36 ms·

A big part of this is the fact that learning information (about a disease or condition) early is often NOT the same thing as learning information that will chan

A big part of this is the fact that learning information (about a disease or condition) early is often NOT the same thing as learning information that will change a patient’s clinical outcome.

radu_floricica 5y ago
I never understood this. I mean I totally get the concept, but I don't get how it came to influence policy. Put it like this: how many people are you willing to kill to make sure people don't find out they're sick? Because that's what it boils down to. And it sounds... sociopathic.
- sbelskie 5y ago
  I’m not sure I understand what you mean. It seems like you’re framing my motivation as not wanting people to know that they are sick and being willing to trade peoples lives to achieve that outcome, which I don’t think is at all what my above comment says.
  - radu_floricica 5y ago
    It is how I understood it, sorry about that. What did you mean?
- dabbledash 5y ago
  The idea is that treatment also has risks and false positives are an issue. If you gave a mammogram to every 20 year old woman, you would end up doing a large number of unnecessary biopsies and you’d find almost no real cancer. In the end, you’d lose more people than you would save. Or that’s the idea. I’m no expert in this but it makes sense to me conceptually.
  - JumpCrisscross 5y ago
    > If you gave a mammogram to every 20 year old woman, you would end up doing a large number of unnecessary biopsies and you’d find almost no real cancer If you gave a mammogram to every 20-year old, you wouldn't do a biopsy when you got a positive. You'd increase monitoring and maybe suggest lifestyle changes. The same way we don't immediately catherize everyone who comes back with high cholesterol.
    - vidarh 5y ago
      If you gave a mammogram to every 20-year old, you'd cause a significant increase in cancers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4878445/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4878445/
- vidarh 5y ago
  How many people are you willing to kill with too excessive testing programs? Because that is part of the equation too. It's easy to ignore because chances of harm from non-invasive-ish testing like a blood test is low. The problem is that it compounds quickly. Consider: * For a rare condition you need a lot of tests to find them. Let's say you look for something that 1 in a million can be expected to have. * When you find those 1 in a million, the testing needs to save them, which means a proportion of them need to otherwise be significantly affected. Let's say 1 in 10 of them die. * When you find those 1 in 10 million that has the condition that would have died without intervention, early intervention needs to actually make a difference. Let's say 1 in 10 of those actually survive because of early detection. Now you have to do ~100 million tests to save one life. Suddenly 100 million:1 odds of dying as a result of a visit to do a blood test are enough to neutralise the benefit, be it infections or accidents etc.. And that includes secondary effects such as delayed diagnosis because a false negative leads a patient to delay seeking a second opinion once symptoms present. Additionally there's the opportunity costs in terms of saving lives by spending the money elsewhere, such as e.g. awareness of symptoms and the like, or addressing entirely different issues. Of course, if you have a more common condition, and/or a condition that is much more lethal, and/or a condition where early intervention makes a difference, this all changes. But it's worth noting how little mass screening we do - as it turns out, it's hard to find conditions where the benefits are substantial enough to be worthwhile. In some cases, such as the use of mass screening for breast cancer, there has been calls to scale it back some places because it was unclear whether there was a net benefit.
