Not at all. I'm very good at my job, and I worked very hard to get where I am. Sick people should be happy that I'm reading their CT scans. If the pay wasn't good I would have chosen a different field, and somebody else would be doing my job, less well.
For reference, before medical school I worked in IT, and at the age of 23 ( 15 years ago ) I was making six figures. I left that to go back to school, partly because I knew I would be financially rewarded.
Again, see my response below. I don't order the studies I read, and I don't set the reimbursement rates. Additionally, it is Medicare fraud for me to read a study and not charge for it. In short, I have very little say over what I earn.
Do you worry about being replaced by computer software? You said you make $800k a year, and for that amount of money you could higher 2-3 PhDs in machine learning and computer vision for each radiologist. Do that maybe 5 times over and give them 5-10 years and productivity and costs for CT scan analysis will probably go way down.
After all a radiologist armed with software that prescans each CT scan looking for interesting areas could work way faster then otherwise.
There's a small army of PhDs doing this right now, NIH-funded and in commercial research shops.
A radiologist still must look at every single slice for liability reasons for the foreseeable future. That's the real time cost.
I work for a neurosurgeon in image-guided surgical planning research. One of the challenges is segmentation (labeling) of target areas to use in navigation. (radiologists generally don't do this, for various reasons). I've used some of the best commercial software, and seen some of the top research algorithms. With these, for the `easiest` tumors, we still have to semi-manually choose the region on every 3rd or 4th slice. The best algorithms will interpolate the other slices based on essentially fitting along a levelset. For a typical tumor, it can take 20-40 minutes to do this task - using the best available software!
This is `not` radiology, it's image labeling. It's orders of magnitude simpler than radiology.
There are some promising techniques to, for example, automate detection of changes in volume of some radiographically questionable area (after manual labeling for the first scan). At best, this will add information with no extra time cost.
I can assure that there are already an awful lot of PhDs in machine learning and computer vision working on the problem of automatically analyzing medical images, and there's been lots of progress over the last few years. There's a really long way to go, however- it's one of those areas where the problem is a lot harder than it seems.
Working in the field has convinced me that radiologists earn every penny, and that there are good reasons for the lengthy residencies and stringent board exams.
> it's one of those areas where the problem is a lot harder than it seems.
Yes, exactly. To computer savvy people unfamiliar with radiology it looks like something a computer might be good at, but I suspect the best we can hope for is a computer to aid me in my work, not replace me.
I've sort of answered this problem above. I happen to think the problem is not solvable any time soon. If you or someone you know would like to prove me wrong, I will invest in your venture.
Does it bother you that you are profiting handsomely from sick people?