Landmark RCT on colonoscopy screening for colon cancer

[wamcp]

https://www.nejm.org/doi/full/10.1056/NEJMoa2208375

“The trial’s primary analysis found that colonoscopy only cut colon cancer risk by roughly a fifth, far below past estimates of the test’s efficacy, and didn’t provide any significant reduction in colon cancer mortality. Gastroenterologists, including Bretthauer, reacted to the trial’s results with a mixture of shock, disappointment, and even some mild disbelief.

“This is a landmark study. It’s the first randomized trial showing outcomes of exposing people to colonoscopy screening versus no colonoscopy. And I think we were all expecting colonoscopy to do better,” said Samir Gupta, a gastroenterologist at the University of California, San Diego and the VA who didn’t work on the trial. And, he said, it raises an uncomfortable question for doctors. “Maybe colonoscopy isn’t as good as we always thought it is.”

In gold-standard trial, invitation to colonoscopy reduced cancer incidence but not death

“This study provides clear data that it’s not as simple as saying, ‘Colonoscopy is the most sensitive test, and therefore it is the best.’ It still prevented cancers.”

www.statnews.com

---

Members don't see this ad.

 

[drema]

I am disturbed by these results. Because it sounds like we have been doing screening scopes with no clear evidence, or maybe I'm naive.

 

[Gastrapathy]

They did a terrible (calculated?) job of actually getting patients the intervention in the colonoscopy arm. 42% of the patients "exposed to colonoscopy" actually had a colonoscopy. So 58% of patients didn't get a colonoscopy but are included as part of "intention to treat." Pretty hard to show a benefit when you fail most of the patients up front. This was a European study where they favor FIT over colonoscopy due to the cost difference. The cynic in me suspects they designed the trial to fail so they could continue that approach. Enroll folks but don't really try to get them scheduled and then include them in the intervention arm. The 50% reduction in death in the patients who had a colonoscopy at 10 years is a little low but that may be due to demographics and the length of followup.

 

[Ubba]

What on earth is an "invitation" for colonoscopy?

Also, per the protocol: "A centrally located body in each of the participating countries will mail a personal letter of invitation containing date and hour of appointment to every individual randomised to the screening group. The control group will not be approached and will not be actively informed about their status as controls. A letter of invitation will be mailed to eligible individuals 6-7 weeks prior to appointment..." So... did they just basically offer them a single slot and if the individual couldn't make it, too bad?

The truth is probably somewhere between the guidelines of the country with the world's reserve currency and therefore the option to spend lots of money on healthcare (US) and those of countries drowning in debt and pension obligations (European ones).

 

[Fpg1245]

What on earth is an "invitation" for colonoscopy?

Also, per the protocol: "A centrally located body in each of the participating countries will mail a personal letter of invitation containing date and hour of appointment to every individual randomised to the screening group. The control group will not be approached and will not be actively informed about their status as controls. A letter of invitation will be mailed to eligible individuals 6-7 weeks prior to appointment..." So... did they just basically offer them a single slot and if the individual couldn't make it, too bad?

The truth is probably somewhere between the guidelines of the country with the world's reserve currency and therefore the option to spend lots of money on healthcare (US) and those of countries drowning in debt and pension obligations (European ones).

I don’t think practice patterns in either country will be affected by this study. there is plenty wrong with the study

The “less is same” approach is really taking center stage in Europe and the UK probably for a myriad of reasons though

 

[sloh]

Intention to treat/screen is how you should present the primary analysis. It reflects what happens in the real world in this population if you deployed this intervention.

There could be a significant selection bias going on with those who went through screening. The 42% of those that agreed to participate in the study essentially self-selected, and a population that self-selects for a colonoscopy may also have certain characteristics and lifestyle traits (exercise, diet, general greater concern for one’s health) associated with lower cancer risk.

This is why randomized controlled trials use an “intent to treat” methodology rather than a “per protocol” methodology. That is, you have to analyze everyone who entered the study and was assigned to receive a treatment (or placebo.) Restricting attention to only those patients who decide to comply with the treatment would likely lead to biased results.

