Not Either/Or: Doctors and Algorithms

Almost ten years ago, I wrote a thesis for a Master’s degree at Georgetown. It was about the artistic representation of the machine in human form. I was curious about the human drive to create cyborgs and robots and machines that look and behave like us, and what it means for us to also have the fear that seems to crop up in so many movies and books: The fear that the cyborgs and robots and machines that look like us are going to turn around and destroy us. I looked at a couple of texts, and I considered the Prometheus myth and, of course, Frankenstein’s monster. Naturally, I called the thesis “Doomsday or Desire: Do Androids Dream of Robotic Lovers?” (Download it here.)

It never occurred to me that a decade later I’d be considering a very similar question, albeit in a completely different industry and from a different angle (and without all the critical theory).

The other day, there was a big media flurry about a pretty cool and exciting announcement. You may have heard about it: The Qualcomm Tricorder X Prize. It’s a competition with a $10 million prize for the first group to design a Star Trek Tricorder: A device that can monitor a set of health metrics and diagnose 15 diseases. It’s pretty cool, right? Absolutely. Talk about potential!

But something was nagging at me. So I got to thinking. I got to thinking about a self-scanner and diagnoser. I got to thinking about the rise of self-tracking apps and health technology in general. I especially got to thinking after I read Vinod Khosla’s piece on TechCrunch, with the unfortunately polarizing title Do We Need Doctors or Algorithms? (NB: He addresses some of my concerns at the very end.)

Here’s what’s been bugging me: There’s a particular strain of thought, or maybe even of person, that says, “I have a set of constructs that solve these particular problems. Hey, you know what? I bet I can use them over here to solve this totally different problem!” Now, I’m not talking about using a theory from one sphere to help understand something in a totally different sphere. I’m talking about the applied version of my favorite aphorism: Running around looking for nails while wielding a really big hammer.

There are a lot of amazing ideas and possible solutions to big problems coming out of the health tech sphere right now. But there’s also what I see as a lot of this hammer/nail-ness too, this “I’ve figured out how to reengineer this over here, so I’m pretty sure I can reengineer this as well” - even though the bigger problems behind and connected to “this” haven’t really been explored. I see “my best guess” and “I’m going to assume here” more than I see “patients want to self-diagnose these types of ailments, but they want specific types of physician support in these medical cases.” 

Let me explain. As I mentioned, I think the idea of a self-scanner could be great. Empowering patients is really important. But once a patient is empowered, what then? How do we work with physicians and nurse practitioners to give up some of the control they may (or may not) want to have over their professional knowledge? How do we get them to work with patients and use this technology in a mutually beneficial way? How do we educate patients so they can use this technology to have better healthcare experiences and make healthier choices? How do we go rebuild the system to train and create better doctors, period, so labor isn’t being duplicated and patients can get the best out of doctors and doctors can focus on necessary clinical tasks and the personal aspects of patient care?

Just creating the technology isn’t enough. Like I’ve said before, sometimes we look at technology as this silver bullet. Ah ha! This - this one - this is what will fix this problem. We’ll get rid of crummy doctors, right? But isn’t that treating the ecosystem of healthcare the way many of us treat our own bodies? Rather than engage in some preventative care and understand the root cause of some our problems, we allow it to get bashed around and then hope technology will save it. Technology can help it, most surely, but we have to fix the whole system. Including the human parts of it.

More than that though, if we see doctors as replaceable by algorithms, what does that mean? Yes, I know, part of being a doctor already is using an algorithm, so why not just let a computer do it since computers are more powerful. I’m not arguing against that. What I’m saying is: If we take much of the diagnosing and clinical decision-making away from doctors, and then the surgical duties, what’s left? Yes, there’s personal connection. But so far no one’s offered a solution for making sure that personal connection is good, for helping both clinicians and patients be better at making and sustaining it. We’re focused on building the technology but we need to also dig into in medical schools to make sure we’re creating a system that can support these technologies in positive ways. I’ve seen some people argue it’s just an expensive placebo. I disagree, because medicine isn’t just a quantifiable science. With the best doctors, it’s also an art, based on observation, an ability to synthesize knowledge and experience, an understanding of people as well as science and medicine. 

Think about what would happen if we took away diagnosing entirely from clinicians. Think about what would be left - for clinicians to be the janitors of health? If we gave the creative part of your job to a computer - because it’ll be entirely possible to do that in almost any field - what would be left for you? What would your value be?

So let’s not make it an either or. Let’s build powerful algorithms and more powerful doctors and more powerful patients. Let’s create a better system.