To Automate or Not to Automate: Risks and Rewards

10/3/2021

Lots of Oppurtunity

Physicians just don't have the resources to meet the business requirements to deliver healthcare and deliver the best healthcare to their patients. As Dr. Z explains in the video over there (points), these folks who have spent the better part of their adults lives training and preparing to provide care to others simply cannot. They can't because they are in a system that won't allow them to. So what can we do if we can't chante the system?

I absolutely love this video by Dr. Z

Robots to the Rescue?

It goes without saying that physicians would love some assistance on tasks, specifically administrative tasks. They didn't sign up for this, as noted above, and they certainly would prefer to not spend half their day or more on such things.
This presents a very interesting opportunity for applications of recent advances in AI and machine learning. Since the actions and thought that go into these undesirable tasks can be thought of as faily straight forward - certainly more straightforward than performing procedures or interacting with patients - it may be a ripe area for improvement by automation.
In her book, An American Sickness, Elisabeth Rosenthal describes in depth the many ways the American healthcare system is not working for its patients. A system ripe with inefficiencies, fraud, malpractice swpept under the rug, and penny-pinching amounting to millions or billions in fees, it makes me wonder if radical disruption by automation could be an answer to woes named here.

With administrative tasks optimized, clinicians would have more time to practice their craft, to honor their passion for healing people. It's easy to see how this could alleviate lots of problems in healthcare today.

With that said...

We need to be careful here. When it comes to disrupting healthcare, it's literally life or death.
While automation is exciting, and open doors to shift efforts from the menial to the meaningful, it also opens the doors to unchecked disaster. A good example of this, perpetrated by one of the (if no THE) largest electronic health records in the nation (or in the world?) showed us just how wrong AI can get it.
I think the lesson we can take from this is: let's let the experts do what they spent their lives training to and passionately wanting to do: diagnosing and treating illness. Let's unleash AI elsewhere, at least to begin with.