Too Much Health Tech, Not Enough Health Tech

In healthcare today, it looks like there is a tension between too much technology, and not enough technology.

And simultaneously, too much humanity, and not enough humanity.

It seems like these things, the movement towards technology and away from humanity, are accelerating at an incredible pace, potentially bringing us to a place we don’t necessarily want to be.

Across the entire spectrum of healthcare, we’re seeing inroads in care by technological advances that were unheard of just a few years ago.

In fields like radiology and pathology, artificial intelligence (AI) systems are assisting the doctors performing these services, proving to be more efficient or at least adding to their efficiency, probably still nowhere near replacing them but definitely able to serve and enhance their care. And AI is now reading our patient’s messages in the electronic medical record and offering up suggested responses, helping with billing and coding and scheduling, and guiding doctors toward the best evidence.

Importantly, this future holds promise to relieve clinicians of the burdensome administrative tasks that prevent us from really taking care of patients.

Patients are using these systems as well, engaging with chatbots and agentic AI programs that explain what might be causing their symptoms, what their labs might mean, what that scary CT scan report in the patient portal might mean.

As these systems get smarter and smarter, we are hearing that they’re getting closer to looking like something that will approximate the practice of medicine.

First we heard that they were able to do a reasonable job of passing the national board exams for medicine. Now there are companies with applications where patients can pay a fee and interact with an algorithm, where they can voice their complaints and symptoms, run through a series of questions, and come up with a differential diagnosis, which they are free to bring to their doctor and have them take over from there. Some are apparently close to being able to prescribe medications and order tests, but for now, state medical boards have slammed the brakes on these.

As these things get better and better, there is the terrifying potential of the computer taking over the job.

On the humanity side, our patients are complaining that we’re not spending enough time with them; we’re not listening to them; we’re not examining them; we’re sending brief terse responses to their detailed portal messages; we’re not available to see them when they want to be seen, that we’re spending far too much time referring them out to specialists rather than addressing their symptoms and concerns and tackling the problems on our own.

We’ve gotten to a place where doctors want to use this technology, but they’re also really scared of it, and patients want to use it as well, but they seem just as worried about letting it completely take over their healthcare delivery.

Someone told me recently that we’re living in a time between two types of healthcare systems.

The first one is where a doctor is more likely to be sued for using AI to take care of a patient. And the other one is where a doctor is more likely to be sued for not using AI to take care of the patient.

So how do we address this tension, how do we safely get across to the other side?

In thousands of sci-fi movies, we’ve seen future visions of healthcare where robots and scanners and other devices deliver most of the healthcare to patients.

A valiant injured space explorer is brought in, laid down on a sterile white table, a clear plastic shield slides over them, lights flash and laser beams scan up and down, and then they emerge healed.

This is only a few generations away from when an injured farmer was brought in to the country doctor, the same one who managed their diphtheria and delivered their babies, stanched their bleeding, removed their appendix, sewed up their lacerations, and got them back to work.

Practicing now in this uncomfortable middle, we’re all dipping our toes into the waters, at times tenuously, at times plunging headlong into the icy abyss below.

Being part of the conversation, bringing our knowledge and expertise and experience, will be the difference between success and failure. While the companies that make these things are really smart, unless they have a whole lot of doctors and nurses working alongside them, I’m worried they’re going to get so much of it terribly wrong.

Just as every new app and widget in the electronic medical record needs the approval of multiple levels of governance, so too should a stamp of approval be required from the people who are actually going to be using it — the doctors, nurses, and of course, the patients.

I understand it’s hard to show the inner workings of many of these programs, that sometimes there’s a big black box we put data into on one side and out pops an answer on the other, and it’s hard to really explain what happened in that quiet dark space in between. But unless they let us take it for a test drive, help them work out the kinks, we are unlikely to end up with a system that really moves things along in the right direction.

My hope is that these systems get to be good enough that they can run along beside us, listening in on our conversations, collecting data from all over the ether, pharmacy records, billing history, laboratory, old records buried in a file room somewhere, photographs and videos, analysis of a patient’s facial expressions and body language, the results of our physical examinations, and so much more, adding all this into the mix of the doctor’s interpretation, to help get to the truth.

The real truth, which almost always lies somewhere in the middle.