64 replies
1d4h
https://youtu.be/FzFT-KuE4BQ?si=7-EXbRz1TD4a5leL&t=56
The video makes the case that the early detection of cancers isn't always a good thing. See 6:30 and 7:00 for specific references to scientific studies. Some cancers (thyroid and skin) can be detected at a much higher rate, but the associated mortality remains constant (we are detecting benign cancers).
To be clear: obviously this is dependent on the cancer. Really my point is that we need studies that show that the screening improves outcomes.
This is how Dutch public healthcare motivates its general aversion to medical testing, as many a foreigner finds surprising and incongruent to the generally OK level of healthcare outcomes provided by the system. You can test and detect, but the error margins are often large (so many false positives and/or false negatives), would generate way too much workload to follow up and thereby cost time and money for better leads. Plus, as you say, knowing early doesn't always or even often mean you can actually change the outcomes.
Now, this is what they say... I have tried to find the actual literature they use to motivate their protocols, and apart from occasionally, have not been able to find that evidence. I'd love to have a resource that demonstrates these sorts of things.
It's not just about keeping costs down, but also about increasing quality of life. If you detect a benign cancer in someone, and they then go on to receive chemotherapy, you've massively decreased that person's quality of life for a significant period with no upside.
A benign tumor doesn’t spread and doesn’t warrant chemotherapy I believe? And if it’s actually cancer than I don’t understand how you would not want it gone as soon as possible to avoid metastasis?
Well, this is talking specifically about the case where the cancer is benign. Of course you want a metastatic one gone.
For the benign ones, that's going to hurt the person's body quite a bit
I think the jump from "protein blood test" to "Chemotherapy" is a bit of a stretch. There are almost surely additional screenings and diagnoses going on there. And I don't think anyone is going into chemo for benign tumors.
Makes sense. Is it possible for the two types of cancer to be confused?
Should also point out that tests will negatively affect your life seriously either way, especially if these benign cancers are common. Think of the time spent going to/from the doctor, the incredible stress (am I dying?), the tests itself: it is damaging the person's life. Detection is good, false positives are unacceptable though.
I recently read Outlive, and he makes the claim that almost all of the increase in survivability for cancer comes from early detection. I think projects like this are extrapolating that out. I think a little heartache and worry is excusable if it means 10 years added to your life. Over time, we'll develop a callous there and take it in stride.
e.g.,
My two friends who got early screening had a few false positives, and one true positive treated early. One later died of heart attack and one lived long enough to get dementia.
My two friends who didn't get early screening, one died of heart attack and one died of cancer that could have been caught. Both died earlier than two above.
Sadly, at this point, you want to die of cancer, but you want to do it when you're 85. It beats a sudden heart attack and it beats dementia. You just want to prolong the outcome.
Precisely.
I don't know any "did early detection which lead to unnecessary chemo" stories..
I know people who died due to the lack of early detection.
I also know people who survived due to accidental pre-symptomatic detection because of unrelated ER scans where they'd have been in stage 4 before having any symptoms.
So more screenings for more types of cancer are absolutely needed. It's more treatable early, and you will have better outcomes.
Patients might worry about lab results is not a good reason not to have early detection. Patients can self select if they are so sensitive.
How would you distinguish a necessary and unnecessary chemo after the fact?
Nobody would give chemo if the tumor is known to benign, after all.
An unnecessary chemo will be one where they couldn't tell, and the chemo ensures they wont tell.
The only way you'll know is from aggregate mortality statistics not budging.
The other aspect is that even if a tumour isn't, earlier detection is often not reducing mortality any more once you get to a decent baseline.
The evidence often does not support early screening in general. Early screening for specific subsets of the population have better results, but are also not unambiguously helpful.
I have seen a few people get benefits from a cancer scare: a refocus onto what matters in their lives. Agree: I would guess most people just get costs. Of my middle-aged friends with health scares only a few addressed the underlying cause (and even fewer are proactively avoiding health issues).
i think you are severely underestimating the human species bias for action
once the train leaves the station theres a lot of steam to go in 1 direction only
Prostate cancer. Lots of nerve endings there. The procedure to remove it can lead you to be incontinent. Let's say you treat the cancer but get damaged by the procedure and can't be as active. Your seditary life style leads to a blood clot and an early death ...in the end you may have lowered your life expectancy as prostate cancer is slow growing
Prostate cancer also came to my mind first. Doctors generally stop screening for prostate cancer after a certain age (70 and older is the recommended cutoff from the US Preventative Services Task Force), because, if the cancer wasn't causing symptoms, it's unlikely to impact quality of life or cause death before something else.
