only because it’s readily available, it’s affordable, it has great evidence behind it and it lends itself nicely to programmatic screening. So in Europe, the predominant modality is the fecal occult blood, and they actually do it every other year. Whereas in the U.S., our recommendations are every year.
JENA: And why is it that we rely more heavily on colonoscopy in the U.S. compared to other countries?
SHAUKAT: We are a resource-wealthy nation, and we like to use the best and the strongest resources we have. And in 2001, Medicare agreed to pay for colonoscopy as a screening tool and truly that’s when the use soared.
JENA: Is there variation in the quality of colonoscopies or in the people who perform them. And is that useful for telling us something about whether or not the identification and removal of polyps may have a causal effect on outcomes?
SHAUKAT: Yes, and that’s a very, very crucial point. So colonoscopy, as effective as it can be, is heavily, heavily operator dependent. And as a result, about 20 years ago, we developed a set of quality indicators for colonoscopy. One of those indicators is, for instance, completion of the colonoscopy. So we’ve set a bar that the completion rate for a screening colonoscopy needs to be 95 percent or higher for an endoscopist. Another bar is the number of pre-cancerous polyps detected by an endoscopist needs to be, at the minimum, 25 percent or more. So those are all indicators that tell us if the exam was able to detect the kinds of things that we want to use the exam for.
JENA: Prior to the last month, what was the evidence base for the benefits of colonoscopy?
SHAUKAT: There was great observational data from pretty large studies that suggested the benefit of screening using colonoscopy could reduce cancer incidents by about 85 percent and the risk of dying from colon cancer by somewhere between 60 to 80 percent. But as you know, observational studies inherently have design issues and they have a lot of biases because the comparison groups aren’t balanced. So therefore, some of the results you’re getting with a group undergoing colonoscopy may be not just the colonoscopy, but other factors that make them healthier or less likely to develop cancer or die from it. And the field has been looking for randomized control-trial evidence on colonoscopy. And we’ve been waiting for a trial on colonoscopy and asking for one, and our wish was answered. And it’s one of those, be careful what you ask for because you may or may not want to know what it says.
What does it say? After the break, we’ll talk with the lead author of this long-wished for randomized controlled trial on colonoscopy that has stirred strong emotions — on both sides of the debate.
BRETTHAUER: People start doing things because they are convinced for some reason, with limited data, that this is the right thing to do.
I’m Bapu Jena, and this is Freakonomics, MD.
* * *
BRETTHAUER: My name is Michael Bretthauer. I am a professor of medicine here at the University of Oslo and the Oslo University Hospital, where I am also a gastroenterologist.
Dr. Michael Bretthaur lives in Norway, which is different from the U.S. in a lot of ways, but for the purposes of this discussion, there’s one difference that stands out.
BRETTHAUER: Most countries here in Europe do not actively recommend colonoscopy as a primary screening test for the general population because people think it’s too invasive, it’s too costly, and there was a lack of randomized trials that can quantify the benefits of that screening test, colonoscopy, until we published our study.
That study is titled, “Effect of Colonoscopy Screening on Risks of Colorectal Cancer and Related Death,” and it was published in October in the New England Journal of Medicine. Its findings rattled gastroenterology, especially in the U.S., because they suggest colonoscopy may not be as effective as previously believed when it comes to detecting colorectal cancers, and also reducing deaths from the disease.
On its face, this runs counter to what previous research has shown. Is this study a game-changer?
As Aasma Shaukat told us earlier, to do a proper randomized controlled trial of colorectal cancer screening methods, you’d need a lot of time. Which Michael and his team had.
BRETTHAUER: We started to plan this trial 15 years ago. We had our first meeting in 2005, and already then, we had a strong interest in finding out what the benefits and the harms are of colonoscopy screening for the general population to be used as a colorectal cancer screening tool. So we wanted to find out because we thought, well, we would like to know, with the best methodology available, which are randomized trials, how large is the benefit as compared to the harms? And then we set out to plan this trial, which happened to be very, very large and needed to run for a long, long time
JENA: Tell me about the study design, where the patients were, how they were recruited, what was done — all that.
