Comment by thaumasiotes | Hacker News Reader

Comment by thaumasiotes | Hacker News Reader

Parent Full thread

> Invited or not, if you don't show up to a colonoscopy you don't get one. Which means the study's data on invitations alone cannot be used to evaluate the effect of colonoscopies. Seems the study does provide data that allows this so, and shows clear differences in cancer rates and survival rates for those that did get colonoscopies.

This is not an oversight on the part of the study designers. It's called "intention to treat", and studies are done this way to get past the problem that, when you invite a bunch of people to get colonoscopies (or tell them to comply with any medical procedure), the people who go through with it are statistically very different, in all kinds of ways, from the people who don't. This makes the direct comparison of "received treatment" vs "didn't receive treatment" mostly meaningless.

You're here advocating for everyone to ignore the study and do a direct comparison of "received treatment" vs "didn't receive treatment".

Don't do this.

I am not doing this, I'm saying the study cannot be used to say colonoscopies don't help.

You just did it again.

Whatever

It cannot be used to say that the help either, but you wrote “Seems the study does provide data that allows this so, and shows clear differences in cancer rates and survival rates for those that did get colonoscopies.”

I read it that the invite doesn get you differnces, the colonoscopy itsrlf does. And if that isn't statistically a gold standard, randomization and all tgat, I don't care. What the study says, IMHO, is that invoting people doesn't do any good when people don't show up.

Which is an interesting problem, isn't it? Not inviting people might be an option, one that doesn't provide any benefits, except an number of letters not being sent. So I'd say keep the invites, because every single additional person showing up for a screening is a net win. Additional educating efforts will definetly help so, no idea how those can look like so.

> I read it that the invite doesn get you differnces, the colonoscopy itsrlf does.

Differences with what?

Not with the counterfactual where those people didn’t get a colonoscopy! (At least as far as I can tell from the results being discussed here.)

> every single additional person showing up for a screening is a net win.

Maybe. This study doesn’t tell us much about it.

> because every single additional person showing up for a screening is a net win

We can't say that from this study.

Colonoscopy does appear to be one of the screenings that reduces mortality. (That's why this is such a surprising result and is getting a lot of attention). But that does not mean that colonoscopy screening is useful for people in their 40s. There's a lot of discussion about the benefits and risks of colonoscopy screening for people in their 50s. Colonoscopy, especially as it's practised in the US with heavy sedation, is not risk free.

Now that is a useful discussion. The more I think about the study (the article covering it is just sh*), the more I think it would have been better to compare how many people had an actual colonoscopy between the two groups, compare everything that was compared between a) people with invite who did show up and those without invite that showed up b) people with invite who didn't show up vs. those without invite who didn't show up and c) those who showed up vs. those who didn't (the most important question). The latter can then be used to gauge overall usefullness by age brackets, gender, region and so on. Obviously, as soon as there are risk factors or other test results, we talk a different game all together.

As it stands, this study, while statistically very good, doesn't help anything to answer the really important questions and only confuses people by causing discussions about the study methods, then used to derive conclusions about colonoscopies (utter nonsense, but first principle thinking using supporting science /domains is really en vogue at the moment) instead of discussing the usefulness of colonoscopies in increasing survivability, early detection and decreasing probability of colon cancer.

[removed - beating a dead horse]

The whole study is about the usefulness of regular colonoscopies as a tool to reduce mortality in the general population, and the surprise comes from the already known "usefulness of colonoscopies in increasing survivability, early detection and decreasing probability of colon cancer" not leading to a decrease in overall death rate. See the other comments in this thread about the intent-to-treat principle and why you can't compare only people who accepted the exam.

Don't be a jerk, they are suggesting a repurposing of this studies data, which I wouldn't be surprised to learn is already in the works. That's not to say this study was done wrong, it is assessing a different probe. The thing about invitations is that there are more than one way to make one, this is a single snapshot in time of how a population responds to one particular kind of invite.

You were both to fast for me to edit, and you put it better than I did.

No idea if the study was done correctly or not, I am no expert in medical studies. So to rephrase it, I'd like to see a follow up study as outlined above. Reason being that the corrent one requires a lot of nuance to properly interpret (and I somehow fail to see the point of the results so far, but that ir purely on me), and we all know that technical and scientific nuance is impossible to come by in public discurs. So the risk I see is, that the current study can be seen as showing colonoscopies are pointless and needless, a piint I don't see the study making. And that is not the studies fault, it is us laypeoples and the medias fault on how we cover and consume reporting about medical studies, or scientific studies in general.

I think people need to understand that colorectal cancer isn't one thing. There are a range of cancers that occur in the colorectal system.

Finding a cancer earlier (in terms of staging) is probably a good thing. But finding a cancer earlier (in terms of age of the patient) possibly means they have a harder to treat, more aggressive, cancer. This might be why there's diminishing returns on population screening in younger people. Older people have easier to treat cancers and pre-cancerous polyps.

> I read it that the invite doesn get you differnces, the colonoscopy itsrlf does.

No, it doesn’t, because the two groups of people you’re comparing aren’t actually similar, and therefore any outcome difference can’t be tied to the intervention.

The difference in outcome between those two groups may merely be that the people who go through with the colonoscopy (not fun) are also willing to go through with their doctor’s recommendations on, say, diet restrictions (also not as fun!) so are healthier and have better outcomes.

Reply on news.ycombinator.com