 

[lifeofpi1]

This is a very nicely done study and can certainly be the trigger to change practice patterns. It showed no difference in cancer-specific mortality in the ITT population, period. The goal of this RCT was to evaluate the population-level intervention of screening, not to evaluate the benefit of colonoscopy per individual patient. Hence, "invitation" is the appropriate intervention to measure.

The two major arguments against the study's conclusion that colonoscopy has no benefit are that 1) the rate of participants actually getting colonoscopy was low and 2) there actually was a reduction in death in the per-protocol population.

The 42% of patients in the invitation arm actually receiving colonoscopy is actually not far off from reality. The estimated proportion of healthy individuals meeting criteria for screening that actually undergo colonoscopy in the US is around 60%. Thus 42% should be enough to show a benefit. Also the Norway participant population had a 60% rate.

As sloh pointed out, nobody should be making a conclusion based on the per-protocol population. Otherwise, what's the point of doing an RCT? Matching covariates to compare the per-protocol population is just not going to be enough and that per-protocol population will be self-selected. This will be an observational study.

 

[Ubba]

Intention to treat/screen is how you should present the primary analysis. It reflects what happens in the real world in this population if you deployed this intervention.

There could be a significant selection bias going on with those who went through screening. The 42% of those that agreed to participate in the study essentially self-selected, and a population that self-selects for a colonoscopy may also have certain characteristics and lifestyle traits (exercise, diet, general greater concern for one’s health) associated with lower cancer risk.

This is why randomized controlled trials use an “intent to treat” methodology rather than a “per protocol” methodology. That is, you have to analyze everyone who entered the study and was assigned to receive a treatment (or placebo.) Restricting attention to only those patients who decide to comply with the treatment would likely lead to biased results.

One question (of many) here is: how realistic was the intervention that was deployed? Were patients assigned to the trial arm educated by their physicians about the presumed importance of colonoscopy? As described in the study, it's almost as though they received random fliers or letters in the mail and had to take the initiative to pursue the given appointments. Sounds pretty half-@ss to me.

 

[sloh]

One question (of many) here is: how realistic was the intervention that was deployed? Were patients assigned to the trial arm educated by their physicians about the presumed importance of colonoscopy? As described in the study, it's almost as though they received random fliers or letters in the mail and had to take the initiative to pursue the given appointments. Sounds pretty half-@ss to me.

In the US, which pushes colonoscopy much more strongly than any other nation, 2018 national screening rate was only 61% for ages 50-64. Norway's screening rate results in this study were 60.7%, which is similar to the US. That 61% US screening rate number is also based off interview surveys, which we can expect to be inflated.

 

[chessknt]

I was hoping for someone to post this, the threat of taking the money scope away always brings a lot of interest. I think the idea is that a global population based screening regime might not be worth the cost. Obviously at an individual level the calculus will vary but to date it doesn't look like there is anything better to refute this. I doubt the lobbying interests of GI or hospitals that benefit from Endo will allow anything to change however so if I were GI I wouldn't lose sleep because your interests align with the hospitals.

 

[bronx43]

I was hoping for someone to post this, the threat of taking the money scope away always brings a lot of interest. I think the idea is that a global population based screening regime might not be worth the cost. Obviously at an individual level the calculus will vary but to date it doesn't look like there is anything better to refute this. I doubt the lobbying interests of GI or hospitals that benefit from Endo will allow anything to change however so if I were GI I wouldn't lose sleep because your interests align with the hospitals.

I’m not sure lobbyists have much say here, as a lot of it will depend on what USPTF puts into their screening guidelines and then most pcps will follow that going forward. I doubt insurers will stop covering scopes based on this. However I do think Medicare would likely start to look at them as an area of potential cost cutting, as the budget will likely start to matter in the near future.

 

[chessknt]

I’m not sure lobbyists have much say here, as a lot of it will depend on what USPTF puts into their screening guidelines and then most pcps will follow that going forward. I doubt insurers will stop covering scopes based on this. However I do think Medicare would likely start to look at them as an area of potential cost cutting, as the budget will likely start to matter in the near future.

Cms and usptf are both governmental bodies that can be influenced (it not outright controlled) by lobbying. Just look at that stupid alzheimer drug that was going to bankrupt cms--that only got shut down because of huge political backlash over the price tag. I bet a saavy researcher could even source the donations to people who had influemce over the cms appointees that were responsible for approving it initially.

