I've been wondering if face to face time with a physician will becoming a competitive factor in the US oncology market. Patients still value the opinion of a physician over other healthcare providers.
Obviously there is selection bias, but I have a growing number of patients that are unhappy with the very impersonal care offered by the large university network in town. My own family has pointed this out to me on their own (non-oncology visits).
How many times do you have someone thank you profusely for sitting for an hour and explaining their disease? Chat bots will be able to do this accurately, but we are very far away from a chat bot being human "enough".
Everyone is discussing this future, Im not sure patients will want it. They may not have a choice, but just some thoughts.
This is the future.
Putting aside all the other conversations about our job market, and the disclaimer that ANY prediction, by me or ANYONE, for more than 1-2 years into the future is essentially worthless given the pace of change:
I fully expect that, 10 years from now, I will be making minimal medical decisions. Much of the diagnosis and treatment will be handled automatically.
For us, that means software (AI) will, at minimum, extract relevant data from the patient's chart, determine if workup is guideline concordant, tells us what's missing, suggest the orders we need to place to complete the workup (there's case precedent right now that if the orders are placed by the computer itself it's not "medically necessary", but perhaps that changes), once workup is complete we will be recommended the best guideline concordant treatment.
Then, at sim, the software (AI) will delineate targets and OARs (approved by the human), then do the treatment planning (approved by the human), QA the plan, etc. The rate limiting step will be how quick the humans click "OK", but this will be done within 2 hours.
On treatment, online imaging will optimize inter- and intrafraction motion (combination of various wavelength sensors, not just x-ray or visible spectrum). Various side effect management recommendations will be made based on ambient microphones detecting human speech and, cross referencing the patient's chart (and medical history), sensor data from the linac, and conversations between patient and staff.
Follow-up strategy will be recommended at the end of treatment.
I would say two things:
1) No young person should go to Dosimetry school. That career is the equivalent of a lamplighter right as the first residential light bulbs were installed.
2) In the near (10-20 years) future, robotics seem unlikely to get to the point to demonstrate open world labor capabilities. Meaning, if an MLC motor dies and needs to be replaced, human hands will do it.
The one thing I don't see changing is an elderly, scared person just diagnosed with cancer needing a real, breathing human to interact with and guide them through the process. With NCCN guidelines, how many "decisions" are we really making now, anyway?
I'm OK with a future where I exist as a Human Experience Navigator. I can translate what the mysterious AI wants us to do. I can hold a hand.
Maybe I'm wrong. But this is the worst our technology is going to be. And we didn't have this stuff even 5 years ago.
Just...don't go to Dosimetry school.
