BLOG: Are we more accepting of human than robotic error?
Click Here to Manage Email Alerts
When complications occur during cataract surgery, there can be many different reasons.
Most often, some frailty of the patient’s tissues does not stand up to the rigors that surgery unavoidably exerts.
Sometimes a patient moves unpredictably, or a poor view through a compromised cornea makes it impossible to protect all the delicate surfaces inside the eye. And sometimes conditions like intraoperative floppy iris syndrome make it impossible for even the most experienced surgeon to achieve a perfect result.
Experience helps. Looking back, I’m certainly a better surgeon today than I was 5, 10, 15 or 20 years ago. Sometimes I take actions in surgery, moving the instruments or adjusting the machine settings, and I’m not consciously aware of why. It’s evident that I’ve experienced so many patterns of tissue moving during surgery and learned how to stay out of trouble somewhat automatically. Any highly experienced surgeon will report the same evolution of technique.
It’s been postulated that AI-driven robotic systems can learn effective surgical technique by “watching” surgical video of both trained and inexperienced surgeons. Learning just like humans, they can observe subtle cues to learn what motions of experienced surgeons yield success and create an algorithm for any imaginable scenario. I want to believe that is true, but even surgical video recorded in 3D doesn’t fully show the spatial relationships that the human brain is capable of decoding. It doesn’t convey the subtle tactile feedback that comes from operating on actual tissue. As a result, I’m very excited about robotic systems but prepared for this to take some time.
We all remember the ambitious predictions about self-driving cars taking over the roads quickly and driving more safely than humans. The problem is that not all roads are clearly marked. Sometimes we have to navigate through an unmarked grass parking lot while workers are pointing the way. How do you teach a robot to handle the occasional bizarre scenario like this when the solution is obvious to a human but the robot has never quite seen it before?
Lastly, if surgical complications are sometimes inevitable, how will we feel about it when a robot has one? It’s easy to forgive a human, but consider the uproar when fatal crashes occur in self-driving cars. Defenders of autonomous vehicles may be right when they say the accident rate is much lower than with human drivers, but try telling that to the mother of a victim.
We owe it to our patients to support the development of robotic surgery, but we must also accept that, especially early on, robots will have complications, just as we do.
Follow @DrHovanesian on X, formerly known as Twitter, and Instagram.
Collapse