  - radu_floricica 5y ago
    You're missing a step. There are several responses saying basically the same thing (testing kills), and more or less randomly I'm going to answer yours. All have the same step missing. Testing doesn't kills by itself, not in numbers worth mentioning. Treatments - sure, a whole different ball game, they're positively dangerous. But between testing and treatments there should be a specialist that crunches the numbers and comes to a decision. Which, like in the mammogram example above, is not always going to be more aggressive testing. What testing does is give you more information, which in a remotely sane system should lead to better decision making. Of course, I can imagine insane systems where, for example, the patient decides, the insurance pays and the doctor can be sued for discouraging treatments. In this particular combo you probably want to avoid doing mammograms to 20 yo, because the chance of a false positive is 1%, the chance of a true positive is 0.01%, and you end up with perverse chains leading to healthy people doing chemo. Like I said in the original comment, I GET the phenomenon. What I don't get however is how it can get even close to conventional wisdom that you want to avoid testing, as a rule. That's a particular fix to a particularly insane incentive combo, and common sense should make everybody rail against the insane incentive combo, and not act like the niche fix is actually a goal. I can't really explain how this came to be. Maybe people stumbled on an explanation of how extra testing _may_ be harmful, and the idea was so cool that it got stripped of context and became a meme in itself.
- bildung 5y ago
  You assume that the treatment works and has no side effects, I think that is the main source of misunderstanding. A great example is prostate cancer. It often gets detected now, people are informed they have cancer, but most often the correct answer is "watchful waiting", i.e. no treatment, probably forever. But people now know they have cancer, and are frightened, because cancer, and thus press for treatment. But this comes with a 10-90%(!) chance of incontinence and 50-90% chance of erectile dysfunction - for a cancer that most probably wouldn't have caused them bigger problems for their whole life.
- colinmhayes 5y ago
  How many people are you willing to give chemo to even thought they don't have cancer to save a life? Over zealous testing makes outcomes worse.
refurb 5y ago
Yup. Let’s say the test shows “very likely” for pancreatic cancer. Now you do imaging. Ok, nothing there? Now what? Biopsy? That’s general anesthesia now and costly (for the patient or govt). Biopsy is negative. Now what? Start chemo? Watch and wait? For how long? Do a biopsy every 6 months? None of these tests are 100% accurate. If broadly used, a false positive of 0.1% will result in tens of thousands getting unnecessary testing.
- someguydave 5y ago
  exactly, this screening is an invitation for overtreatment.
- kokey 5y ago
  I think for pancreatic cancer early detection just means it gives you more time to enjoy a slightly longer bucket list. For bowel cancers however, this could potentially buy many extra years since it's currently difficult to detect early.
  - refurb 5y ago
    One reason why pancreatic is so fatal is it doesn’t tend to get diagnosed until quite late, so this test could help here. But really early cancer (no identified mass) is more of a watch and wait thing.
- postingawayonhn 5y ago
  > Now you do imaging. Ok, nothing there? Now what? Put the patient on a more frequent screening schedule (blood tests and imaging) to pick up any growth.
  - refurb 5y ago
    Ok, now multiply this scenario by thousands. This sounds like a great way to bust your healthcare budget.

