Hacker News Reader

Comment by adamredwoods | Hacker News Reader

We should be able to test our blood for PFAS on a regular basis and our doctors should be able to let us know if we need to filter our blood.

But PFAS testing is not the norm?

Per a Tedx talk on this subject: https://www.nrdc.org/experts/anna-reade/pfas-blood-tests-nee...

>> Unfortunately, a PFAS blood test is not a routine lab test that can be processed by your local lab and covered by your health insurance. In fact, there are only a handful of labs certified to test for PFAS in blood and the out of pocket cost for this test is out of reach for most people.

View on HN

Nobody knows the threshold above which intervening is a good idea or worth the effort, so universal testing seems like jumping the gun a little.

People said similar things about lead...

We don't preemptively test everyone for lead either!

This is true, although I would prefer to still test everyone for plausibly important things to feed data analysts who might eventually discover something useful.

I'd love for that to be true, but there's a good chance that in doing broad tests you are incurring costs (both to patient health and cost [whether paid by patient, gov, or insurance company]) without getting anything beyond noisy data without useful signal

That's a lot of blood draws to ask of people, for something that doesn't even have a hypothesis.

There are studies like the UK biobank that are trying to collect longitudinal information about a huge number of people, but they're volunteers.

Related PSA: If you do a lot of target shooting, especially with indoor ranges, it's worth asking your doctor to occasionally test your lead exposure.

The people vested in using lead in fuel said so out of profit motive, but it seems lead was widely known to be harmful for hundreads of years [0] before it was using it fuel.

[0] https://www.youtube.com/watch?v=IV3dnLzthDA

We should all get annual blood tests so we can track variance from out baselines.

Probably not, honestly. For most tests, the risks associated with a false positive outweigh the benefits of marginally earlier detection.

Humans aren't precise machines, each person has different values and different value ranges for standard measurements - and asymptomatic deviations frequently mean nothing whatsoever. On the other hand investigations can be invasive, painful, risky and expensive.

There's a reason most tests aren't recommended for folks outside certain risk categories (age, family history, etc). And in a litigious society like America, it's not because it's expensive. There's some specific diseases where there's a lot of data backing this thesis - certain cancers for instance.

> For most tests, the risks associated with a false positive outweigh the benefits of marginally earlier detection.

If these tests aren't run, then I am not sure if it would be "marginally earlier", it might be much later or even never that the problem is discovered.

If they are run, the results could be uninformative (https://en.wikipedia.org/wiki/Base_rate_fallacy). Too many false positives could lead to risky treatment and be dangerous. Testing each positive again would still create false positives. I don't know the numbers in this case, but I wouldn't be surprised if accurate mass testing were impractically expensive.

This should only be a problem if you don't adjust the models. If you run a huge number of tests, you can create far more detailed models and thus reduce the false positive rate.

Ideally, all chemical processes in the body could be known and we could have an almost "mechanical" model of the body. How else but with more testing can be reach that knowledge?

> This should only be a problem if you don't adjust the models. If you run a huge number of tests, you can create far more detailed models and thus reduce the false positive rate.

Can you? The problem is the test's sensitivity and specificity, there's not much statistical work to do after the test on a single case. It's not an advanced model deciding whether you're sick or not.

> Ideally, all chemical processes in the body could be known and we could have an almost "mechanical" model of the body. How else but with more testing can be reach that knowledge?

Yeah, ideally we would. Not a doctor, but as far as I can tell, we're a thousand years away from that.

The video in this comment provides some good examples for why the risks associated with a false positive can be pretty significant: https://news.ycombinator.com/item?id=31126368

That's the whole point of establishing a baseline.

I don't know why this comment is downvoted. Let me elaborate a little on what I think they meant.

All the risks associated with false positives (other than the side effects of actually drawing blood once a year, which I think is negligible) can be removed by simply not overreacting. Find a cancer? Well, consider not subjecting yourself to chemo just to get rid of it. Some cancers will kill you by winter, some will just stay there until you die of something else. How will we know which is which? By testing lots of people over long periods of time to see what's harmless and what's actually really bad. Or as the previous commenter said: "That's the whole point of establishing a baseline."

The parents premise is wrong. You can't just "establish a better baseline" and "if the test says you might have cancer don't worry about it maybe?" The former is hampered by base rate fallacy [1] (specifically the part about disease in a low-incidence population) and the in the latter, if you're going to disregard the outcome why are you testing?

A concrete example is the NHS recommends against the PSA prostate cancer antigen screening test completely for everyone of any age because this test does more harm than good. They have a good write-up on why. [2]

Generally I agree this data should be collected, but aggregated and not used for individual treatments.

[1] https://en.wikipedia.org/wiki/Base_rate_fallacy

[2] https://www.nhs.uk/conditions/prostate-cancer/psa-testing/

I guess I should have been more clear that by establishing a baseline I did not mean take one test and then have surgery.

I think the value in regular blood tests is to get an idea of what normal values are for you specifically. For some diseases/conditions the blood levels have such a big range that it is more useful to know the delta.

We hardly ever measure the “good” or stable cases so when we start testing stuff when we are ill there is very little data on what is normal or even normal for you.

Ive heard of cars where due to family history someone preemptively screened on certain blood values. When those changed they knew something was up, even though the value was still within normal range it was up 100x for them.

Here's a good explanation of why it's not a good idea from Medlife Crisis. [1] h/t wirelesspotat.

[1] https://www.youtube.com/watch?v=rW3DGnHO2iY

How well do we actually know "normal" values if we don't have the broad data that regular blood tests for everyone could give us?

There's a great video from Medlife Crisis about why this might not be a good idea

"The Problem With Silicon Valley Medicine": https://www.youtube.com/watch?v=rW3DGnHO2iY

Even is the US, a metabolic panel of about 15-20 basic blood tests is only about $40. You an order it online and go to a local lab to get blood drawn.

What do they cost?

If you were curious, a quick search gave me results in the range of $300-$400

https://empowerdxlab.com/products/product/pfas-exposure-test

I just called Quest Diagnostics and they charge $679 w/out insurance for the PFAS panel [0].

[0]: https://testdirectory.questdiagnostics.com/test/test-detail/...

Reply on news.ycombinator.com