Pocket Doctor
Siri: “Hey Jill, how are you? Anything you would like to talk about?”
Jill: “I fell while hiking today.”
Siri: “Were you hurt?”
Jill: “Left ankle is somewhat swollen and sore from jamming the foot into a rock while trying to stay upright.”
Siri: “Sorry to hear that, Jill. Any idea why you fell?”
Jill: “I don’t know, just fell.”
Siri: “Did you feel dizzy?”
Jill: “No.”
Siri: “Did you trip?”
Jill: “I guess so.”
Siri: “Why do you think that, Jill?”
Jill: “I think my left foot may have caught on the log that I was stepping over.”
Siri: “Were you distracted?”
Jill: “No, not really.”
Siri: “I think you should have your podiatrist take a look at your left ankle; especially the tibialis anterior tendon.”
In addition to always being available to ask the right questions and give answers to your medical questions, your pocket doctor, in conjunction with soon-to-be-announced heart and brain monitors, can provide around-the-clock monitoring and automatically transmit medical alerts to doctors and dear ones in case of falls, heart attacks, blackouts, strokes,…
Artificial Intelligence (AI) is here. There is a lot to be excited about. Healthcare will be one of its biggest beneficiaries. Soon, everyone can/should have a pocket doctor. Soon, everyone will be able to access affordable psychological screening.
A lot of conversation is around the mistakes that AI makes. Meanwhile, AI reads certain types of radiological images more accurately than radiologists.
In your example, Siri is patient, persistent and thorough. Overworked doctors can rush through the patient interview process and overlook diagnostic procedures that AI will consistently flag. For the time being, AI/physician collaboration looks like a good idea.
Joel:
What it knows based upon an established foundation of fact and reality is not a concern. What it learns from those conversations is more of a concern as each of us is unique.
And then the application of the latter?
@Bill,
AI can and does already adjust conversation based on answers. Meantime, AI will continue to follow up and won’t be distracted by the clock or hunger or exhaustion.
Of course, the application of Siri’s information is up to a healthcare professional, just like current AI reading of radiology data, while more accurate than a trained radiologist on average, is still just a guide to action for a healthcare professional. The goal of AI in medicine is to aid more accurate decisions by healthcare professionals, not to make those decisions unilaterally.
“still just a guide to action for a healthcare professional”
Administered healthcare today is driven by bracketing of patients into categories. I find myself as such. Assertions are made and not questions of present-day health. It can catch you by surprise and your nonanswer results in another checkmark of observation and perhaps medicine to take care of a patient.
Coding is the norm to extract payments from insurance, Medicare, etc. I finally got my doctor to reduce my intake of Metoprolol. I have my own uncertified blood pressure machine which equals their uncertified equipment. If my blood pressure was any lower 110/65 at times they would need to do other experiments on me to discover why.
Lipitor at 10mg is another one. When I had open heart surgery, my cholesterol was at 114. Can’t blame that one for sure. It had been lower for years from the time they asked me to lower it decades earlier. It is always low because of my exercise and eating habits (oats, oat bran, etc.). The issue for me is more likely stress.
AI will more than likely be used as a tool for directing or reinforcing (hopefully) a healthcare decision. I suspect the former as opposed to the later. Costs, how much time per patient, and profits resulting from coding of patients could moreso result from AI. It will be easier to acquiesce to a machine’s determinations rather than a human. Less time spent and potentially more profit.
And yes, I am seeing the negative to the use of AI. (Sigh, keep calling it A1). Shold have never made that mistake.
@Bill,
What y0u are describing is not medical decision-making, it’s insurance administration. Perhaps AI will also have a role in that. I don’t know. I was just commenting on the role of AI in patient care and diagnostics.
There is already robotic surgery, and I expect AI will have a role in that. Machines do your blood analysis already, and AI may improve on analysis of those data. As I mentioned, AI is superior to human judgement in reading certain imaging data.
Since you cited personal anecdotes, I’ll cite mine. I went to see my PCP back in 2017 about having to pee every 30-60 minutes around the clock. He asked me some questions and became fixated on the mention of having passed a couple of tiny kidney stones a few months earlier. Meanwhile, he didn’t palpate my bladder or use a hand-held ultrasound, which would have disclosed >1 liter of urine in my bladder, secondary to prostate enlargement. Fortunately, he did send me to a urology clinic, which made the correct diagnosis. I’m confident that AI would have caught it in my PCP’s office, if he’d been using it, and would have saved me a couple additional weeks of misery.
It isn’t that they can’t see the solution.
It’s that they can’t see the problem.
~ G. K. Chesterton