The USPSTF references a lot of meta-analyses dealing with screening outcomes. They make decisions by whether a specific screening practice decreases mortality rates. They explicitly don't even include the financial cost of a screening practice.
It’s more complex than this. There is a pretty narrow sweet spot where early detection actually helps.
If the cancer is very fast growing, it could be too fast for treatment to help at all. Even if treatment helps there likely not a very long period of time before you develop symptoms that would have lead to treatment regardless.
If it is very slow growing, you might outlive the cancer and it doesn’t require treatment. It is effectively but not actually a benign tumor.
You also have to deal with false negative and positives, that could be an order of magnitude higher than the Goldilocks true postives that earlier detection actually made a difference. It’s easy to see how population results will not show much of a benefit.
Probably the most famous slow growing tumor is prostate cancer. As per my friend who is urology surgeon, basically all men eventually catch it, unless they die young. But it goes so slowly and symptoms are rather mild in most cases no invasive treatment is done.
yet, many a malignant tumor may not metastize until you are 90 and dead from other causes, either
My nurse friend said she only discharged 2 people after chemo in about 7 years of service. People have a misguided notion about the odds of survinvg a deadly cancer. They also found the diagnostic procedures for breast cancer was causing the cancer.
I assume you're referring to mammograms. You do get exposed to a significant amount of ionizing radiation in mammography, about 0.4mSv, about 40% of the EPA's annual radiation limit for a member of the public.
That's one of the very good reasons why guidance is women wait until age 45 to get annual screenings and switch to biennial at 55.
At that point the rewards outweigh the risks.
the scenario you're describing would never happen, you don't prescribe chemo based on a blood test of protein markers
I am not sure that is unique to the Dutch, it is the line I have heard in several countries.
Heavy agreement on your last part - if there is substantial evidence to back this up I would like to see it.
As it stands I personally prefer the idea of constant mass testing in order that we learn as much as possible as quickly as possible, including improving the tests from the resulting feedback loop.
When I see the prices of some of that testing, I know that that cannot be the reason, and indeed, establishing personal baselines by regular testing can only be helpful. The aversion to it grounded on taking some average patient, I am convinced.
I really wish there was more transparency, because test aversion is exactly the same protocol you'd invent if you were trying to save money. I want to be able to see which of the two we're dealing with.
Given how the Dutch system seems to be designed to maximise paper filling and busywork instead of healthcare - and of course to make the insurance companies rich - it's no surprise that they're against it.
Medical experts here have very little say in how things are run. It's all bankers and bureaucrats.
Do you have any sources for that?
I've worked for a hospital and I've never been able to find anything approaching a complete balance sheet. Financing it utterly opaque, but I'd love to have something solid before I accuse anyone.
I'm far from an expert on this topic, more on the contrary.
The surgeons use to run the hospital. In contrast with mba's they knew things. I don't know what the difference is precisely but I hear the ziekenfonds use to have people to divide money over treatments (set prices) without their salary depending on their choices.
I don't know about the scale but longer ago we would just build hospital buildings and house a workforce of nuns nearby who had their own garden. Now we some how cant afford to put down a building and with realestate prices on the rise the salaries need to follow. We might not like the factory village concept but if you have to be on call all of the time it seems fkn convenient to me. Cut the salary and give the employees a house, seems a great perk.
Employee shortage is also costing a fortune.
We've created ambitious labor protection laws then we created a loop hole where all you have to do is pay 190% of the salary to a job agencies (uitzendbureu) and no laws apply, anything goes. If you don't like it you can go home. This didn't need to cost 90% of the salary. Unless our labor taxes are now that complicated(?) but that doesn't work as an excuse either.
Our taxes not paying for education doesn't mean we don't have to pay for it eventually (+interest)
I read they are also lacking the money to streamline the processes.
