301–310 of 389 posts
Earlier quoted context omitted.
For every genuinely smart person on HN, there’s a handful of Google warriors who hardly know what they are talking about. I guess that’s not unique to HN, but it can be aggravating just how many commenters can be so “know-it-all” with their comments.
Try making an actual argument rather than simply attacking the attitude or character of the person you're responding to.
For those interested in reading the actual paper, this appears to be it (open access too): https://www.annalsofoncology.org/article/S0923-7534(21)02046... I’ve only skimmed it, but seen quite a few limitations, notably those with “non-malignant conditions at enrolment” (it would be nice to know more what that means), previous cancer or recent corticosteroids use were excluded. Additionally, it’s a case controlled tri…
If you’re tested positive, what’s the probably that you actually have cancer? (The article states 0.5% false positive rate and about 50% true positive rate, but I would need to know the the prevalence of cancer in the population to compute what I am asking for).
I believe the name for what you describe is "positive predictive value" or PPV - defined in the paper as the "proportion of true positives among those with a positive test result". According to the paper, their PPV for cancer detection is 44.4% (28.6%-79.9%, presumably the 95% CI).
As a point of comparison, one source I found reports a much lower PPV for the mammogram - single digits on initial screening, rising to 28% post biopsy: https://www.bcsc-research.org/statistics/screening-performan...
The paper notes that PPV can be a more useful metric than sensitivity. Their multi-cancer approach includes some hard to detect cancers that decrease the overall sensitivity, but increase PPV.
Edit: the paper also states: "The extrapolated PPV reported here based on SEER cancer incidence and clinical stage distribution was 44.4% in the screening-eligible 50-79-year age group, which is higher than that of currently recommended screening tests, as PPV is driven by specificity and population incidence."
They also add the caveat that "studies in intended-use populations that will provide more accurate PPV estimates are ongoing".
Earlier quoted context omitted.
Actually you don't want this, and you are describing a nightmare scenario that everyone who studies health policy understands all too well. Mass screening of healthy people will result in extreme iatrogenics and unnecessary psychological damage and stress (which leads to physical effects as well), not to mention overwhelming the medical system. The fact is that many things that could be detected will never result in…
The argument you’re making has always bothered me because it’s hiding the ball. If finding something that otherwise carries no symptoms is best left untreated, then the fact that you found it should make no difference to the decision. The doctor should say that the best course of action is to do nothing. More information can never be harmful. If you know information is best not acted on in the abstract, then you also…
In my particular field of biomedical research, there has recently been a push for "Diagnostic Stewardship" because more information has demonstrably been harmful to patient well-being.
This is one of those medical revolutions that I am waiting dearly for. Facilities that are not hospitals(to avoid the risk of occupying medical devices that sick people need) built to _regulary_ check up otherwise healthy people for preventive care. Heck, I have so many alerts defined on my monitoring setup for servers to watch for signals of failure before they get too big. But, my own body is not observed until som…
Actually you don't want this, and you are describing a nightmare scenario that everyone who studies health policy understands all too well. Mass screening of healthy people will result in extreme iatrogenics and unnecessary psychological damage and stress (which leads to physical effects as well), not to mention overwhelming the medical system. The fact is that many things that could be detected will never result in…
Perhaps it is true - in the US, anyways - that there's a tendency to overly prescribe more aggressive interventions (read: more expensive ones), but my understanding is that the US model is the exception, not the rule, when looking at the rest of the world.
0.5% false positive rate is actually quite high. I don't have the numbers, but I'd be surprised if more than 1% of the people that would receive such a screening actually already have cancer , which would mean that it produces about as many false positives as false negatives. Picture a population of a million people, all receiving the test. 1% of them have cancer (unknown to them) - half of those people get a negativ…
There's an interesting discussion raging on elsewhere in the thread about the possibility for cultural adaptation that would enable us to cope with statements like "the cancer test was positive, but there's still a decent chance you don't have cancer". There's also an interesting question, taking your roughly 1:1 true to false positive ratio as an example, of whether the marginal true positive does more good for the…
But yes, the higher you drive the prior, the less the false-positives cost. That's a fairly normal situation - it's why they have to evaluate the cost-value of further-investigation against so many different population groups to determine the actual optimal usage.
And yes, the question I was trying to prompt really is "what information would we need to have in order to know when this test is a net positive, from the patient's perspective" (and separately, "from the insurance company's perspective", since that's likely to have a very different answer).
Earlier quoted context omitted.
Your belief has no representation in medical science, which is mostly evidence based. A lot of data has been collected on this, studies (many) have been conducted and the general consensus is that more testing absent symptoms does not lead to improved patient outcomes. That you want to have some kind of theoretical argument in the face of this evidence might be interesting to you but it isn't to me.
Less information is better? That would be a bit of a paradox in science. I suppose we might take 2 steps forward and 1 step backward in the short term. As medical science advances, hopefully we can address any shortcomings from the additional early knowledge. I noticed that pancreatic is on the list. This cancer is almost always fatal because we can't detect it early.
For example, a large number of unnecessary antibiotic treatments, which fuel resistance, are triggered by doing diagnostic testing on patients with no symptoms.
"Diagnostic stewardship" is a concept that exists for a reason.
Earlier quoted context omitted.
> but doctors can't always predict this and aren't perfect decision making machines So mass screen healthy people, collect the data, run models, and get better at it. People like you would rather not try, and this is the same reaction I get from doctors. Applying the same tech that we use to improve ad targeting to disease prediction is a no brainer to anyone whose cushy job doesn't depend on the current medieval sta…
Same vibe as "I would just simply..." What do you think the entire medical, pharmaceutical, bioinformatic, etc. industries have been doing for decades? Are you really so arrogant to think that other people are too lazy and stupid to think of your brilliant idea and that you are simply more intelligent than everybody else? Do you really think people haven't been trying?
Well, let's remember that the AlphaFold team at Google solved the protein folding problem with a relatively small team in a relatively small number of years after extremely large, well funded companies whose primary business was drug development failed to do so for decades.
So yeah, it's been demonstrated to be possible that the current leaders in a field might be significantly less capable than a different team.
Earlier quoted context omitted.
You're talking about praxis, we're talking about research. I agree the praxis can be pretty bad, but in regards to the above commenter's remarks on why people don't just simply collect data and apply it, medical professionals simply cannot try out new treatments the way they are suggesting.
To go further, a large issue/distraction in biomedical research has been Big Tech types coming in assuming that the roadblock all along has been lack of smart computational people in the process. I will not name names, but have seen so many instances of personal tech heroes coming in & claiming the underlying problem was that some of the brightest computational I've ever encountered simply didn't know how to computer…
I am skeptical of the “more data is bad” meme of screening hesitancy. It cannot be scientifically true in the strictest sense, and to the degree it’s an accurate assertion, it really seems to reveal an unscientificness to how screening data is used today in practice rather than that in principle more data is bad.
0.5% false positive rate is actually quite high. I don't have the numbers, but I'd be surprised if more than 1% of the people that would receive such a screening actually already have cancer , which would mean that it produces about as many false positives as false negatives. Picture a population of a million people, all receiving the test. 1% of them have cancer (unknown to them) - half of those people get a negativ…
I feel it's a net win anyway, if the procedure is not invasive and cheap, you could take one every so often (yearly?) and if something shows up you just double check with a better method or w/e. Better than nothing, which is the current alternative.