Oh my goodness.... so much to respond to I don't know where to start... I don't want you to think I'm being defensive about my field... I'm not. Also, it's a stupid political move as a radiologist to start an interventional spine service if you know many of your referrals are coming from pain docs. The least the rad could do is talk with you and offer you injection slots for patient's you find difficult (super obese, or need for CT guided (pudendal) injections or just a PITA) or those who may need to get in sooner but can't fit in the schedule.
So being a radiologist and a pain doc, I can tell you that I can derive the patient's primary pain generator on a spine MRI (if it's present on the scan) and that correlates better than the clinical exam. I always practiced "treat the patient and not the image." After doing that enough, I figured out it's not always true and, in fact, is probably wrong most of the time. It doesn't stop me from doing a comprehensive exam, though. I also wonder if some forget that once you send the patient to get an MRI... it's not JUST sending the patient to get an MRI... you are asking for a consultation from a radiologist on diagnostics and management. Now some feel it's only for diagnostics, but I can tell you, that's not how the field works. "Clinical correlation advised" I know is a cop out... but most radiologists are fairly good clinicians and understand patient management. I had 8 months of IR plus primary care sports medicine continuity clinic throughout radiology residency. The rads who make injection recommendations are not doing it without some background.
Also, there are very subtle things I've learned to look for that are not taught during radiology residency or neuroradiology fellowship or in pain fellowship (subtle edema on MRI in various locations, subtle cystic changes, findings I was always taught were "in everybody", so let's not get all high and mighty about how the radiologist didn't "see or examine the patient" when in reality, he/she literally saw INSIDE your/their patient and if he/she did due diligence, there was a detailed pain history intake form prior to MRI. I think we all know what an acutely herniated disc looks like in a hot radic. If I could inject those patients in the MRI scanner I would because they are so easy to treat. So, if that's seen on MRI, you can sure as heck bet that I'll be putting in an injection recommendation "without seeing the patient". Same goes for a hot facet (edematous, enhancement). Those aren't clinical decisions, as much as they are easy, quick, pattern based, reflexive decisions that tend to have good outcomes the earlier they are treated. They don't need a full H and P with separate billing codes (driving up healthcare costs) and delayed time to treat in order to fit into the office procedure or ASC schedule. With that said, I wouldn't do it on a chronic spinal stenosis MRI with chronic facet and disc... that's a clinical decision between patient and doc with lots of diagnostic injections, possible RFA and stim.
I realize that you all really, really want those hot radics... but sometimes, earlier is just better for the patient. There's enough to go around and it is about patient care and not always about the wallet. The truth is that most radiology practices actually lose money on doing spinal injections. It's literally chump change compared to other procedures and it's offered to increase referrals to the practice and demonstrate genuine interest in patient management and treatment. Many of those injections are not even reimbursed... think all of those medicaid patients who "aren't eligible" for injections... guess what, a lot of them are done for charity in radiology practices... how you say? Because the fluoro techs are already there, you have a motivated doc who really enjoys the procedure, and the patient is already familiar with the site... the few hundred bucks from the injection is a drop in the bucket compared to a repeat customer who will need follow up MRI and other imaging needs. I can guarantee you that if you told the radiology practice that if they had a hot radic on the table and you could guarantee their patient an injection later that day, they would JUMP at the chance to send the patient to you. But instead of cultivating the referral relationships with your friendly neighborhood radiologists, pain docs tend to focus on the PCPs or neurosurgeons or orthos.
Now, where you SHOULD be worried is when a rad becomes buddy buddy with a neurosurgeon who implants SCS and then passes off chronic pain management to his employed NP or PA. Those are the people trying to take your business. Not the simple epidurals.
Now, as far as the profession itself goes... radiologists take care of their own MUCH better than pain docs do. Competition is intense among pain docs (even competition that is bread WITHIN a single practice). In the area I practiced pain there were docs flat out lying about their credentials in order to get business. Buy ins for partnership are insane... vacation is non-existent and quite frankly those in the field who are most financially successful are outright criminals. I went from two weeks vacation to 7 months vacation (I work nights) and INCREASED my salary by moving from pain back to rads. I don't worry about drug addicts busting through my secretaries threatening to shoot me and I don't have extreme production pressures to the point where I'm entering false codes for procedures *cough* epidurograms *cough* or unnecessary procedures/medications *cough* sedation *cough*. I'm still shocked when I talk to friends of mine in the field about how they were pressured by their "senior" partners to do 3 of every type of injection on every patient or that every patient needs sedation at the ASC owned by the boss. I've never once come across that kind of crap in radiology but yet admin at my pain job was telling me how I wasn't coding "well enough" for my procedures and not using sedation so I was going to get paid less... uh huh and the urine tox schemes and the excessive opioids because "return customers are good business." None of that kind of stress in radiology. There are different types of stressors but not nearly as bad as I experienced in pain.
For the IBM Watson topic above... artificial intelligence is being trained to interpret imaging data. However, it's far from taking over the field. There will always be a need for a physician to interpret the cluster of findings. And as of now, there isn't enough "big data" to train an AI to learn all the diagnoses (especially the zebras) and be reliable. There are interesting discoveries, however, coming from machine learning. For example, there are algorithms now that can predict a genotype for specific brain tumors based on imaging... without any discernable imaging characteristic... no one knows how it's doing it but it ends up correct well over 90% of the time. The best I can hope for is that AI will be able to pick up and describe some of the incidental CYA findings I'm always having to dictate rather than focusing on the major clinical problem for the patient.
TL;DR
Radiologists are mostly cool. Talk to them. They also enjoy treating patients.
Radiologists treat radiologists better than pain docs treat pain docs.
AI is not taking over for radiology; the tech isn't there yet and won't be for a looooooong time.