The reason this rings wrong for people is because the reasons early testing is dangerous are entirely social. People are motivated by fear to have every test available, doctors are motivated by fear of being accused of neglect by the patient, doctors are also motivated by the profit that they make from the tests, the manufacturers and patent-holders of tests are motivated to have them done as much as possible, the labs that do tests make money on the number of tests that are done, the nonprofits that campaign based on diseases are expected to message to increase testing for those diseases and accept money from manufacturers and patent-holders, there's motivation to exaggerate the danger of what's detected by the manufacturers and patent-holders of treatments for the disease, and there's motivation by researchers who formulate the criteria for determining whether a particular feature of something detected is potentially dangerous/deserves treatment, and a motivation to give them the most expensive treatment, regardless of whether that treatment is unpleasant; in fact if the treatment is dangerous, it opens up secondary markets.
There's just an enormous number of tailwinds pushing overdetection and overtreatment. The sum of that is what's important, which is that when you test earlier, you often objectively end up with more death and suffering. Which is what a state-run national healthcare systems needs to look at, they can't get lost in the trees. You pick an optimum age for testing that shifts the balance to less suffering and death (and costs), and you look for specific exceptions (genetic, lifestyle, comorbidities) and test just those people early.
Could there be a way in which all testing would help instead of hurt? Yes, but it's political and psychological and not likely to ever happen. You'd have to (as a patient) trust probability in general, and additionally you'd have to trust the probabilities that they're handing you haven't been distorted by the self-interest of others. Not likely for the foreseeable future; maybe 1000 years from now.
If you want to do the study, all you have to do is compare the number of deaths from a thing when people are tested early to the number of deaths when people are tested late. Or just look for other people who have done them. If early testing obviously saved lives, the people who sell testing would tout them everywhere. Instead, they're stuck trying to look for angles to argue that lessening death and suffering isn't the biggest consideration. They recently did this to push breast cancer screening earlier again, by arguing that if you specifically look at black American women, they benefit from early cancer screening. So overall, breast cancer deaths go up, but that's just your privilege talking. Woke conglomerates. Ignore that black people have a unique, neglected, discriminated against, and poverty-ridden situation in the US; in fact, it might even be racist to point that out (in backwards land.)
Of course, you definitely don't have to do the studies, plenty have been done. Anybody saying that they've looked and haven't found them has not looked or has seen them and is not telling the truth. Both positive and negative, about every test. The studies that support earlier and more testing are press released and marketed, though, while the others can be suppressed or simply ignored, unless some public health system or insurance company champions them, and of course they would.
I fully agree.
However, the converse is also true: we don't actually know which protocols are there because they're social, and which are there to benefit the insurers. Both could motivate cost cutting.
I want to see the difference. I want whatever the full story is laid out in front of me.
The evidence is out there in published papers that do health economic analyses of various interventions.
On a population based level it makes perfect sense to not do things even if it could save lives.
This is exactly what NICE does in the UK. If it costs more than 30,000 GBP to give one person one more full quality year of life, then it’s not worth it.
To the point about NICE, it's worth pointing out that this is 1) not how NICE decides. Average cost per QALY being below the preferred level is one factor. Analysis of treatments they approve suggests the "actual" threshold is closer to 45k.
2) These are averages per treatment or protocol and in most areas the NHS spends less than the threshold by simply delivering based on clinical assessment of benefits, and so the threshold is limiting treatment mainly on the fringes where you find a few very expensive treatments with marginal benefits with no cheaper alternative.
In practice NICEs threshold is what any insurer would do: You know how much money there is in the system and try to maximise health output with what is available to you.
In practice NICEs threshold is what any insurer would do
This is what makes Americans getting upset their insurance rejects coverage for medicines so ironic.
Single payer systems do the exact same thing. The difference is in the UK is the option is just isn’t brought up in the first place because the doctor already knows it’s not paid for.
Even then, they'll happily bring it up if you mention you're fine with private referrals or recommendations for something that isn't covered. Almost all single-payer systems are single-payer only in the sense that everyone is covered by the public system, but without preventing people from going outside it. A very few - Norway used to - strictly limits what you can go outside the system for. In the UK about 10% have private insurance on top.
Right. So if you have the $$$
Having lived a decade in NL, my impression was that keeping costs down is the top priority. Unless you have a serious chronic condition or were in an accident, good luck getting somebody to take a look at you.
(irl, after a while you learn to push, exaggerate symptoms etc. or just go back home to get tests and treatment).
Yep, the loudest people get the most help. There's no good solution for that other than to become a bitchy 'client'. It's unfortunate that despite the promises, you still have to 'use it correctly' if you want those good outcomes as a patient.
I can't watch a video at work but I have seen this argument before.
I just find it fundamentally hard to believe that having more data is a bad thing. What we choose to do with that data is a different story, and the actual source of these bad outcomes.