 

[DrMetal]

They did a terrible (calculated?) job of actually getting patients the intervention in the colonoscopy arm. 42% of the patients "exposed to colonoscopy" actually had a colonoscopy. So 58% of patients didn't get a colonoscopy but are included as part of "intention to treat."

Eureka. Someone actually does a critical analysis and demonstrates that another 'critical', 'landmark' paper is bullsht. Reminds of the Rivers EGDT crap, where everyone thought it great to place central lines and pound people with fluids (a study done on a whopping N=260).

It's amazing how quick we are to believe our own BS literature, without any critical analysis. I learned to analyze papers carefully in graduate school (a form of schooling that actually teaches critical thinking) . . . to never read a paper without a pad of paper, pencil, calculator, work it out yourself and see if it makes sense. If you do that, you might find that some ~50% of the basic science literature is BS. That percentage is likely much higher in the medical literature.

Most physicians are too dumb to notice, and that's what makes studies like this dangerous.

For my money, scope me, find precancerous lesions, cut em out, once every 5 to 10 years . . .sound good to me. GIs are making too much money?! Good, let 'em do so (while they can . . . in 10-20 years we'll have mid-levels and robots scoping).

 

[bronx43]

Cms and usptf are both governmental bodies that can be influenced (it not outright controlled) by lobbying. Just look at that stupid alzheimer drug that was going to bankrupt cms--that only got shut down because of huge political backlash over the price tag. I bet a saavy researcher could even source the donations to people who had influemce over the cms appointees that were responsible for approving it initially.

Maybe, maybe not. Going by this logic we would still be doing PSA since it provides OR revenue from urologists.
I’m giving it 50/50 that USPTF may downgrade their recommendations.
Another way this may play out is that insurers (private plus MA) may still “cover” scopes but it won’t be a free for all buffet where patients get it for free. They can shift some of the cost onto the consumer via a co-insurance and requirement of paying into a deductible. This move itself will prob drastically reduce the number of scopes per year, since patients will then have to shell out cash for a procedure they didn’t really want done in the first place.

 

[NotAProgDirector]

Maybe, maybe not. Going by this logic we would still be doing PSA since it provides OR revenue from urologists.
I’m giving it 50/50 that USPTF may downgrade their recommendations.
Another way this may play out is that insurers (private plus MA) may still “cover” scopes but it won’t be a free for all buffet where patients get it for free. They can shift some of the cost onto the consumer via a co-insurance and requirement of paying into a deductible. This move itself will prob drastically reduce the number of scopes per year, since patients will then have to shell out cash for a procedure they didn’t really want done in the first place.

If that type of cost shifting happens, it's because insurers want to do it to save money. They may use the study to explain the change, but that has nothing to do with it.

The main finding of this study is that mailing an invitation letter for a free colo to a group of people who don't really want it doesn't change many people's minds. Or, if you prefer a more positive spin, that it's difficult to convince people to get a colo and simple mailings are not likely to be very effective.

Otherwise, the study shows exactly what you would expect. There is an initial increase in incidence of CRC in the intervention group -- because you're doing colos so finding those cancers early. There's a late marked decrease in CRC due to prevention, and the slope of those graphs are very different suggesting the benefits will grow with time (although that's an assumption).

The per-protocol sub analysis which says "what would the likely outcome been if everyone had actually had a colo?" shows a 30% reduction, which is just in line with other studies.

Even with the poor colo uptake, the NNT of the study was 455. If you send 455 invites to get a colonoscopy (of which the largest cost would likely be the stamp), you find one more CRC. That's well within the zone that's considered high value.

 

[bronx43]

If that type of cost shifting happens, it's because insurers want to do it to save money. They may use the study to explain the change, but that has nothing to do with it.

Uh, ya that’s the point. I’m not talking about the validity of the study here, simply what I think may transpire on the backend of this which will likely hinge upon dollars and cents. We all know industry participants don’t care about actual outcomes. Only the money talks.

 

[HemeOncHopeful19]

The estimated proportion of healthy individuals meeting criteria for screening that actually undergo colonoscopy in the US is around 60%. Thus 42% should be enough to show a benefit.

What? That is not how science or math works… in fact I think it should be enough to say this is clearly a flawed study if they barely got 2/3 of the national average.

