TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

The madness of reduced medical diagnostics

87 点作者 dynm将近 3 年前

29 条评论

h2odragon将近 3 年前
If a test shows &quot;narrowed arteries&quot; or whatever but it&#x27;s not currently impacting you; <i>inaction is an option</i>.<p>Knew a fellow who was told he had a &quot;weak blood vessel in his brain&quot;, had probably had it forever, but now he&#x27;s retired they can &quot;go in and fix it&quot; and so he goes in for a surgery to remove a threat he was unaware of and that had not hurt him yet.<p>He got out after a week, never really recovered, and died within 2 months. Apparently there was another aneurysm they failed to see or fix; because it <i>couldn&#x27;t</i> have been a <i>direct</i> result of his procedure.
评论 #31448373 未加载
评论 #31447402 未加载
评论 #31447615 未加载
评论 #31447034 未加载
PaulKeeble将近 3 年前
The NHS has gone the other way and wont run diagnostic tests it really needs to. The end result is growing legal action over failure to diagnose conditions. Long covid is diagnosed in just 28,000 patients and yet the office for national statistics says by the symptom studies 1.8 million have it. There is no where I know of, certainly not on the NHS, where you can get a test for microclots, a condition found in the bulk of Long Covid cases and easily treatable but doesn&#x27;t show up in normal clotting tests. You can get micro clot tests throughout a lot of Africa but such is the reluctance to run tests in the UK no labs have bothered. Its something we have known can be tested and treated for 17 months at least.<p>Be careful about reducing the testing just because it drives over treatment because under treatment is quite deadly, cheaper maybe but also has a serious impact of life expectancy and disease burden. Life expectancy is dropping in the UK and disease burden is one of the highest in the western world and growing. The grass isn&#x27;t greener on the low test and treatment side, health care is still increasingly expensive and the results are worse.
评论 #31447375 未加载
rayiner将近 3 年前
&gt; What? If the harms of the biopsy outweigh the benefits, don’t do the damn biopsy!<p>But you’ll get sued if the scan suggests something but you don’t follow up and do the biopsy, and it turns out to be cancer. The author doesn’t seem to realize this is all happening against the background of potential litigation.
评论 #31447391 未加载
the_sleaze9将近 3 年前
I think the author is correct in the acedemic sense, but ignorant of the reality of healthcare.<p>If a doctor receives a test result that shows a suspicious mass, then neglects to order a biopsy to investigate further, there are 2 rough outcomes.<p>1. It was nothing 2. It was cancer<p>If it turns out to be (2), the doctor and the practice&#x2F;hospital&#x2F;whatever gets sued for malpractice.<p>On a more ranty note, why does everyone think they are better at other people&#x27;s jobs than those people are? Where is the basic trust in the other? You think a doctor in the United States after ~15 years of school doesn&#x27;t know that most biopsies turn out to be more harmful that useful? Come on. Restore some basic trust in other&#x27;s competence.
评论 #31448680 未加载
评论 #31449434 未加载
评论 #31450768 未加载
tylerrobinson将近 3 年前
&gt; And I’ll wearily pretend to accept that people are emotional and couldn’t understand Bayesian reasoning or false positives and so we need to worry about stressing them out (#2).<p>&gt; Why are we taking as given that a net-negative decision to do a biopsy will be made<p>The author should stop wearily pretending that anxiety isn’t a common result of a positive test, and recognize that the average person is not comforted by a Bayesian analysis when their health could be at risk.<p>If continuing to a biopsy is as common as the author says (I don’t have any data on it), that suggests anxiety IS a common response and we can’t simply wish it weren’t the case. Mental health is health.
评论 #31448203 未加载
评论 #31448119 未加载
评论 #31447860 未加载
评论 #31447938 未加载
qgin将近 3 年前
We&#x27;re in an uncanny valley in medicine right now where it is good enough to give the impression that we have more control over human health than we actually do.<p>More and more, the general public has come to see medicine as the Genius Bar service department for the human body. Everything that can be fixed, should be fixed. Let&#x27;s get this thing back to like-new condition as we can, right?<p>But in reality, medicine is still much closer to the older model of simply seeking to reduce human suffering. You will die of something one day and most likely you will have 8 other things wrong with you when it happens. Curing all cancer in all humans would only at about 2 years of average increased lifespan because of all the other causes of death that are creeping up right behind it in the background.<p>So there is a tension here of what medicine should conceptually be doing. If, in the effort to &quot;fix everything&quot; (which is impossible), you actually introduce more suffering (by treating a diagnostic finding that may never have resulted in symptomatic disease), then many would say that&#x27;s a failure of medicine.
gumby将近 3 年前
There’s a baseline issue here which is insurance companies and actuaries.<p>If you can afford to be outside the insurance system in the US you can get excellent care. That can include having tests done that aren’t covered by insurance and <i>avoiding</i> tests that an insurance company would normally require just for CYA reasons, but aren’t worth the risk.<p>But the incentives of the insurance-driven care are not aligned with optimal patient outcome.