BRETTHAUER: The research question we ask is: What are the benefits of introducing a screening program for a general population of people around 60 years of age — which is usually a good screening age for colorectal cancer — what is the effect of such a population screening program? We looked for cities and regions in Europe where such screening was not yet introduced because we wanted a control group of people who did not get that test. And we found four regions in Europe — one region in Norway, in the south of this country, and one region in Sweden, in the middle of Sweden, one region in Poland, and two regions in the Netherlands. We set up colonoscopy centers and we built an infrastructure and then we randomized all the people to either get an offer for a colonoscopy or no offer at all. And there were about 95,000 people living in these cities where we were with the trial. And one third got an invitation for a colonoscopy, and the other-two thirds did not get an invitation. And then of course we did the examinations, which took us about four years. And then we followed everybody in both groups, the people who got the invitation and the people who didn’t get an invitation, followed them for 10 years with regard to the risk of getting colorectal cancer and the risk of dying from colorectal cancer.
JENA: And what was the standard of care for patients in those regions before the trial was conducted?
BRETTHAUER: The standard of care was no screening, which was important for us. There was no screening program set up by the government or the local authorities, and there were no private endoscopy clinics as you would have in the U.S. This is different. This is Europe, so it’s a public healthcare system, and so there will be no availability for people in the control group to get screening. Obviously there were services, if they were referred by their G.P. for some complaints. That was available, but not colonoscopies for screening in the usual care group.
JENA: So in the usual care group, no access to colonoscopy really, unless they had symptoms like bleeding from below or something like that. Were they doing stool testing or any other form of noninvasive testing?
BRETTHAUER: No, no testing at all. No. Nothing.
JENA: How many people got colonoscopies in the invitation group?
BRETTHAUER: The participation rate of the people who got an invitation was different in the four different countries. So here in Norway, where I’m sitting right now, 60 percent, six-zero, of all the people who we invited said yes and underwent a colonoscopy, which is a very high number. In the other countries it was lower. In Poland it was 33 percent of the people. In Sweden, it was about 42 percent. And then the Netherlands, which is not part of the current paper, it was down to 22. So between 22 and 60 percent of the people showed up for their colonoscopy.
JENA: And then what did you find?
BRETTHAUER: We did all these colonoscopies, about 13,000 colonoscopies. We removed a lot of polyps. We found some cancers, obviously, some early cancers. And over here, we have all these registries that follow people over time. So it was easy for us to check if people got cancer or if they died of cancer. We found that after 10 years, the people in the Sweden group, the folks who did not get an offer for screening, 1.22 percent got colorectal cancer, as compared to 0.98 percent in the screening group, which is a relative risk reduction of 18 percent, one-eight percent. That’s the main outcome. So that of course includes all the people who were randomized to screening but did not show up. We also did so-called peer protocol analysis where you only look at the people who actually got the colonoscopy as compared to the people in the control group. Now that is a tricky analysis to do because you always have selection bias of the people who show up versus the control group. There is more uncertainty to these estimates. However, the results were, of course, more favorable for screening. So for risk of colorectal cancer, the effect was 31 percent reduction.
Then we looked at death of colorectal cancer. There, the risk overall of dying of colorectal cancer was low. Even in the control group, it was not higher than 0.3 percent, which is a very low risk. And the reduction in the screening group was not different. And it was 0.15 percent in the peer protocol analysis, which is a 50 percent, five-zero, decrease. But all on very low levels. The risk of dying from colorectal cancer was very low in both groups.
JENA: And then what did you find for the overall or all-cause mortality?
BRETTHAUER: The all-cause mortality was not different between the groups. About 11 percent of people died of any cause in both groups, so there was no difference.
JENA: So how would you then put together the findings? You found an 18 percent reduction in colorectal cancer for those people who were invited, but you didn’t find a statistically significant reduction in colon cancer mortality for that group.
BRETTHAUER: What I would say for colorectal cancer risk is that the real effect is somewhere between 18 and 31 percent risk reduction. And then, I think it’s very important to talk about the absolute risks and the absolute risk reductions. If I want to know, okay, should I do this? I would like to know, so what is my risk to start with? What is my risk of getting this disease? If I buy a car, for example, and I walk into this shop and the car dealer says, “Hey, you can buy this car for 50 percent off” — I would still like to know the price of the car, right? I don’t buy the car just with that piece of information that it’s 50 percent reduced. So I think we should start out saying, well, look, the risk of getting this disease over 10 years is 1.2 percent. If you find this risk interesting enough to do something about it, if you go to a colonoscopy, you can probably reduce it to, let’s say, 0.8 or 0.9, somewhere around there. And that, I think, is the information we need to give to patients or to people to make a conscious decision about if they want to do this or not.
