As for computer programs doing radiological diagnosis -- which ones are you referring to? The only ones I know of are mammography programs and they have not been very successful -- certainly they have not lived up to the expectations people had for them.
It's getting close in virtual colonoscopy.
http://www.thieme-connect.com/ejour...ssionid=F808B51464A26DFDA91E9A38D80081FA.jvm2
I've considered the issue a lot, and unfortuantely I don't have a lot of answers, but I suspect for a lot of screening tests that are looking for a singular type of well-defined pathology like mammograms or VC, computers might be able to outperform humans very soon. How this will translate into medico-legal issues surrounding practice, I don't know, but I certainly don't want to be counting on legal barriers to be protecting my career against patient interests or economic forces.
That said, these issues don't scare me off radiology. It will be a much, much longer time until a computer can be as effective as a human radiologist for reviewing scans that don't fit into neat little boxes. I have a little bit of background in neural-network simulation and even image processing via neural-networks, and they can do some amazing things, but they've got nothing on the flexibility and reasoning that a human can provide, and I suspect they never will.
If computers do manage to have that capability in my lifetime, it still doesn't scare me off of rads because every other field will be similarly affected, perhaps even earlier. Let's take endocrinology for example, a field that's pretty far away from rads. What if you could provide really precise complete mass-spec data about every single compound in someone's blood, fit that in with medical history and demographics, and you could probably have some sort of computer that would spit out the most likely pathology in endocrinology.
Basically, when computers and/or robots do finally rule the world, the economy will have altered so dramatically that it's really absolutely impossible to predict what field you should be in and maybe we'll all just lounge around on our vibrating couches watching holographic television in our Jetson's houses by then anyway.
Put another way, I think that diagnostic radiology is just about the most intellectually rigorous undertaking beyond pure science available to me as a career choice except for maybe law (which I don't want to do anyway), and therefore probably the last to get well and truly replaced by computers.
Similarly for teleradiology. It will definitely affect the market, but there's a severe shortage of appropriately trained radiologists period. Right now that works to the American patient's benefit (and American radiologists detriment) because of the income gap because they can afford to suck up radiologist intellectual resources from India, but in 20 years when the average Indian or Chineese person makes $25,000 a year, then it will be added market too. And we can do teleradiology for the Indian market while the Indian radiologists sleep and they can do our prelim-scans while we're fast asleep. Hell, even the Indian radiologists are pulling in $120 K, and that goes a lot, lot further in Bangalore than it does in New York.
My thoughts on it.