评论 #31447412 未加载
评论 #31447311 未加载
bell-cot将近 3 年前
Amusing Idea: Let people apply for &quot;Always Full Diagnostics&quot; status, and accept ~0.01% of the population. The &quot;lucky&quot; AFD&#x27;s would get every test, scan, biopsy, follow-up, exploratory surgery, etc. that was remotely plausible. The anonymized results from all those (mostly unwise) things would be a gold mine of public health data. And I suspect there&#x27;d be more than enough obvious adverse outcomes to convince most of the other ~99.99% that &quot;more is better&quot; does not apply to medical diagnostics.
edarchis将近 3 年前
That&#x27;s all nice but we used to recommend systematic PSA screening for all males over 50. We later realized that there was a lot of false positives and that the intervention often brought more risks than it cured actual cancers. So it&#x27;s only recommended for patients with risk factors.<p>Now, consider the GP who has healthy 50yo, does a PSA anyway and it&#x27;s positive. The GP could still decide not to act. But the patient is not so likely not to do anything. Or even understand the odds. At that point, if the patient dies, the GP is likely to get sued because he should have acted. Even if over all of his patients, not acting was more efficient. If the GP decides to act, however, regardless of whether the patient survives, he won&#x27;t be blamed for it. So doctors have a nasty incentive to act.
xsor将近 3 年前
One reason behind why doctors can’t simply do tests then not follow up is simply medical malpractice suits. Often times, once you do the test you’re locked into figuring out everything even to the detriment of the patient for fear of getting your license revoked.
IG_Semmelweiss将近 3 年前
Incentives have a powerful effect in healthcare:<p>- Physicians want to avoid getting sued - Physicians want to get paid , just like everyone else. - The payor in the system is not the patient, but a faraway entity removed from the situation on the ground.<p>In such instances, all 3 factors influence a decision. If a decision to perform procedure X is close to 50&#x2F;50 (test&#x2F;no test)<p>Do you think &quot;no test&quot; will win ?<p>Now put that in the macro context of millions of patient visits. now, since we are talking diagnostic testing, you may need to be tested at a hospital. So now, you are voluntarily exposing yourself to a hub of tired doctors&#x2F;bureaucracy&#x2F;germs.<p>That starts looking like madness , for sure
sudden_dystopia将近 3 年前
I have never understood this rationale to not do more comprehensive diagnostics. Sure, it costs more up front. But wouldn’t catching problems earlier save more money and suffering in the end?
评论 #31446955 未加载
评论 #31446982 未加载
评论 #31447561 未加载
评论 #31447033 未加载
评论 #31450809 未加载
psyc将近 3 年前
Not for the reasons central to the article, but my favorite is the No HPV Test For Men stance. Because men usually don&#x27;t have symptoms and have far less risk of cancer than women..... Just take 10 seconds to ponder how myopic that is. But that&#x27;s how the medical establishment thinks about health care.
maxfan8将近 3 年前
One slightly reasonable common argument I hear for reduced medical diagnostics is: “we don’t have the resources; it’d overwhelm the medical system”. While this may be the case for some tests, there are probably a great number of tests that could be scaled to be done yearly on the whole population.
