Comment by theptip
2 days ago
> The theory here is largely mechanical
I’ve long felt that the reliance on population-statistics (RCT) rather than individual diagnosis highlights how little we really know about medicine.
A mechanic wouldn’t try to fix a car based on a checklist of symptoms interventions that work X% of the time across the population of cars; they would actually inspect the pieces and try to positively identify e.g. a worn/broken component. Of course, this is harder in the human body.
I’m hopeful that as diagnostics become cheaper and more democratized (eg you can now get an ultrasound to plug into your iPhone for ~$1k), we’ll be able to make “medicine 3.0” I.e. truly personalized medicine, available as standard rather than a luxury available to the 0.1%.
This kind of thinking, that everything can be broken down into pieces and studied in isolation... it has brought us very far, but it has some hard limitations. Especially in biology, where any leak you find may serve some function.
In medicine, the primary goal is to help, rather than understanding why something works exactly. Sure, understanding is an important goal too, it's just much harder to achieve than being able to help. And less important than knowing that your treatment will work, without any major side-effects, for the kind of patient in front of you.
Atomization vs gestalt is mostly orthogonal to the point here.
Taking an extreme example to illustrate the point, if you could image every atom in the body you’d have the opportunity to understand the whole-system dynamics, or try to isolate subsystems if that helps.
> In medicine, the primary goal is to help, rather than understanding why something works exactly.
Right. My point is that each phenotype is different, and so understanding the individual will allow you to help them more.
As the joke goes, doctors/surgeons have to keep the "engine" running the whole time they are trying to fix it. I'm not sure how useful the mechanic analogy is.
My brother is a mechanic and he definitely goes on technet to review a checklist when doing diagnostics. Especially on cars he doesn't see a ton
I don't know about the name "technet" in particular but there are services that aggregate technical service bulletins released by car manufacturers to make them more easily available to independent (non-dealer) mechanics.
It's more than just service bulletins. It aggregates service bulletins as well as common failures, steps to diagnose, and common repair procedures. Most of the data is via 3rd party repair centers as dealers quit sharing this data ages ago. Shops across the country use it to generate estimates and pay their techs. This ensures all work gets logged and that diags and standard procedures improve over time.
What is Technet? I've googled it, but there are just a bunch of generic sites.
I don't know if that's an official term, I think it might refer to a legacy platform that doesn't exist any longer. It's a software package that aggregates common failures and repair procedures across 3rd party repair shops. They use it to do estimates and pay their techs as the techs are on contract and paid per the job.
Yeah, how can this take possibly be on HN when this approach works on software. We even manufacture the population samples ourselves (fuzzing).
What makes professional mechanics fast (and therefore makes them good profit) is knowing from experience when you can shoot from the hip like that.
But yeah, you won't find people doing things that way in any setting where it Actually Matters(TM) (e.g. expensive things where you really can't justify not fixing it right the first time)
Nitpicking on the mechanic point, but this is pretty common, just not at the same level of detail as medicine. Certain brands, models, and parts are more likely to fail in certain ways, so if a model comes in with symptoms of a known, high frequency problem, many times that work will be done first rather than taking more of the car apart to inspect individual parts.
Certainly I didn't think there's huge bodies of work on those statistics the same way there is for medicine, but any car repair forum online will give you some sense of this
Please, do not compare an engineer working on something built by humans with publicly available documentation, and a scientist working on something built by nature without event a single hint of documentation, nor assurance of any logic behind.
They work in different contexts, with wildly different constraints and wildly different expected outcomes.
I worked at an auto parts retailer growing up in high school. I had very basic knowledge of cars. After working there for a year, someone could come in and describe what they were seeing and hearing and 80% of the time I could tell them what is wrong and what is needed. You'd be surprised at how a large population is similar for different 'paths' through the problem and solution space
Diagnostics only really tells you that something is happening but not why; and it is so impractical if not impossible to observe mechanisms of action for much of the body.
Heck, we don’t even know the mechanism of action for acetominophen, and that’s one of the most popular, oldest OTC drugs out there.
The other difference, of course, is that the mechanic can just RTFM
This sounds a lot like the pitch for Theranos
Which was a great pitch because it’s what people want. It just has to be based in reality.
So?