ASK THE DOCTOR: Dr. Google
If you have taken part in medical care in the last year, it is likely you have encountered artificial intelligence (AI). It is increasingly used to write medical provider notes, sometimes by directly listening in on visits.
It is also widely available to lay people and doctors alike to answer questions about symptoms and disease management. You can ask AI to summarize a visit and your recommendations in normal language for a patient or family member to take with them after a visit. If you have had surgery, you might be receiving text messages from a bot checking in on you to see if you have any questions. Even as I write this article (the old-fashioned way), an AI bot offers to rephrase things for me.
Having been a doctor for almost 30 years, I find myself a little unsure about the reliability of the technology. Studies do show AI diagnostics perform well, and perhaps better than a human brain ever could. I do not want to discount the value of having assistance when the world of medical diagnostics and treatments is changing so rapidly. But what of the ethics, the nuance and the context?
One of the concerns people have is that AI can’t actually “care.” It can quickly assimilate a large amount of data to narrow down the possibilities of diagnoses and treatments, but it does not care what the person’s life has been like and what is most important to them. It relies on the data coming in to be correct and might miss something not making sense for the individual person being robotically assessed. Granted, human doctors also can sometimes be robotic and uncaring. Also, humans can be frustrated by things while AI doesn’t take anything personally. At least, not yet.
Stacy Abrams wrote a thriller called “Coded Justice” that looks at the good and potential evil of technology in medicine. Reading it might make you anxious about using a smart phone for a while, not unlike how the movie “Jaws” made many of us leery of swimming pools, much less the open ocean. For Abrams, the good is technology’s potential for improving and giving more consistent care to even the most vulnerable people. The bad is technology overriding physicians’ more nuanced ideas to the detriment of the same vulnerable patients.
At an annual national internal medicine meeting this year, I heard a lecture by an ethicist about AI. What stuck with me was the reminder to “use the tools, not be used by them.” Many healthcare workers spend hours a day on the computer, charting and documenting all the things required. The term for a tired physician who does medical charts at home after the kids are in bed is “pajama charting.” If AI can reduce the hours per day spent clicking computer keys, the benefit could be seeing more patients in a less rushed manner that leaves little time for relationships, proper physical exams and nuanced care.
When AI is used to write a patient note, the physician or PA or NP or nurse or MA still has to edit to make sure the occasional weird things thrown in by AI are fixed. This was also an issue in the days when we dictated our notes and someone had to listen to our taped comments and transcribe them. Once I was asked to PLEASE not have Beethoven playing in the background while dictating. Once I read a dictation from a surgeon on a mutual human we were caring for that described patient’s “Bologna amputation.” Below knee. All of which is to say humans can make odd mistakes as well.
“Dr. Google” has been a phenomenon for years, which is when a patient comes to an appointment with the internet’s ideas about what is going on with them. AI has significantly improved the online information, though it often is still out of context and may not apply to that person at all. How can a doctor know this? Hopefully by knowing the person and having seen lots of similar “cases” and understanding the nuances of our amazing physiology.
I like it when people have questions and find information. This is a place to open real discussions and sometimes Dr. Google has a great idea. There is a place for technology in medicine. I just hope the use of it is not considered a way to bypass actual healthcare workers. Humans have a special ability to care about people who are affected by the science, and to respond with compassion as well as expertise.
The U.S. has some of the worst health outcomes statistics of any comparably wealthy nation in the world. This is not for lack of expertise or technology, but rather the predictable outcome of a system geared toward a business model that has fragmented health care. Now AI can compute answers much faster than any human, but we still need to make sure those answers make sense for the person in front of us. Health care is not improved by further isolating people from each other. And AI has the potential to stifle and undervalue the creativity and compassion of physicians and other healthcare workers who have spent years training and studying to do their work.
Let’s find a way to use the tool of AI without forgetting that the real heart and excellence of medical care come from a thoughtful and caring approach, with human contact and relationships based on trust and mutual respect.
Dr. Jennifer Heidmann is chief medical officer at Redwood Coast PACE (707-443-9747) at the Humboldt Senior Resource Center. This column should not be taken as medical advice. Ask your medical provider if you have health questions. Send comments to SN@humsenior.org.