评论 #31450730 未加载
OrvalWintermute将近 3 年前
This article is rather funny because it misses out on key facts of how doctors practice medicine these days.<p>I listen to my supremely well educated plastic surgeon wife regularly discussing and giving guidance to office staff around patients in their private practice. Their general approach is to get patient statistics, and that is all, without the aggregated vital signs. I asked her once, &quot;Why don&#x27;t you get the patient vital signs too?&quot;<p>Her response was unexpected, &quot;If we get patient vital signs, we immediately become liable for patient overall health. Our malpractice insurance is specific to the practice of medicine for plastic surgeons, and plastic surgeons only. Our front office staff are normally front office only, or aspiring doctors on a break between their undergrad and entry to medschool. Our practice uses no Nurse Practicioners, Physician&#x27;s Assistants, or nurses. We&#x27;re not set up, or staffed appropriately to deliver general medical care.&quot;<p>Conclusion 1: Medical care is characterized by liability control; there are logistics&#x2F;staffing aspects to medicine around specific types of care.<p>Medical Diagnosis we think of as scans, remote tests. You get a CAT scan, you get an MRI, or X-ray, bam, 15 seconds, 5 minutes, or an hour later (after waiting in a hospital for hours) you&#x27;re done and some radiologist or other specialist is interpreting the results and you have a diagnosis. That is true in some cases, but not others. If you get a positive mammogram, or a suspicious lump what is the next step? Welp, that is regularly one of two things - a needle biopsy, or an investigational biopsy. Needle implies small, thin, fairly painless. Not this needle - instead imagine a needle designed for tissue harvesting, like a horse sized needle. This big giant thing needs to go into your breast to suck up enough of the suspicious lump for a pathologist to examine it. Alternatively, you could have an investigational biopsy where a surgeon takes a small amount of tissue from inside your breast with a scalpel. Investigational biopsies via scalpel can be a big scam too, where a car accident occurs, a cut tendon in the hand, and the surgeon decides to open up the arm past the wrist to visualize all the soft tissue up the arm. Lucrative billing enhancements.. But, when it comes to breasts, we know that certain life changes predispose towards a positive mammogram for a limited time window. Stopping breast feeding being the foremost change that can cause suspicious lumps.<p>Conclusion 2: Diagnostic tests can be quite painful, and, at the wrong time, quite unneeded. Some of them are damaging. Good medicine is about as much when to intervene, and when to test, just as much as it is about when not to intervene and when not to test.
评论 #31450959 未加载
daveguy将近 3 年前
This entire argument of it being &quot;madness&quot; hinges on the fact that a rational actor will know they are part of a low risk demographic and while they might worry, they <i>won&#x27;t seek medical care</i> when they are informed they are low risk and &quot;it&#x27;s probably nothing&quot;. This is not a reasonable assumption.<p>I&#x27;m not sure if the author has ever had a biopsy, but most people will seek out care if they get some concerning result, like a nodule on the lungs that&#x27;s &quot;probably benign because of your priors.&quot;<p>Maybe you could opt in to &quot;I am fully rational and will only make decisions using bayesian statistics with accurate priors.&quot;<p>Maybe you could hide test results or generate false results based on probabilities as not to steer someone into an unnecessary procedure. (Yes this it tongue in cheek)<p>Those who really really want to do that next test because they want the information are free to do that. But it seems like the exact mentality that will make you more prone to a botched biopsy. If you keep following the trail of more information you will get to a test that may not be best for your health outcome. That may be the initial x-ray.<p>It seems that if you have not-fully-rational human beings and an obligation to truthfully inform them of the results of medical tests, then maybe not doing the test to begin with has significantly better health outcomes. Can you not chain bayesian probabilities?
dataangel将近 3 年前
I think this article misses that there is a statistical conditioning issue.<p>If the research literature says that patients who get scans showing lumps in their lungs have X% chance of having lung cancer, there is an implicit conditioning there in the patient having complained to the doctor of some kind of symptom that prompted the scan. X% isn&#x27;t the probability of having cancer given lump on scan, it&#x27;s the probability of having cancer given a lump on a scan in patients have breathing trouble. If you test too much you run the risk of there not being data about what the baseline is! Studies for a lung cancer treatment usually take 1000 random people and give them all lung x-rays and then wait to see how many get cancer later and survive the experimental treatment, instead they usually find 1000 people already diagnosed with lung cancer that had it diagnosed by traditional means (e.g. scans prompted by breathing problems). The same reasoning applies to deciding whether to do a biopsy -- most biopsies are going to be prompted by scans from a relevant patient complaint, so statistics computed to determine the risk&#x2F;benefit tradeoff are implicitly conditioned on that. You need a separate study to determine the risk&#x2F;benefit for incidental discovery.
IG_Semmelweiss将近 3 年前
The big problem with testing is that it assumes the human body can be averaged.<p>We are not.<p>One person&#x27;s natural lump is another&#x27;s cancer.<p>One person&#x27;s low heart rate is another&#x27;s natural genetic gift.<p>Modern medicine , particularly compensation, wants to put everything in neat black&#x2F;white boxes.<p>Which is why tests are more dangerous than they seem. You are not average, and you shouldn&#x27;t compare yourself to any perceived average because the human body is not average.<p>You should compare yourself to you, and that&#x27;s all the doctor should use.
gumby将近 3 年前
There is an easy fix but nobody would stand for it: do full longitudinal surveillance of all tests. That is, the gent with the “narrowed” arteries could be followed to see if it becomes a problem (a lot of the metrics are, to simplify, fairly arbitrary and based on a small observational <i>n</i>). Perhaps he does have a stroke, perhaps not.<p>Of course no one would stand for such an intrusive system. But over time it <i>would</i> improve diagnosis.
