could somebody rewrite this blog post about algorithmic doctors, citing some papers?

- Edit
- Delete
- Tags
- Autopost
could somebody rewrite this blog post about algorithmic doctors, citing some papers?

there's a post on tech-crunch here about how machines and algorithms will be able to take over from doctors in diagnostics and treatments.
http://tcrn.ch/yz3V9c
i think it's a nice idea, but i'd like to see the evidence. as it happens, there's a large literature on decision support tools, and shared decision making, and all that. stretching that data to the extreme, going close to the hopes of this fantasy piece, i'd be interested to know, where you would get?
essentially, if someone could rewrite that tech-crunch piece, but citing relevant papers where people have tried to do it, and reported results, that would be a really interesting read (i think, for a lot of people). because this idea comes up a lot, and it would be great if it worked.
my personal view on the future of medicine is that a lot more of it will revolve around doctors being experts in evidence search, and risk communication: there are a zillion academic papers published every year, nobody can keep up with it all, or store it all in their head, and things change over a decade of post-medical-school practice, let alone a life-time (also: prices change, and efficiency in healthcare matters more than ever). the future is finding info, communicating risks and benefits to patients, discussing priorities, and making decisions together that optimise the patients stated goals.
ho hum. but i'd love a "state of the field", and an informed sense of how close it could get to the tech-crunch ideal.

EDIT

if you want somewhere to start, btw, there will be some interesting stuff in this book, from a brainstorming meeting i was at a year or two ago (chapter from me in there too), on shared decision making etc. but it's expensive, and i don;t know if the written chapters are as good as the sessions were, as ever with mtg reports!

http://amzn.to/xYkYSf
- 18 January 2012
- Views
- Permalink
- Tweet
- 6 responses
- Like
- Comment
4 months ago Arin Basu responded:

Ben ,
There are some serious mis-assumptions in this article that I am afraid Techcrunch editors never examined or were too afraid to disute. Khosla (the author) writes,
" The vital signs could all be determined with the help of mobile devices, the operation of which do not require years of training and a certification." He does not seem to understand that the years of medical education and training is not about "how to put a stethoscope" but precisely "where" to pu that stethoscope and what to do "next". I am not sure anyone has quite figured out an algorithm or an artificial intelligence based solution to do _that_ for the myriad of physical complaints (including fake complaints) people present with.
I think what you write in your post needs to seriously done: a good evidence based review written from a joint perspecive of artificial intelligence and medicine as is practiced. In our enthusiasm for predicting the future and moving ahead particularly in medicine, we often forget that there is a place and reason for physicians in the system, often deus ex machina do not work. Just don't.

4 months ago mrstrellis responded:

Your posts would be a lot easier to read if you followed the convention of starting each new sentence, and the "i" in "I'd", with a capital letter. Can't see why you don't :( .

4 months ago anonymous-coward responded:

The tech-crunch post seems to completely miss the point, exemplified by the growth or alternative medicine, that a large amount of people visit a doctor for reassurance.
People seem to fall largely into two groups, people like me who only go to the doctor when the smell of gangrene is so overpowering that the next door neighbour starts to shun me, and people like my partner who goes to the doctor when she or one of the kids has the slightest hint of a cough.
Both of us are quite happy with the actual medical treatment available, but that's not actually what she's after.

4 months ago Adolfo Neto liked this post.

4 months ago jamougha responded:

Anyone who writes software knows that it's impossible to idiotproof a simple smartphone app, the idea of doing it for a medical diagnostic program, with all the legal issues that entails, sounds laughable to me.
I expect to see medical diagnostic AIs that outperform humans -- as he says, it's just a matter of enough data, and the NHS records would be ideal -- but I believe they'll need expert clinicians to act as an interface.

4 months ago colaboramos (Twitter) responded:

Hi there, is there any article or post (here or in www.badscience.net, or any other scientific blog) related to blood eletrification, which claims to "100% cure AIDS, cancer and all of the other known diseases to humans, including some which we have not tested like ebola"?
http://www.healingcancernaturally.com/energetics5.html
I need it to help someone I know who is making the wrong decisions.
Thanks
Ben Goldacre's Space

I like stats. I'm a doctor. I write about dodgy scientific claims in my spare time.
This isn't my main blog. Find me here:
Blog:
www.badscience.net/
Book:
http://www.amazon.co.uk/Bad-Science-Ben-Goldacre/dp/000728487X/?tag=bs0b-21
Tweet:
www.twitter.com/bengoldacre
TED talk:
http://www.ted.com/talks/ben_goldacre_battling_bad_science.html

bengoldacre - secondary blog

could somebody rewrite this blog post about algorithmic doctors, citing some papers?

Ben Goldacre's Space

About Ben Goldacre

Archive

Follow Me