This is (was?) the most accurate crowdsourced medical site I know of.
Only providers, i.e., ER doctors, may edit, and it only targets providers and residents.
ER doctors focus on quickly segregating acute issues into urgent, critical, and all the rest, making this a great site for differential diagnosis from symptoms/presentation and initial labs. Because ER docs could be dealing with anything, they need quick and brief references. So it's not for pathology or mechanisms, explanations in any depth, and not tailored for patients.
It's almost sad to see such a carefully tended garden, fruit of so much effort, and so open in the spirit of the early internet. It will be soon be consumed by AI and then displaced. The incentive to write or edit might completely disappear when the information is available primarily via regurgitation.
> The incentive to write or edit might completely disappear when the information is available primarily via regurgitation.
This is a strange non sequitur. The incentive to write and edit has always been the same: you want to share information you found with the world. Information availability has nothing to do with editing incentives.
I don't even see how this has anything to do with AI. The doomerism is getting quite annoying.
Given the quality of even bleeding-edge AI it will take a several epoch-making quantum leaps in exponential improvement before I trust it over a dart thrown at cards randomly placed on a dartboard when it comes to medical matters.
I have it from the horse's mouth that some doctors are already experimenting with it regularly (GPT-4, not ChatGPT). It's not always right, but medicine is about memorizing unreasonable amounts of information, and the bot sometimes does catch less common things that multiple humans have missed.
Of course the difference between regular users and doctors is they can tell if GPT is confabulating or giving explanations that make no sense. I wouldn't recommend anyone go and blindly trust what comes out of the pile of linear algebra, it's not quite there yet.
Wow, I remember when WikEM was just getting started (2009 it looks like). I was in medical school trying to teach people how to use the iPad, which was also brand new, and sharing (self-made) flashcards for SRS software via Dropbox.
Surprised to see that they're now a 501C3, seems like they've grown a lot in the interim. Funny that I went into EM but haven't spent much time there, I'll have to take another look.
Personally I use it more as a reminder than an official reference. So, for example, reminders of less common diagnoses or less common presentations of common diagnoses that I’ve learned and know but may not remember the details of but will come back to me with some prodding. If I actually need to dive deep, I usually use UpToDate
> Our android app has currently been removed from the app store. We are working on addressing Google’s concerns so that we can restore the app. In the meantime we encourage everyone on android to use our website.
This is (was?) the most accurate crowdsourced medical site I know of.
Only providers, i.e., ER doctors, may edit, and it only targets providers and residents.
ER doctors focus on quickly segregating acute issues into urgent, critical, and all the rest, making this a great site for differential diagnosis from symptoms/presentation and initial labs. Because ER docs could be dealing with anything, they need quick and brief references. So it's not for pathology or mechanisms, explanations in any depth, and not tailored for patients.
It's almost sad to see such a carefully tended garden, fruit of so much effort, and so open in the spirit of the early internet. It will be soon be consumed by AI and then displaced. The incentive to write or edit might completely disappear when the information is available primarily via regurgitation.
> The incentive to write or edit might completely disappear when the information is available primarily via regurgitation.
This is a strange non sequitur. The incentive to write and edit has always been the same: you want to share information you found with the world. Information availability has nothing to do with editing incentives.
I don't even see how this has anything to do with AI. The doomerism is getting quite annoying.
Given the quality of even bleeding-edge AI it will take a several epoch-making quantum leaps in exponential improvement before I trust it over a dart thrown at cards randomly placed on a dartboard when it comes to medical matters.
I have it from the horse's mouth that some doctors are already experimenting with it regularly (GPT-4, not ChatGPT). It's not always right, but medicine is about memorizing unreasonable amounts of information, and the bot sometimes does catch less common things that multiple humans have missed.
Of course the difference between regular users and doctors is they can tell if GPT is confabulating or giving explanations that make no sense. I wouldn't recommend anyone go and blindly trust what comes out of the pile of linear algebra, it's not quite there yet.
Not sure what the "was" is for, it seems like it is still active, several edits in the past few days: https://wikem.org/wiki/Special:RecentChanges?hidebots=1&hide...
Wow, I remember when WikEM was just getting started (2009 it looks like). I was in medical school trying to teach people how to use the iPad, which was also brand new, and sharing (self-made) flashcards for SRS software via Dropbox.
Surprised to see that they're now a 501C3, seems like they've grown a lot in the interim. Funny that I went into EM but haven't spent much time there, I'll have to take another look.
If someone wanted to volunteer to package that Android app as open source app in F-Droid, that would be great. https://wikem.org/wiki/WikEM:Mobile_Apps
(Another option would be to package the site as HTML5 Offline, but that's a longer discussion.)
Interesting, but it seems to be a bit buggy (and is also exposing PHP errors):
https://wikem.org/w/index.php?title=Special%3ASearch&search=...
Interesting but unclear how to determine the quality of the data and control over edits.
Personally I use it more as a reminder than an official reference. So, for example, reminders of less common diagnoses or less common presentations of common diagnoses that I’ve learned and know but may not remember the details of but will come back to me with some prodding. If I actually need to dive deep, I usually use UpToDate
Funny you should mention that:
> Our android app has currently been removed from the app store. We are working on addressing Google’s concerns so that we can restore the app. In the meantime we encourage everyone on android to use our website.
How is this relevant to what the gp asked?
The quality of the information is such a concern that Google decided to remove the app from their app store.
> The quality of the information is such a concern that Google decided to remove the app from their app store.
Is there any evidence that Google removed the app for quality concerns? Google often removes apps for spurious reasons:
https://news.ycombinator.com/item?id=19429193
https://news.ycombinator.com/item?id=21268389
https://news.ycombinator.com/item?id=24416578
But it is up? https://play.google.com/store/apps/details?id=wikem.chris&hl...
Too bad there isnt an AI to take multiple symptoms and narrow the search a bit.
On it.