HBR Article: Computers in Medicine

Put Doctors at the Center of Health Care Tech. By Daniel Marchalik. Harvard Business Review.

Interesting article from Harvard Business Review looking at the current state of computer technology in healthcare. The COVID epidemic has highlighted the promise of healthcare tech. For example, utilization of telehealth services has skyrocketed. Yet healthcare technology rates an “F” on facility of use by doctors, contributing to burnout. “Put patient care first,” advises Marchalik.


In healthcare, maximizing shareholder value boils down to a throughput problem. Sick people go in at one end of the assembly line, and well people come out the other. Healthcare is highly labor-intensive, so productivity is the name of the game.

In every other business in the world, computer technology has improved productivity. But not in healthcare. Quite the opposite, in fact.

I thought computers made us more efficient?

On average, I spend seven and a half minutes in any patient encounter thrashing around with computer software that is ill-conceived for the task at hand. I would estimate that has reduced my productivity by about 20-30% compared with the days when things were done on paper.

There are several potential reasons for this. The primary reason, it seems to me, is that electronic health records (EHR) are, primarily, billing systems. With record-keeping appended almost as an afterthought.

The VA’s VISTA/CPRS systems stands out as a notable exception. It’s fast, clean, and was developed with a great deal of physician input. To be clear, it is showing its age. The prescribing interface is clunky at best, as is its imaging interface. And it lacks interoperability; among other things, it can’t talk to DOD systems. It’s due to be replaced by an off-the-shelf solution (at considerable taxpayer expense).

In fairness to commercial products, when EHR’s first hit the marketplace, doctors were not at all tech-savvy. In fact, most did not know how to type. To this day, data entry typically consists of some combination of point-and-click and lousy voice recognition. VISTA/CPRS requires the doctor to know how to type. Learn that skill, and the computer can, potentially, become your friend. Otherwise, your days (and often your nights) are spent watching the spinning wheel of death after every mouse click (because systems evidently have trouble keeping up in the hardware and bandwidth departments).

Beyond that, EHR’s collect a prodigious amount of data, most of which might satisfy some regulatory requirement, but little of which has anything to do with making sick people well. Meaningful information is buried in a haystack of stochastic noise. To top it off, we are expected to be part-time data entry clerks, inputting that stochastic noise.

Productivity-wise, as a general rule, it’s probably best not to make the $150/hour person do work a $20/hour person could do. I can see having me enter my own prescriptions — it’s no skin off my back, and it substantially reduces the probability of medical errors. Personally, I can type faster than I can dictate, so I’m fine with entering my findings as well. But somewhere along the way, the government got the idea in its head that they should only pay for a visit if I ask about a hundred random questions they define. Current practices allow a medical assistant to ask some of those questions, and record the results in the record. That still leaves a chunk of time for me to ask a bunch of useless questions.

So, start with a fifteen minute followup appointment. Subtract seven and a half minutes for computer-thrashing, and a couple of minutes for the exam. That leaves the patient with five and a half minutes to blurt out what the problem is, and essentially no time for the doctor to explain what’s going on. I’ve noticed recently I’m typically scheduled a half hour for appointments these days; but to maximize charges, I have about fifteen minutes of “busy-work” — asking a slew of mandatory questions, and recording the results point-and-click-style, spinning wheel of death and all, so we are back where we started.

I type fast, and figured out how to do time-based billing, so I still have time to talk to my patients. And I’m getting the feeling, I’m the only one. I hear that a lot from my patients. “You are the only doctor who ever listened to me.” (Meaning, I talk to them, and show that I was listening).


Marchalik is right: the human comes last in this equation. Improving the technology might help, but a couple of other factors also need adjustment:

  1. Abandon the Resource-Based Relative Value Scale, a reimbursement system based on Marx’s Labor Theory of Value (I kid you not). Indirectly the source of most of our busy work. Prospective payment schemes, time-based billing — anything would be better. Part of this is the government’s problem, and part of it is organized medicine’s problem. Hint: Marxism doesn’t work.
  2. Doctors need to learn how to type (a problem that, I imagine, will take care of itself in time).
  3. Administrators need to understand that you can’t improve productivity by griping about it. Among other things, doctors should be doctoring, not doing data entry.

Leave a Reply