评论 #31447871 未加载
评论 #31447362 未加载
mreiner将近 3 年前
From what I understood about how the system works in most European countries:<p>A test has a specific sensitivity and a rate of false positives. So for screening (CT in this case) you would want a high sensitivity and low cost, while the false positives would not be your priority. The positives should go through a confirming test (biopsy in this case) with a high specificity. Here cost and sensitivity are usually secondary.<p>Now insurers look at the screening tests rate of false positives. If they deem it too high, they don&#x27;t want doctors to do those tests on a population with a low probability of having the condition you are screening for (low base rate). If the patient belongs to a subgroup shown to have a high enough base rate of a condition, then it makes sense to do the screening.<p>Then you have different patients, some want to get one MRI each year, some only want to run diagnostic after they experience symptoms. I believe most doctors respect that individual risk tolerance within the given framework.<p>Now the thresholds obviously should be revised regularly as cost, test properties and even base rates of diseases change, but I don&#x27;t see a systemic defect here, my blind spot?
chrismeller将近 3 年前
I’ve been shocked at the night and day attitude between the US and Estonia. The US definitely over-tests and Estonia just waits to see if you die and they don’t have to worry about it.
teekert将近 3 年前
This is so true. And I think we need more Bayesian models in Hospital IT systems to help with this. The probability needs to be updated, you shift to a new population with new evidence. It seems so clear, yet somehow it is so intuitive. Why? In many respects, Bayesian reasoning is just what we do, but then sometimes it fails us. We are just not good at accepting risks of any kind I think. Better to not put ourselves into this position seems to be the reasoning. Avoid that &quot;The test told you so!&quot; feeling...
norswap将近 3 年前
&gt; If you do a CT scan and it shows a mass, you’ll order a biopsy.<p>&gt; But because that patient was low-risk, the harms of that biopsy will outweigh the benefits.<p>&gt; Thus, you shouldn’t do the CT scan.<p>&gt; What? If the harms of the biopsy outweigh the benefits, don’t do the damn biopsy!<p>I&#x27;m not sure what the point of doing the CT scan if we&#x27;re not going to act on its result.<p>Sure, you gain the confidence that everything&#x27;s gucci if it turns negative. Is that really worth the stress of the test going positive and the not doing the biopsy? Seems dubious to me.
评论 #31448459 未加载
评论 #31448337 未加载
mdrzn将近 3 年前
I am in a similar boat, I <i>feel</i> healthy but I would like to do a full body and blood checkup to know if there&#x27;s anything coming down the line or if anything&#x27;s out of order at the moment but I do not yet know.<p>Is it better to know? To not know? To know when it&#x27;s the moment? BTW I&#x27;m in Italy so it wouldn&#x27;t even be <i>that</i> expensive to do every couple years.
评论 #31447911 未加载
cflyingdutchman将近 3 年前
I was squarely in the author&#x27;s camp and am now a little less so having seen the impact of false positives on my family. Anxiety is real and harmful, and humans are generally very bad at responding to tiny, explicit risk of big harm (small nodule in lung that may be cancer or not, covid, vaccine side-effects, shark attacks).
评论 #31447429 未加载
a_c将近 3 年前
What about the other extreme, let everyone do all the tests imaginable? Or course it doesn&#x27;t make sense. The balance is on how mych evidence available before a test is suggested. Evidence is the incident rate among the population and the false negative rate(specificity). I&#x27;m using very loosely defined terms here because I&#x27;m too lazy to quote the definition. But no need false dichotomy.
cryoz将近 3 年前
Why do the CT scan in the first place if you are not going to do a biopsy if you find something?
lobocinza将近 3 年前
That&#x27;s why I avoid reading stock market news during bear periods.
SemanticStrengh将近 3 年前
More generally I had read on HN a while ago that this is the reason why doctors don&#x27;t do a Full-body radiography&#x2F;scan. Because many humans have bugs&#x2F;anomalies that are mild&#x2F;non-issues but are ambiguous and anxiety prone. The latter (and most salient reason tragically) is that the doctos by covering much more data, would become legally responsible for not seeing (e.g. a tumor found in a place that wasn&#x27;t the tissue&#x2F;region target of the original investigation) I find those really sad and I want to have a whole body radio (and disclose that I take the anxiety cost willfully and that I abandon my right to sue the doctor for this specific radio), is it possible? Is there a place where this can be done?
评论 #31447431 未加载
评论 #31447351 未加载
评论 #31454779 未加载