View Single Post
Old Posted Aug 30, 2019, 3:04 PM
MonctonRad's Avatar
MonctonRad MonctonRad is offline
Wildcats Rule!!
Join Date: Jun 2008
Location: Moncton NB
Posts: 22,866
Originally Posted by yaletown_fella View Post
I bet that call center work, radiology, low level accounting/bookkeeping and low level legal work will be almost totally automated in a few years before the approvals are there for automated trucking.
I think the future of radiologists is still fairly secure (although an interesting discussion could be made over how many radiologists will be necessary).

Image interpretation is only a portion of the work that radiologists do. In addition to reading radiological examinations, we also:

1) - serve as departmental administrators.
2) - supervise technologists
3) - evaluate requisitions for imaging examinations to ensure that the test ordered is actually indicated
4) - serve as consultants to other physicians if they have any questions regarding a report, or want our advice
5) - perform imaging guided procedures (sometimes as sophisticated as surgical procedures, but performed with needles and catheters rather than scalpels).
6) - attend various multidisciplinary case management rounds at the hospital
7) - teach
8) - patient advocacy
9) - patient interaction (yes, I occasionally talk to a patient regarding the results of their imaging study, and give them bad news).
10) - involvement in imaging equipment purchasing decisions.

These are the things we do that I can think of right off the top of my head. There are probably others.

Even with image interpretation, Watson can be quite sophisticated in detecting findings on the images, but is a lot less sophisticated is determining the significance of these findings. We already use CAD (computer assisted diagnosis) in areas like mammography, but in general, all the CAD program does is flag the finding for your attention. It is up to the radiologist to determine if the finding is significant or not. Many of the findings the CAD program observes are inconsequential, and are ignored by the radiologist in the final report. This is where experience and human judgement come in to play. If we left the whole process up to Watson, we would be performing 10x the number of biopsies as we are doing now.

Finally, once a diagnosis is made (by the radiologist), recommendations are made regarding management and follow up. Should the imaging finding just be followed to see what happens or is treatment or biopsy necessary? What additional imaging modalities are necessary to narrow down the diagnosis? If a biopsy is necessary, how should this be accomplished and by whom?

Many people think that it is the attending physician who makes these decisions. Let me assure you that while the attending is the responsible physician, and makes the final decision, they usually are simply following the advice of the friendly neighbourhood radiologist.

And one last note. The radiologist, when he interprets the imaging study, is the person responsible for determining the acuity of the patient's problem, and whether or not a stat report is necessary. In other words, does the patient's problem constitute a medical emergency? How should the attending physician receive the results of the examination? Is a simple mailed report sufficient, or should the attending recieve a stat fax (with or without a receipt acknowledgement). If the patient has an unanticipated acute life threatening emergency, then we call and talk directly to the attending, while at the same time sending the patient to the ER for management.

Can Watson do all this??????
Go 'Cats Go
Reply With Quote