 

[chessknt]

Eureka. Someone actually does a critical analysis and demonstrates that another 'critical', 'landmark' paper is bullsht. Reminds of the Rivers EGDT crap, where everyone thought it great to place central lines and pound people with fluids (a study done on a whopping N=260).

It's amazing how quick we are to believe our own BS literature, without any critical analysis. I learned to analyze papers carefully in graduate school (a form of schooling that actually teaches critical thinking) . . . to never read a paper without a pad of paper, pencil, calculator, work it out yourself and see if it makes sense. If you do that, you might find that some ~50% of the basic science literature is BS. That percentage is likely much higher in the medical literature.

Most physicians are too dumb to notice, and that's what makes studies like this dangerous.

For my money, scope me, find precancerous lesions, cut em out, once every 5 to 10 years . . .sound good to me. GIs are making too much money?! Good, let 'em do so (while they can . . . in 10-20 years we'll have mid-levels and robots scoping).

You realize CMS operates in a zero sum game--they can't take out loans to sustain spending at a level above their funding, their only mechanism to balance their budget is cuts to provider reimbursements. GI making more money means less money to spread to other fields. They designed it this way so we could all try to kill each other off Mortal Kombat style while they sit back and watch. So under this premise yes, I am definitely excited to see someone actually ask if the endo suite needs to be running 14 hours a day 5 days a week doing colos on low risk people. I'll be more excited to see them gut TAVR coverage (or at a bare minimum the reimbursement). We need to get away from pro-procedure medicine and get people to actually think again. Maybe the subspecialists might actually think about the questions I ask them if they made more doing that than a procedure and actually help me.

 

[HemeOncHopeful19]

You realize CMS operates in a zero sum game--they can't take out loans to sustain spending at a level above their funding, their only mechanism to balance their budget is cuts to provider reimbursements. GI making more money means less money to spread to other fields.

That’s not a good reason to get excited about bad science.

 

[DoctwoB]

If that type of cost shifting happens, it's because insurers want to do it to save money. They may use the study to explain the change, but that has nothing to do with it.

The main finding of this study is that mailing an invitation letter for a free colo to a group of people who don't really want it doesn't change many people's minds. Or, if you prefer a more positive spin, that it's difficult to convince people to get a colo and simple mailings are not likely to be very effective.

Otherwise, the study shows exactly what you would expect. There is an initial increase in incidence of CRC in the intervention group -- because you're doing colos so finding those cancers early. There's a late marked decrease in CRC due to prevention, and the slope of those graphs are very different suggesting the benefits will grow with time (although that's an assumption).

The per-protocol sub analysis which says "what would the likely outcome been if everyone had actually had a colo?" shows a 30% reduction, which is just in line with other studies.

Even with the poor colo uptake, the NNT of the study was 455. If you send 455 invites to get a colonoscopy (of which the largest cost would likely be the stamp), you find one more CRC. That's well within the zone that's considered high value.

True about the NNI (number needed to invite) for prevention of once case of CRC, but the real issue is the NNI to prevent one death from CRC was actually about 2500 given the very small mortality difference in the paper. OF course this is likely underestimating the mortality benefit due to poor screening compliance and too limited follow up. I'd love to see 20 year mortality data, 20 year mets free survival, 20 year colectomy free survival, etc, but the data is too immature for that.

As a urologist, I hope we don't see the same thing as with PSAs, where an early backlash from immature data (even more so with PSA, since prostate cancer is slower then colon) led to a big drop in screening, only for more date to emerge (longer term screening data plus info about the poor quality of the initial PLCO trial, where more people in the control arm got PSA tests then in the intervention arm) saying we should switch back except now the USPSTF has its foot in its mouth.

For example, the ERSPC (big, multinational european randomized PSA screening RCT) initially published with 9 year follow up in 2009, found you needed to offer 1410 men PSA screening to prevent one prostate cancer death. Lots of headlines about PSA screening bad, grade D recommendation, etc.

The latest update, published with 16 year follow up in 2019, found you needed to offer 570 men PSA screening to prevent one prostate cancer death. The Swedish cohort of the ERSPC, with 22 year follow up, found the number needed to invite was down to 221. The real benefit is likely even higher, because the same caveats of limited screening compliance, and in the case of PSA screening, significant PSA contamination of the control arm (testing in the control arm) which blunts the observed benefit.