JENA: It’s a hallmark of a clinical study when I see economists tweeting about a study, and one of the points that I saw a lot of my colleagues making is that — so 42 percent of those people ended up getting a colonoscopy. If you look at that group on average and you compare it to the usual care group on average, and you don’t find a statistically significant reduction in colon cancer mortality, is it because the intervention, in this case, the colonoscopy wasn’t effective or because only 42 percent of people in the group got it? And, in economics, the way we deal with that is we view that randomization as an instrument or an instrumental variable. It’s not perfect. But you basically are randomizing people to a higher probability of getting the treatment. And if you do that, the quote-unquote “effect” of colonoscopy screening is actually quite large.
BRETTHAUER: I understand that economists think mostly about death, mortality. And we have no difference there. So you could just say, “Well, chapter closed. There’s no difference. This doesn’t work.” I don’t think it’s that simple. Here in this study, the most interesting endpoint, I think, is incidents, so the risk of disease. And why do I think that? Well, two reasons. Number one, I think, it may be a little early to see the full effect of colonoscopy on death because it takes time from people getting the disease until dying from it. Everybody thought when we designed the trial that 10 years would be enough. We were probably wrong — not just we, making the study, but everybody, the field. It probably takes longer. We will see that because we will follow these people longer. The other argument, however, is that colonoscopy screening is intended to prevent colorectal cancer. So it’s intended to reduce the risk of getting the disease. Therefore, the death end point is only a logical consequence of reducing the incidents. So incidents is, for me, the more interesting endpoint for this study at this time with this screening instrument.
JENA: I would guess that if you follow these patients another five years or 10 years, you might actually find, in that case, that there is a statistically significant reduction in death from colorectal cancer. And I’m curious, what’s your prediction?
BRETTHAUER: My prediction also would be for colorectal cancer death that the effect will get larger until a certain point in time. I don’t know where that time point is. If it’s 12 years or 15 years or 20 years, nobody knows. But My guess would be that we may see one in two years time or in five years time. That will be my prediction, although I’m far from certain. And how large that will be, I have no idea.
JENA: What has been the feedback that you’ve gotten about the work? Like what are people saying and how would you respond?
BRETTHAUER: There were some heated discussions the first day safter the publication, especially in the United States. But things have calmed down and we have had some nice conversations, What everybody says, and I certainly appreciate, is that this is a good study and we did it well, and it was well designed and executed — which we like, obviously, because we put a lot of thinking and emphasis into that — and that it moves the field forward because this is the first randomized trial with this screening tool. I think in the G.I. community, there was disappointment because, especially in the U.S., everybody thought, at least in the field, that colonoscopy would have a far higher benefit. People start doing things because they are convinced for some reason, with limited data, that this is the right thing to do.
SHAUKAT: The study was what we have been waiting for and asking for. The biggest criticism about colonoscopy screening is that we did not have randomized clinical trials showing the benefit or how effective it was.
Dr. Aasma Shaukat again.
SHAUKAT: Observational studies are always optimistic and more rosy than real life. And we’ve seen this time and time again. So, that’s why we do randomized trials. We do expect the differences from observational studies to shrink, but nobody quite expected it to be as low as 20 percent. The lack of reduction in colon-cancer mortality, I think, is somewhat premature and perhaps shouldn’t have been included in the main study because for colon-cancer mortality, the study is grossly underpowered. So just because we don’t see a difference doesn’t mean there isn’t one. But that’s lost when you read the results.
JENA: Do you think people are going to shun colonoscopy now or could it even have like the opposite effect ’cause of all the attention it got?
SHAUKAT: So there was a large public health outcry, and also from the medical community. What I’m seeing is — exactly what you alluded to — is actually such a strong reaction that it might put screening on the forefront of people’s minds and invoke a discussion with their providers, give them that extra nudge to think about it, and then perhaps even schedule it. So we might see some unanticipated benefits of this.
And how would Dr. Michael Bretthauer advise his patients?
BRETTHAUER: I would start with explaining them their risk. And there are very good calculators online where you can actually calculate your risk of getting colorectal cancer. And they would come up with the risk of let’s say, 1 percent over the next 10 years. If they say, “Tell me what I can do and how much I can reduce.” And I would say, “Okay you can go from 1 percent to let’s say a 0.8 percent, which is a 20 percent risk reduction, if you come to me next week in the office and we do a colonoscopy.” And then I would explain to them the harms they need to take into consideration — so perforation and bleeding. And then, finally, I would explain to them what a colonoscopy entails, all the things that are involved with it. And then I think at the end I will tell them, “Now you make your decision.” And some people would say yes, and some people would say no. And that is really what is called shared decision making, between a doctorand patient They need to understand the harms and the burdens, and then they need to make a decision, and that’s not my decision, that’s theirs.
So, what happened when it was his decision?
