One to two times a week I go through my junk mail folder. Usually it’s a collection of, well, junk: ads for CME, office software, car warranties, gift cards, dating sites, eyeglass or razor sellers, etc.
But there are usually a few items I’m glad I found, ones that I’m not sure how they ended up there. Bank notifications, package-tracking updates, a few other things. By the same token, every day a few pieces of junk land in my inbox.
Dr. Allan M. Block
This is, however, what we do for a living in this job.
Some patients are straightforward. The story is clear, the plan obvious.
Some require a bit more thinking.
And some are all over the place. Histories that wander everywhere, a million symptoms and clues. Most are likely red herrings, but which ones isn’t immediately obvious. And it’s up to the doctor to work this out.
With my junk folder, though, it’s usually immediately obvious what the useless things are compared with those of value. In medicine it’s often not so simple. You have to be careful what you discard, and you always need to be ready to change your mind and backtrack.
Artificial intelligence gets better every year but still makes plenty of mistakes. In sorting email my computer has to work out the signal-to-noise ratio of incoming items and decide which ones mean something. If my junk folder is any indication, it still has a ways to go.
This isn’t to say I’m infallible. I’m not. Unlike the algorithms my email program uses, there are no definite rules in medical cases. Picking through the clues is something that comes with training, experience, and a bit of luck. When I realize I’m going in the wrong direction I have to step back and rethink it all.
A lot of chart systems try to incorporate algorithms into medical decision-making. Sometimes they’re helpful, such as pointing out a drug interaction I wasn’t aware of. At other times they’re not, telling me I shouldn’t be ordering a test because such-and-such criteria haven’t been met. The trouble is these algorithms are written to apply to all cases, even though every patient is different. Sometimes the best we can go on is what I call “spidey sense” – realizing that there’s more than meets the eye here. In 24 years it’s served me well, far better than any computer algorithm has.
People talk about a natural fear of being replaced by computers. I agree that there are some things they’re very good at, and they keep getting better. But medicine isn’t a one-size-fits-all field. And the consequences are a lot higher than those from my bank statement being overlooked for a few days.
Dr. Block has a solo neurology practice in Scottsdale, Ariz.