All this is to say we shouldn't make any sweeping changes or generalizations based on what is likely immature data.

 

[sloh]

If that type of cost shifting happens, it's because insurers want to do it to save money. They may use the study to explain the change, but that has nothing to do with it.

The main finding of this study is that mailing an invitation letter for a free colo to a group of people who don't really want it doesn't change many people's minds. Or, if you prefer a more positive spin, that it's difficult to convince people to get a colo and simple mailings are not likely to be very effective.

Otherwise, the study shows exactly what you would expect. There is an initial increase in incidence of CRC in the intervention group -- because you're doing colos so finding those cancers early. There's a late marked decrease in CRC due to prevention, and the slope of those graphs are very different suggesting the benefits will grow with time (although that's an assumption).

The per-protocol sub analysis which says "what would the likely outcome been if everyone had actually had a colo?" shows a 30% reduction, which is just in line with other studies.

Even with the poor colo uptake, the NNT of the study was 455. If you send 455 invites to get a colonoscopy (of which the largest cost would likely be the stamp), you find one more CRC. That's well within the zone that's considered high value.

Screening recommendations should be based off all-cause mortality

Why cancer screening has never been shown to “save lives”—and what we can do about it

The claim that cancer screening saves lives is based on fewer deaths due to the target cancer. Vinay Prasad and colleagues argue that reductions in overall mortality should be the benchmark and call for higher standards of evidence for cancer screening Despite growing appreciation of the harms...

www.bmj.com

"Using disease specific mortality as a proxy for overall mortality deprives people of information about their chief concern: reducing their risk of dying. Although some people may have personal reasons for wanting to avoid a specific diagnosis, the burden falls on providers to provide clear information about both disease specific and overall mortality and to ensure that the overall goal of healthcare—to improve quantity and quality of life—is not undermined.

Discrepancies between disease specific and overall mortality were found in direction or magnitude in seven of 12 randomised trials of cancer screening. Despite reductions in disease specific mortality in the majority of studies, overall mortality was unchanged or increased. In cases where both mortality rates were reduced the improvement was larger in overall mortality than in disease specific mortality. This suggests an imbalance in non-disease specific deaths, which warrants examination and explanation. A systematic review of meta-analyses of cancer screening trials found that three of 10 (33%) showed reductions in disease specific mortality and that none showed reductions in overall mortality.

Such “off-target deaths” are particularly likely among screening tests associated with false positive results, overdiagnosis of non-harmful cancers, and detection of incidental findings. For example, prostate specific antigen (PSA) testing yields numerous false positive results, which contribute to over one million prostate biopsies a year. Prostate biopsies are associated with serious harms, including admission to hospital and death. Moreover, men diagnosed with prostate cancer are more likely to have a heart attack or commit suicide in the year after diagnosis14 or to die of complications of treatment for cancers that may never have caused symptoms.

The overall effect of cancer screening on mortality is more complex than a disease specific endpoint can capture, owing to the harms of further testing, overdiagnosis, and overtreatment. Realisation of this has led to reversal or abandonment of a number of screening campaigns, including chest radiography screening for lung cancer, urine testing for neuroblastoma, and PSA for prostate cancer. Screening for lung cancer and neuroblastoma increased diagnoses and harms without decreasing disease specific mortality. PSA screening increased harms without changing overall mortality; disease specific mortality remains debated."

 

[HemeOncHopeful19]

Screening recommendations should be based off all-cause mortality

Why cancer screening has never been shown to “save lives”—and what we can do about it

The claim that cancer screening saves lives is based on fewer deaths due to the target cancer. Vinay Prasad and colleagues argue that reductions in overall mortality should be the benchmark and call for higher standards of evidence for cancer screening Despite growing appreciation of the harms...

www.bmj.com

"Using disease specific mortality as a proxy for overall mortality deprives people of information about their chief concern: reducing their risk of dying. Although some people may have personal reasons for wanting to avoid a specific diagnosis, the burden falls on providers to provide clear information about both disease specific and overall mortality and to ensure that the overall goal of healthcare—to improve quantity and quality of life—is not undermined.