Are you prepared to make the decision, "I have cancer but statistically it has a relatively low likelihood of killing me before I would die of other causes, so I won't do anything about it"?
Are most people?
It might provide enough impetus for getting you to avoid sugar and processed meats though.
Has avoiding those things been shown to stop existing cancer? I thought those foods just increased the chance of getting cancer in the first place. But I'm totally ignorant here, it sounds plausible that some carcinogens work by worsening cancer which would otherwise have been benign, I just haven't heard about that
Depending on the type of cancer, there are correlations with diet changes and slowing growth of cancer. Causation isn't known yet, with the speculations about reasons veganism works I've seen all refuted. Low carb effects seems about increasing ketones and reducing glucose levels in the blood, which I think is being studied with some lung cancers. Type of cancer is important, with different diet changes positively correlated different with cancers not returning (and the wrong change could possibly even promote growth, as different cancers 'feed' on different things). Prostate, breast and lung cancer I'm aware of there being these correlations.
There is a spectrum between invasive treatment and not doing anything.
You could have a scan 3 months later and if there is no progression the doctor schedule a scan 6 months later and then 12 months... If there is a progression he schedule an appointmentwith an oncologist.
The problem is that once the odds of benefiting from early detection is low enough, and the rate of a given condition is low enough, it takes very little for other things to dominate.
E.g. how many of them end up leading to unnecessary biopsies, or scans that are themselves introducing a risk?
For any condition there will be a threshold where too much screening becomes harmful because you're doing so many unnecessary tests that rare errors / accidents come to dominate the benefits.
The question isn't if there's a threshold where more scans do more harm than good, but where it is.
For some things, increased screening will be unambiguously good. For some it takes work to figure it out.
If you're male and live to over 60, you are going to be in this camp regarding the PSA and intervention for Prostate Cancer. Two GPs, a Urologist and an Epidemiologist (none of whom know each other btw) have all said to me "you will die with this not of this" because they can trace the dynamics of my presentation.
Enhanced imaging and blood tests alone didn't do this: their intuition based on progression and behaviour of the system as a whole did.
Treat the person, have a longterm relationship with your health provider.
I think the same thing.
That said, I think engineers seem to believe other careers should use logic in a straightforward fashion, but they are more complicated.
For example, lawyers cannot use logic in the same way, because the court can almost arbitrarily say some facts must be ignored.
In the same way, medicine has the hippocratic oath. It is "do no harm". It is not "find the cause" or "cure the patient".
It seems doctors have made their peace with this and are aligned with it, since the overwhelming majority 88% have do-not-resuscitate orders for themselves.
Data isn’t necessarily good because medicine and biology are messy and inaccurate. I just went through a scare myself with elevated markers on a typical blood panel. Lots of fuss, anxiety, and cost for zero gain. At the end, I learned that human bodies vary so much that we’re was just no way to know upfront if a finding was a concern or if my body was just on a tail end of a bell curve. Turns out, if you fully scan people, we all have lumps, bumps, and various anomalies. How much do you spend “treating” and investigating this stuff? I wasted my own time and precious time with doctors for nothing, increasing costs to society as a whole.
That kind of data, the costs, we have tons of. That’s why pretty much every medical association regardless of culture has limits on recommended screenings.
If your are choosing not to do anything based on the data, gathering the data is objectively a net negative. There are financial costs related to taking the tests as well as emotion costs related to false positives and even with deciding not to act with possibly true positives.
There needs to be a net positive action on a subset of the cases to outweigh the costs of gathering and sharing the data.
Most data is crap, and you generally can't tell where the needle in the haystack is.
Having more consistency between doctors would already be a change needed to actually use data. You will find it matters more than you'd like. We can't all have the best doctors, but we could use data to level the outcomes.
This is not about early detection in general, but rather a specific test of dubious utility, specifically full-body MRI, which often leads to tons of follow-on tests and invasive procedures that may have zero benefit.
For a test with high enough specificity and sensitivity for early detection, it's likely that it would be quickly adopted, and then studied to show that it actually improves outcomes without undue cost (not merely dollar cost but also health cost) to people in terms of treatment and its side effects.
They specifically talk about using the fully body MRI for the purpose of detecting cancers. I'm not sure how you missed that. You literally had to watch 10 seconds of the clip.