A big part of this is the fact that learning information (about a disease or condition) early is often NOT the same thing as learning information that will change a patient’s clinical outcome.

radu_floricica 5y ago

I never understood this. I mean I totally get the concept, but I don't get how it came to influence policy. Put it like this: how many people are you willing to kill to make sure people don't find out they're sick? Because that's what it boils down to. And it sounds... sociopathic.

sbelskie 5y ago
I’m not sure I understand what you mean. It seems like you’re framing my motivation as not wanting people to know that they are sick and being willing to trade peoples lives to achieve that outcome, which I don’t think is at all what my above comment says.
- radu_floricica 5y ago
  It is how I understood it, sorry about that. What did you mean?
dabbledash 5y ago
The idea is that treatment also has risks and false positives are an issue. If you gave a mammogram to every 20 year old woman, you would end up doing a large number of unnecessary biopsies and you’d find almost no real cancer. In the end, you’d lose more people than you would save. Or that’s the idea. I’m no expert in this but it makes sense to me conceptually.
- JumpCrisscross 5y ago
  > If you gave a mammogram to every 20 year old woman, you would end up doing a large number of unnecessary biopsies and you’d find almost no real cancer If you gave a mammogram to every 20-year old, you wouldn't do a biopsy when you got a positive. You'd increase monitoring and maybe suggest lifestyle changes. The same way we don't immediately catherize everyone who comes back with high cholesterol.
  - vidarh 5y ago
    If you gave a mammogram to every 20-year old, you'd cause a significant increase in cancers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4878445/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4878445/
vidarh 5y ago
How many people are you willing to kill with too excessive testing programs? Because that is part of the equation too. It's easy to ignore because chances of harm from non-invasive-ish testing like a blood test is low. The problem is that it compounds quickly. Consider: * For a rare condition you need a lot of tests to find them. Let's say you look for something that 1 in a million can be expected to have. * When you find those 1 in a million, the testing needs to save them, which means a proportion of them need to otherwise be significantly affected. Let's say 1 in 10 of them die. * When you find those 1 in 10 million that has the condition that would have died without intervention, early intervention needs to actually make a difference. Let's say 1 in 10 of those actually survive because of early detection. Now you have to do ~100 million tests to save one life. Suddenly 100 million:1 odds of dying as a result of a visit to do a blood test are enough to neutralise the benefit, be it infections or accidents etc.. And that includes secondary effects such as delayed diagnosis because a false negative leads a patient to delay seeking a second opinion once symptoms present. Additionally there's the opportunity costs in terms of saving lives by spending the money elsewhere, such as e.g. awareness of symptoms and the like, or addressing entirely different issues. Of course, if you have a more common condition, and/or a condition that is much more lethal, and/or a condition where early intervention makes a difference, this all changes. But it's worth noting how little mass screening we do - as it turns out, it's hard to find conditions where the benefits are substantial enough to be worthwhile. In some cases, such as the use of mass screening for breast cancer, there has been calls to scale it back some places because it was unclear whether there was a net benefit.
- radu_floricica 5y ago
  You're missing a step. There are several responses saying basically the same thing (testing kills), and more or less randomly I'm going to answer yours. All have the same step missing. Testing doesn't kills by itself, not in numbers worth mentioning. Treatments - sure, a whole different ball game, they're positively dangerous. But between testing and treatments there should be a specialist that crunches the numbers and comes to a decision. Which, like in the mammogram example above, is not always going to be more aggressive testing. What testing does is give you more information, which in a remotely sane system should lead to better decision making. Of course, I can imagine insane systems where, for example, the patient decides, the insurance pays and the doctor can be sued for discouraging treatments. In this particular combo you probably want to avoid doing mammograms to 20 yo, because the chance of a false positive is 1%, the chance of a true positive is 0.01%, and you end up with perverse chains leading to healthy people doing chemo. Like I said in the original comment, I GET the phenomenon. What I don't get however is how it can get even close to conventional wisdom that you want to avoid testing, as a rule. That's a particular fix to a particularly insane incentive combo, and common sense should make everybody rail against the insane incentive combo, and not act like the niche fix is actually a goal. I can't really explain how this came to be. Maybe people stumbled on an explanation of how extra testing _may_ be harmful, and the idea was so cool that it got stripped of context and became a meme in itself.
bildung 5y ago
You assume that the treatment works and has no side effects, I think that is the main source of misunderstanding. A great example is prostate cancer. It often gets detected now, people are informed they have cancer, but most often the correct answer is "watchful waiting", i.e. no treatment, probably forever. But people now know they have cancer, and are frightened, because cancer, and thus press for treatment. But this comes with a 10-90%(!) chance of incontinence and 50-90% chance of erectile dysfunction - for a cancer that most probably wouldn't have caused them bigger problems for their whole life.
colinmhayes 5y ago
How many people are you willing to give chemo to even thought they don't have cancer to save a life? Over zealous testing makes outcomes worse.

refurb 5y ago

Yup. Let’s say the test shows “very likely” for pancreatic cancer. Now you do imaging. Ok, nothing there? Now what? Biopsy? That’s general anesthesia now and costly (for the patient or govt). Biopsy is negative. Now what? Start chemo? Watch and wait? For how long? Do a biopsy every 6 months? None of these tests are 100% accurate. If broadly used, a false positive of 0.1% will result in tens of thousands getting unnecessary testing.

someguydave 5y ago
exactly, this screening is an invitation for overtreatment.
kokey 5y ago
I think for pancreatic cancer early detection just means it gives you more time to enjoy a slightly longer bucket list. For bowel cancers however, this could potentially buy many extra years since it's currently difficult to detect early.
- refurb 5y ago
  One reason why pancreatic is so fatal is it doesn’t tend to get diagnosed until quite late, so this test could help here. But really early cancer (no identified mass) is more of a watch and wait thing.
postingawayonhn 5y ago
> Now you do imaging. Ok, nothing there? Now what? Put the patient on a more frequent screening schedule (blood tests and imaging) to pick up any growth.
- refurb 5y ago
  Ok, now multiply this scenario by thousands. This sounds like a great way to bust your healthcare budget.