Discrepancies between disease specific and overall mortality were found in direction or magnitude in seven of 12 randomised trials of cancer screening. Despite reductions in disease specific mortality in the majority of studies, overall mortality was unchanged or increased. In cases where both mortality rates were reduced the improvement was larger in overall mortality than in disease specific mortality. This suggests an imbalance in non-disease specific deaths, which warrants examination and explanation. A systematic review of meta-analyses of cancer screening trials found that three of 10 (33%) showed reductions in disease specific mortality and that none showed reductions in overall mortality.

Such “off-target deaths” are particularly likely among screening tests associated with false positive results, overdiagnosis of non-harmful cancers, and detection of incidental findings. For example, prostate specific antigen (PSA) testing yields numerous false positive results, which contribute to over one million prostate biopsies a year. Prostate biopsies are associated with serious harms, including admission to hospital and death. Moreover, men diagnosed with prostate cancer are more likely to have a heart attack or commit suicide in the year after diagnosis14 or to die of complications of treatment for cancers that may never have caused symptoms.

The overall effect of cancer screening on mortality is more complex than a disease specific endpoint can capture, owing to the harms of further testing, overdiagnosis, and overtreatment. Realisation of this has led to reversal or abandonment of a number of screening campaigns, including chest radiography screening for lung cancer, urine testing for neuroblastoma, and PSA for prostate cancer. Screening for lung cancer and neuroblastoma increased diagnoses and harms without decreasing disease specific mortality. PSA screening increased harms without changing overall mortality; disease specific mortality remains debated."

Cancer can do other bad things to a person before / other than killing them. Whether you think it is worth preventing that or the risk/benefit of trying to can be debated but I would not consider all cause mortality the be-all end-all for cancer screening.

 

[Fpg1245]

You realize CMS operates in a zero sum game--they can't take out loans to sustain spending at a level above their funding, their only mechanism to balance their budget is cuts to provider reimbursements. GI making more money means less money to spread to other fields. They designed it this way so we could all try to kill each other off Mortal Kombat style while they sit back and watch. So under this premise yes, I am definitely excited to see someone actually ask if the endo suite needs to be running 14 hours a day 5 days a week doing colos on low risk people. I'll be more excited to see them gut TAVR coverage (or at a bare minimum the reimbursement). We need to get away from pro-procedure medicine and get people to actually think again. Maybe the subspecialists might actually think about the questions I ask them if they made more doing that than a procedure and actually help me.

Trust me. Engaging in schandenfraude won’t make your paycheck any bigger or your job any easier. CMS will pocket the savings and move on. The rest of it is token nonsense.

 

[bronx43]

Trust me. Engaging in schandenfraude won’t make your paycheck any bigger or your job any easier. CMS will pocket the savings and move on. The rest of it is token nonsense.

Schadenfreude was never about making your paycheck bigger or your job easier.

 

[bronx43]

Cancer can do other bad things to a person before / other than killing them. Whether you think it is worth preventing that or the risk/benefit of trying to can be debated but I would not consider all cause mortality the be-all end-all for cancer screening.

If that’s the position taken, then there needs to be data showing overall reduction in “bad things happening other than death.”
If you want to prevent a particular outcome, throwing out vague terms doesn’t help. First define the desired outcome, then data needs to back that up.

 

[DrMetal]

each other off Mortal Kombat style while they sit back and watch.

I was more of a Street Fighter II kid. I don't know a damn thing about CMS, nor do I want to. The day I get that far into the weeds is the day I quit this b!@h.

 

[Fpg1245]

Schadenfreude was never about making your paycheck bigger or your job easier.

Bitter and obtuse…still still slaving away in the rheum clinic I guess.

Point still stands…A loss for GI will never mean you win. Sorry it just won’t.

 

[bronx43]

Bitter and obtuse…still still slaving away in the rheum clinic I guess.

Point still stands…A loss for GI will never mean you win. Sorry it just won’t.

Slaving away in clinic, damn right. Let me guess. You're cruising around Malibu in a Lambo.

Point still stands. Never said I win. In fact, I lost already. I'm in clinical medicine.
Though, I don't even care about scoping coverage or reimbursement. just pointed out that schadenfreude ain't about winning, but whatever. You do you.