BRETTHAUER: I had a colonoscopy. I thought for me personally, it was a deal that I was comfortable with, with the numbers that we just talked about. It was negative. So, if I’m gonna do another one in 10 years? I’m not sure. I will look up the numbers then and decide.
As Dr. Michael Bretthauer said, once you start doing something — like colonoscopy — based on limited data, it can be hard to slow that momentum. Colonoscopy is pretty safe overall, but it does come with risks, both from the test itself and from the anesthesia. There’s also the preparation, which can be unpleasant. We should hold invasive tests like colonoscopies to the same standard we hold drugs: randomized controlled trials. Studies like Michael’s, and the research Aasma is doing, bring us closer. For that, we owe them and their colleagues a debt of gratitude.
That’s it for today’s show. I’d like to thank my guests, Aasma Shaukat and Michael Bretthauer. And here’s an idea for you based on my conversations with them. If you want to study the effectiveness of colonoscopy, you need randomization, either in a trial like Michael did, or natural randomization, like we talk about all the time on this show. So what about this? We know doctors are affected when their patients have a bad outcome. In reaction, they might change how they practice. When a patient is diagnosed with colon cancer, their physician might start encouraging colon cancer screening more often to other patients, because of the salience of that recent diagnosis. You might then see higher colonoscopy rates in these patients compared to eligible patients the doctor saw before the colon cancer case. If that happens, you’d have a natural experiment to study if that greater screening led to lower cancer-related mortality. You’d need lots of data over lots of years, but it’s doable. What ideas do you have? Did you hear about the colonoscopy study in the news? Did it change the way you thought about screening? Email me your thoughts, because I’d love to hear them. I’m at email@example.com.
And speaking of research, we know that people typically take advice on which podcasts to listen to based on recommendations from people they know and trust. So, go find someone who knows you and trusts you — and tell them to subscribe to Freakonomics, M.D.
Coming up next week:
Ashisha JHA: We had been caught flat footed, We had missed a window of opportunity to really get this virus under control.
If that voice sounds familiar, it’s probably because you’ve heard Dr. Ashish Jha before. First as a public health expert on T.V. and Twitter, and now, as the White House’s Covid-19 Response Coordinator. He’s been studying pandemics for a long time and next week, he’ll tell us what it’s really like to try to manage one, from the front lines.
JHA: One of the things that you realize is you could have absolutely brilliant ideas that just cannot be implemented
That’s all coming up next week on Freakonomics, M.D. Thanks again for listening.
* * *
Freakonomics, M.D. is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and People I (Mostly) Admire. All our shows are produced by Stitcher and Renbud Radio. You can find us on Twitter at @drbapupod. This episode was produced by Julie Kanfer and mixed by Jasmin Klinger, with help from Jeremy Johnston. Our staff also includes Neal Carruth, Gabriel Roth, Greg Rippin, Lyric Bowditch, Rebecca Lee Douglas, Morgan Levey, Zack Lapinski, Ryan Kelley, Katherine Moncure, Eleanor Osborne, Daria Klenert, Emma Tyrrell, Alina Kulman, and Stephen Dubner. Original music composed by Luis Guerra. If you like this show, or any other show in the Freakonomics Radio Network, please recommend it to your family and friends. That’s the best way to support the podcasts you love. As always, thanks for listening.
JENA: The actor Ryan Reynolds got a colonoscopy, and I was thinking, I’m gonna look at what happens to colonoscopy rates after Ryan Reynolds got his broadcasted colonoscopy.
SHAUKAT: His experience went really well, and there were no complications that we know of. And for what it’s worth, he’s protected for a large number of years to come.
- “Effect of Colonoscopy Screening on Risks of Colorectal Cancer and Related Death,” by Michael Bretthauer, Magnus Løberg, Paulina Wieszczy, Mette Kalager, Louise Emilsson, Kjetil Garborg, Maciej Rupinski, et al. (The New England Journal of Medicine, 2022).
- “Current and Future Colorectal Cancer Screening Strategies,” by Aasma Shaukat and Theodore R. Levin (Nature Reviews, 2022).
- “Relation Between Medicare Screening Reimbursement and Stage at Diagnosis for Older Patients With Colon Cancer,” by Cary P. Gross, Martin S. Andersen, Harlan M. Krumholz, Gail J. McAvay, Deborah Proctor, and Mary E. Tinetti (JAMA, 2006).
- “American Cancer Society Guideline for Colorectal Cancer Screening,” by the American Cancer Society.
- “Key Statistics for Colorectal Cancer,” by the American Cancer Society.
- The Colorectal Cancer Risk Assessment Tool, by the National Cancer Institute.