Mohtalim

Science and Medicine

http://blog.wired.com/wiredscience/2008 ... -medi.html

The appalling success of clinical diagnosis makes me wonder why machines like the one in the article that automate common lab tests aren't in common use all over the place. I suspect it is because the business model for an M.D. does not depend upon a patient getting better, just not dying.

On the diagnosis side, a doctor's main job is figuring out which tests to order (this is not true of all specialties, but describes internal medicine pretty well). This machine just makes things a lot more convenient in that it makes the testing available in office instead of having to send out for it to be done.

Medicine is generally pretty formulaic. Someone comes in with a certain pattern of complaints and you draw blood or whatever and send for whatever tests you need to establish a diagnosis. In the meantime you start treatment "empirically" which means you start giving drugs that treat the most common (and/or most serious) causes (yes, plural) of whatever the patient came in with. Then when the tests results come back you adjust your treatment as needed.

Eventually this can be done by robots.

And so I'm going into psychiatry, where I will eventually forget 95-99% of my 4 years of medical education. And it'll take longer for them to come up with robots capable of providing psychotherapy.