It's interesting because they explicitly talk about follow up testing (@2:10) which is to say that multiple methods are used if the MRI indicates a problem. So yes, the initial MRI may produce a misdiagnosis and that is a defect of the test. But the commentary in the video obviously suggests that additional testing for early detection is done. That's partly a problem as you pointed out, but then it clearly also indicates the scope of the conversation is more broad.
They didn't miss it at all.
You've misunderstood the comment's point, which is "just because prospective full-body MRI is bad doesn't mean all diagnostics are".
It's true that some of the commentary is specific to the full body MRI itself (such as misdiagnosis due to an imaging artifact), however many of the claims in the video are very general. And the scientific study they referenced at 6:30 has nothing to do with fully body MRIs. The study is in regards to thyroid cancer overdiagnosis due to screening (using ultrasound and not MRI). This is clearly a statement regarding the effectiveness of screening. And, yes, it is specific to the cancer.
I'm not saying anything that the studies aren't saying. For some kinds of cancers and for some kinds of screening methods, screening can result in overdiagnosis.
At 6:30 it's about a specific test for thyroid cancer, which as discussed in my second paragraph, was not found to actually improve outcomes for that specific type of cancer, not cancer in general.
However, early detection is responsible for greatly improving outcomes in many specific cancers. Full body MRI is not the test to achieve that. GRAIL's gallery test might be one to do it for many classes of cancer, but that still remains to be fully seen.
The general of idea of early detection is still an extremely promising one for most types of cancer, and in particular for some of the deadliest, like ovarian and pancreatic cancer.
OK, so you had to watch a little more than ten seconds.
There's no doubt that some diagnostic tests - like getting a full-body MRI as a precaution - may do more harm than good. Your apparent mistake is thinking that means all diagnostic tests probaby do.
We'll have to figure out which one this is; it's a start of that process. We've demonstrated we can do it; now we have to figure out if we can distinguish between "big bad scary" cancer and "whatever it won't kill you" cancer.
How are we supposed to increase outcomes if we don't have the screenings?
Sure maybe the first few years the outcomes may not change, but what about 20 years from now, when a generation of scientists can detect cancer years earlier?
The defeatist attitude of "well it doesn't fix anything right now" is really strange. Once we have reliable detection, it means we can find reliable treatment as well.
Do you mean improve outcomes? Assuming that’s what you mean, improvement of outcomes for cancer patients seems more reliant on proper treatment rather than detection.
That’s right now, because we don’t have early treatment options. Once we can early detect I bet improved outcomes will increase as we develop new ways to fight them.
Yes, I agree, but it also could lead to over treatment. If you detect and treat cancer that otherwise never would have caused problems, it was all for naught. Cancer rates found in autopsies for other causes of death are probably higher than you think.
You can fix overtreatment but you need early detection first.
The screening target should be for malignant cancers, not all cancers.
I would be wary of a study conflating screening with outcomes — that is unnecessarily coupling too many factors. [1]
The screening isn’t an intervention, so why would it make sense to study its effects on an outcome? Those effects are indirect, mediated by the causal factors of the choice of treatment.
Am I missing something?
Perhaps you are looking for psychological effects of getting an early test?
[1] Treatment outcomes depend on the current state of technology and the practice of medicine.
Because screening leads to interventions when it is the data used for deciding whether to do an intervention.
If a screening procedure leads only to interventions that would be made in time anyway, it doesn't provide a benefit.
If it leads to extra interventions and no drop in mortality, it may be a net negative.
E.g irradiate enough people to find more cancers and not improve outcomes, and some of your screening may lead to cancers.
The problem with this is that they haven't done the long term studies (which they admit). They also don't consider that once these are cheap and regular enough you get the change over time which should get you a lot less false in positives.
The more data the better though. We need to train the models to understand what is worth doing over time.
It’s a risk assessment like any other. Probability/impact.
My wife ultimately lost her life to metastatic melanoma, which was believed to be in remission. Had there been a way to detect the proteins associated with the mets that developed ahead of symptoms, the odds are she we be alive and thriving.
In other scenarios, say most prostate cancers, early knowledge has low or negative benefits.
I think Vinay Prasad (MD) has an excellent book or paper on this topic as well. Here is a discussion about it.
It really gave me a new perspective on diagnostic testing.
https://www.econtalk.org/vinay-prasad-on-cancer-screening/
I don't understand how we don't consider benign cancers false positives? Acting against those cause serious damage to the body for no gain