Several weeks ago, as I was walking through the halls of our hospital, out from an intersecting hallway drove a robotic delivery cart.

Unmanned and adorned with sensors so it could tell where it was in the world, it motored along on its way, heading somewhere with a bunch of flashing lights and audible warning signals.

I was told this was part of a new meal delivery service that had been started in our hospital, that promised more efficient and time-sensitive (read “food still hot when it arrives”) delivery of meals cooked on-demand for our patients.

The little guy (although it probably outweighed me) whirred along just fine, turning corners with ease, avoiding obstacles, coming nowhere near patients or pedestrians in the hallway.

But then, shortly thereafter, we heard that the delivery service had been temporarily suspended, and that made me wonder why.

Had it delivered the wrong food to the wrong patient? Refused to disgorge its meal tray when it arrived at its appointed destination? Maybe it simply wandered endlessly, never making any deliveries at all.

There must have been some bugs in the system, and I suspect that those in charge of the program were closely monitoring it and realized that there was some fine-tuning to be done, and smartly took it off-line.

Today, once again walking through the hospital on the way to see a patient, I spotted this automated creature coming out of a hallway just ahead of me, beeping along and making crisp turns, avoiding an open door and a trashcan, heading along its way.

But then I noticed that following right behind it, a few feet back, was somebody whose job apparently was to keep an eye on this robot and make sure everything went OK, taking notes on a device that I suspect was a progress report and some quality measures that those who sent it out had decided were critical to make sure all went OK before they let the thing out on its own again.

What does this have to do with building a better health care system, and creating a more patient-centered way of delivering the health care we think our patients need?

If delivering food to patient’s bedsides without a human being is the right way to go, then the people who designed the system were doing the right thing, making sure that they did not just turn on the system and let it loose.

All too often, as we practice medicine here in the 21st-century, the providers on the front lines taking care of patients, and all of the other members of the care team — from registrars at the front desk, to technicians taking vital signs and rooming patients, to nurses delivering education and vaccinations, drawing blood, and doing EKGs — are subject to somebody coming up with an idea for how to make our lives better, how to improve care, and then putting these ideas in action, sometimes without fully thinking it through.

What comes to mind are the multiple quality metrics we are asked to satisfy, boxes that we click, certifying that we’ve asked about pain, about smoking, about patients’ fall risks, about what we are doing for a body mass index out of range.

These are the things we’re judged on, but many of us out here trying to take care of people wonder if that’s really where the best bang for the buck is.

Perhaps just as that little robotic delivery service had someone following along, maybe we should be exploring whether requiring providers to put a pain score down for every single office visit actually did anything to change or improve the management of pain in the practice of real-world medicine.

In fact, there are those that argue that the fervent desire to document pain scores, and the regulatory requirements aimed at driving them ever closer to zero, may have contributed to the current opioid epidemic ravaging our nation.

Does checking the box that says the patient hasn’t fallen in the past year improve their health? And if they have fallen, does checking that box lead to a decreased chance of them falling in the next year?

I don’t think we are looking with enough of the critical researcher’s eye at whether these things really improve the quality of care of our patients, and maybe we need to stop and think a bit before we leap at these things.

But if we can do it for robotic food delivery, then we sure as heck should be able to do it for actual care delivery for our patients.

Beep, beep.

Fred N. Pelzman is an internal medicine physician who blogs at MedPage Today’s Building the Patient-Centered Medical Home.

Image credit: Shutterstock.com