On the whole, does the EMR do more harm than good in U.S. healthcare? If a case can be made to unambiguously answer in the affirmative, Elizabeth Toll, MD, does just that in “The Cost of Technology,” a compelling essay currently running in the opinion section of the Journal of the American Medical Association.

“We must pick and click according to the EMR's pathways, rather than by following the patterns of learning and thinking we have internalized over years of training and practice,” writes Toll, who practices pediatrics and internal medicine at a busy inner-city clinic of Rhode Island Hospital (RIH) in Providence while teaching at the Alpert Medical School of Brown University, also located in Providence, R.I. “You survive in this new system by giving the computer complete attention, the kind of attention we used to reserve for a patient. By default, the patient moves down to second place. As sad and horrifying as this sounds and feels, it is becoming the new reality.”

Toll expounded on her points during a phone interview.

You led off by describing and showing a child’s drawing sent to a colleague. What struck you about the picture?
TOLL: There’s a famous painting from the late 1800s by an artist named Luke Fildes called “The Doctor.” It shows a doctor in a home, sitting next to a child who’s extremely ill. Her parents are in the background. It shows the very intense focus of this physician connected with this child and her parents. That picture was used by the American Medical Association during, I believe, the Franklin Roosevelt administration, when physicians felt the government was getting too involved in their business. They had a campaign essentially saying, ‘Don’t let the government into your office.’ This picture appeared all over the U.S. at that time. 

In 2009, we switched over to EMRs—and it has been a completely different job ever since. We had been doing e-prescribing up to then and everyone saw the value of that, where you could easily see the last time something was filled. You could easily figure out who did it and when, and you could do it from a distance and so on. So we knew that this could be a valuable change. But the system is so over-engineered, and there are so many design flaws, that this has just become a really different job.

We have a lot of young residents. They’re really good on computers. All of us have worked really hard to learn the EMR. But it’s made our days 10 percent to 30 percent longer. Everybody charts late at night and we constantly see everything we haven’t done. We never feel finished. Stuff comes in 24 hours a day, our little jelly beans are blinking to let us know ‘more labs, more prescriptions.’

And there are some really serious safety issues that come up, mostly involving prescribing. When you order medicines, you can’t see the full prescription in the window. You just pick the medicine and pick the dose, pick how many refills and so forth, but you can’t see the whole thing. You can open it in another version to see the whole thing at once, but we have patients who have 12 to 15 medications and there simply aren’t enough hours in the day to do things completely and correctly using this system.

Going in and out of all these windows, doing all these different things, I personally felt like I was training myself to have Attention Deficit Disorder. The way that I had learned to think, and the organized way I kept things going, became completely irrelevant. I had to re-learn everything, and the amount of time involved—we have 20 minutes and can’t fit all these things into that.

People have found different ways of getting through it. We will often pre-round on our patients, which means that we’ll go through and look at what immunizations a child will need or what labs a patient will need, and we’ll pre-fill the record with that, so we’re not doing that work during the patient visit. Each of us has different ways. Some people type while with the patient, but I don’t. I print out the last note and write on it so that the only thing I do in the exam room is sit with the patient. Then I excuse myself at the end of the visit, send any medicines that need to be sent, get any labs ready and then I come back in. I just can’t stand to have patients watching me use the computer, because you literally have to shut the patient out in order not to make mistakes.

My husband is a pediatrician on the same EHR system. We’re always on the computer at home, all the time. We’re on the computer at 6 o’clock in the morning, 10 o’clock at night. There are all these reports on what you haven’t done and reconciled, and you just can’t finish.

What percentage of all that time on the computer involves direct interaction with the EMR?
A lot. My average day is two to four hours longer than it used to be. I am on the computer all the time, and it’s almost all EMR-related. I barely do email. It’s dealing with our jelly beans, which is what we call the blinking lights that tell you labs you need to take care of, telephone calls you need to answer, refills you need to do, documents that come through—reports on people in the hospital or emergency room, x-ray reports, specialty reports. There’s just the sheer volume. It’s the mindset of feeling that we used to have a people job and now we have a computer job, in addition to safety issues.

The EMR was designed to demonstrate the pieces of the record that you have to attend to in order to bill at a certain level. If you just enter a few questions and you only enter part of the exam, and you only add medicines and you only do this or that, you can only be reimbursed a certain amount. But if you asked about, for instance, the family history, the surgical history and the social history, then you have all the elements to charge more. So there’s an incredible temptation to just push, push, push and bring forward everything from the previous notes without re-asking the questions.

This creates a huge problem: The records are full of lies. They’re full of things that [physicians] have said they’ve done but truly haven’t. The patient has been in eighth grade for three years. The patients are divorced, but in the record they’re still married. The patient used to work as a nurse and now works as a librarian, but it hasn’t been changed in the record because people are giving quick, push-button answers to save time, and they don’t update the info. You can see this as you go through small things in the social history but also in [clinical histories]. Yesterday, someone sent me a letter about an amputee patient he sent to a podiatrist. He got a report back on both the patient’s feet. This patient only has one foot.

And the motivator behind giving inaccurate, push-button answers is simply the ticking of the clock?
Right. You pull across the standard physical form, but the truth of the matter is that sometimes you completely forget. You forget that the patient had an ear infection and report that the ears were fine because you have to make the change. You can bring the whole ‘Normal’ thing across, or your standard exam across or the last exam across, but you have to modify that if something is different today. In the old days, if you were writing a de novo, when you got to the ear, you described the ear the way you found it. But in this world, you just click click click click click to get everything done, and only then do you realize that you’ve just written something that isn’t true. 

There are also problems with recounting necessary patient stories. Let me give you an example. A patient came in with an asthmatic episode and we gave the patient a breathing treatment. The next week, the kid was back for a well visit and I noticed the nurse going into the room with a nebulizer. I asked why she was going into that room with a nebulizer, and she said an order had been written for a breathing treatment. The order from the previous visit had automatically populated into the next visit.

And that was automatic by the software? That wasn’t someone intentionally or inadvertently clicking something?
It was automatic by the software, and I wouldn’t have dreamed up that problem. It was only because I fell over it that I even realized that kind of thing was a possibility. That has since been corrected, but there are all these kinds of mistakes brought on by the software. You keep thinking, ‘I don’t even know what mistake I’m about to commit.’

There are many dimensions to it. The first is that I like to focus on my patients, and I feel pulled aside with the machine and the work. I am constantly crouched over a machine. We all have carpal tunnel, back problems and eye strain. It’s a very different job physically, and patients feel we care more about their birthdays than their names. Some [physicians] are really good at all this.

What’s been interesting is, I see the value of being able to read stuff and I see the value of being able to find stuff in the record easily, and it certainly makes sense to centralize things in a single place after all those years trying to find stuff. But it’s come at quite a cost.

I speak only from my limited perspective, having used the one outpatient system—we have a different system for inpatients and, of course, these two systems don’t talk to each other, so we’re constantly in and out of these two systems. The administrative staff can’t schedule appointments in a single system, so they’re in and out of the two systems. And nobody wants to get rid of paper, so everything is printed. We’re going through paper at a rate we’ve never seen before even though everything is supposed to be electronic.

Click here to read Part 2. 

(Editor’s note: A few days after the release of Toll’s JAMA piece, Bloomberg Businessweek posted a feature that, in effect, offered something of a counterpoint. It’s headlined “Why Don’t More Hospitals Use Electronic Health Records?” and it’s online here.)