7 ms·
> "Medical school has always been hard. But the amount information one has to intake and regurgitate seems to have steadily increased. Meanwhile, medical school
by chiefalchemist 8y ago
> "Medical school has always been hard. But the amount information one has to intake and regurgitate seems to have steadily increased. Meanwhile, medical school remains four years long..."
It would seem to me the med profession is an ideal place for technology / AI. Docs aren't magicians and House (the TV doc) is fictional.
Furthermore, it would seem to me that the aggregation of patient data / symptoms / outcomes would be more beneficial than the single opinion of the one or two docs that see.
I'm not trying to dismiss the human element or the knowledge / experience of any single doc, but certainly that shouldn't remain so silo'ed.
<sidebar>
My father had a (severe) stroke almost 18 months ago. He's doing well and ultimately got good / great care. That said, my sense was most docs favored their (subjective) optinions over what I presumed would be known best practices, etc. Mind you, my assessment is subjective. None the less it was a pattern that existed across multiple shift and med facilities.
</sidebar>
- soared 8y agoDoctors 100% take advantage of technology, contrary to many comments here. There are search engines that are basically google but for everything medicine related. Your doctor doesn’t know the dosage for most medications, they just look it up on their website.
- jahewson 8y ago> Your doctor doesn’t know the dosage for most medications, they just look it up on their website. And then the local pharmacist corrects it :)
- classichasclass 8y agoMy second day of internship, I could already identify which inpatient pharmacist from the beeper number.
- kendallpark 8y agoYup. UpToDate all day, everyday.
- kendallpark 8y ago> It would seem to me the med profession is an ideal place for technology / AI. Docs aren't magicians and House (the TV doc) is fictional. As someone that works on healthcare AI, you'll have to believe me when I say AI physicians are the jetpack of the 21st century. The populace might think they're right around the corner technologically, but there's a lot of obstacles, both technical and human that make it pretty unrealistic. The need for physicians isn't going away anytime soon. In the same way, the profession of piloting hasn't been threatened by autopilot.
- elliekelly 8y ago> The populace might think they're right around the corner technologically, but there's a lot of obstacles, both technical and human that make it pretty unrealistic. And let's not forget regulatory/legal obstacles. Even if we convince doctors and patients of the benefits of AI in healthcare we still have to teach the old farts in Washington what AI is, what AI isn't, and convince them it isn't dangerous. Anyone who watched the Facebook hearings can see how steep that learning curve will be.
- jackpirate 8y agoIf AI could ever come close to replacing doctors, hospitals and insurance companies will be throwing so much money at congress that they'll get whatever regulations they want passed. The problem is that AI actually has to sort-of-work-ish first.
- rscho 8y agoAI is extremely far from being ready for field hospital work. The back office is hidden from you as a patient, but you have to realize that hospitals are a huge mess. Really, really messy places that remind you that medicine is still in its infancy and that medieval times are not far behind. AI absolutely needs a source of information to be efficient. In fact, this is directly related to why doctors work long hours. The information is so scattered, impredictable and irregular that delegating that to AI today is both physically and financially unthinkable. It will be one day, but not today, and not in the coming years.
- derefr 8y ago> Docs aren't magicians and House (the TV doc) is fictional. House (the TV doc) wasn't, IIRC, a particularly-strong example of an intelligent/clever/omniscient/"magical" doctor. House was: 1. willing to do non-therapeutic experimental procedures for the sake of differential diagnosis. (I.e., he was willing to do small amounts of harm to patients in order to get their bodies to tell him something, where other docs would only ever try procedures in DDx that had some likelihood of being ameliorative in-and-of themselves.) 2. good at synthesizing a true patient history from relatively little—and much of that potentially-false—information. Other than that, he was just a regular doctor (who was a specialist in infectious disease and nephrology) who was good at team-building by hiring experts of complementary specialties, and then focus-firing all that group expertise at the problem at once, rather than doing the standard medical specialist-to-speciaist chart handoff. Really, any doctor could be House, if they set out to do so and practiced the skills involved. The real fiction was a hospital willing to set up the kind of "diagnostic medicine" department that House headed, where you've got five experts working on one patient at a time. > That said, my sense was most docs favored their (subjective) optinions over what I presumed would be known best practices, etc. My impression, from having a few doctor friends, is that "best practice" is generally either old news or p-hacked hooey. There are many therapeutic approaches that very obviously work in clinical practice, even though—according to medical academia—their effect-size is "neligible." (For example: https://slatestarcodex.com/2018/11/07/ssris-an-update/ https://slatestarcodex.com/2018/11/07/ssris-an-update/) Ideally, you'd collect statistics on what doctors believed from informed clinical experience, and then teach other doctors that as clinical best-practice. Right now, though, we just teach them the stuff we've "proven" in academia, and then let the real world slap them in the face later on.
- nradov 8y agoMedicine is still as much an art as a science. Evidence based medicine is expanding and some doctors don't follow known best practices as frequently as they ought to, yet for many areas we simply don't have the large-scale clinical studies necessary to establish best practices. And even when there are established treatment protocols, those often have to be tweaked to fit individual patients. A lot of this is based on tacit knowledge and intuition which can only be gained through experience. AI diagnostic tools could be useful in some limited circumstances. The problem is in gathering the necessary data (especially coded data) to feed into the AI. So it won't necessary save any effort. The need for human radiologists to read images may decline as pattern recognition technology improves.