Physician burnout is a hot topic these days — for good reason. As Steve Strongwater, President and Chief Executive Officer for Atrius Health, notes in his discussion with Tom Lee, 54% of U.S. physicians are experiencing physician burnout. Are electronic medical records worsening the problem? In short, yes. Find out why, and how we can address this to improve not only the quality of care for patients, but also the quality of life for physicians. Read or listen to the interview below.
Tom Lee: This is Tom Lee from NEJM Catalyst, and I’m here with Steve Strongwater, a rheumatologist who is the CEO for Atrius Health in Boston. I first met Steve when he was Chief Transformation Officer at Geisinger Health System, and he came in the summer of 2015 to be the CEO of a very respected integrated delivery system built around physician practices in Boston: Atrius Health.
Now, Steve and his colleagues have been dealing, like so many of us, with clinician burnout, particularly among physicians. We recently had an interesting conversation about physicians and their long love-hate relationship with electronic medical records, which we all know are essential to high-quality care in this day and age, but it does seem like the ratio of love to hate seems to be moving in the wrong direction. In fact, EMRs are being blamed as a cause of physician burnout by many clinicians. So, I wanted to take time with [Strongwater] today to hear from him on his take on these issues and what Atrius is doing about them. So, Steve, what’s your take? Are EMRs a symptom of burnout, or are they the actual cause of the disease?
Steve Strongwater: Tom, that’s a great question, and I really appreciate the opportunity to talk this over with you. First, I would hope that everybody agrees that physician burnout or physician wellness is a real problem in this country. As measured by objective studies, 54% of U.S. physicians are experiencing physician burnout, and there are many reasons for it. It’s not a function that people just are whining and they’re not as tough as physicians used to be.
About 80% of physician burnout is really due to workflow issues, and as it turns out, the way the electronic medical records have evolved — unlike in other industries where automation has made work easier — the electronic medical records have added work. Now, that may be a function of many different things: requirements for more data capture for things like quality, or coding for billing purposes, or regulatory requirements. But the electronic medical record has clearly added work to a physician’s day, and people who are so dedicated and committed are working late into the evenings in what we would call “pajama time.”
In general, what seems to happen is that our docs will work during the day — they’ll work a full day, sometimes eight or 10 hours or longer — they’ll go home for a brief period of time, and then they’ll get back on their record in order to finish the work of the day that evening. So, we are adding work, and over time what we’re seeing is that people, by virtue of wanting to meet the needs of their patients, reduce their amount of FTE (full-time equivalent time), so that they can use that time to finish the work during the week to maintain balance in their lives.
So, the short answer is that the electronic medical record has contributed to burnout as one component of burnout, and I think it’s really important for the designers of EMRs to try and work to improve the user interface, the workflow, in a way similar to the way smartphones work or when you start a search on a search engine and it almost anticipates your needs. We’re just behind. We’re almost in generation one of that electronic medical record.
Lee: Before we go to the solutions, let’s go a little bit deeper into the path of physiology. I know that there’s been tremendous medical progress, and there are many more people involved in care today than when I was coming out of my training, but how is it that electronic records might actually be worsening the problem? Is there something more than just putting in front of me all the work that I need to do to take care of my patients?
Strongwater: The short answer is yes. What has happened over time is we have asked our clinicians to become sophisticated coders. They are clicking through screens that are cluttered, that are not designed with human factors in mind. They are filling out forms that at one time would have been triaged to a medical assistant or health assistant. They’re having to respond in their inbox to messages that otherwise historically would not have come to their inbox, that would have been filtered away, and so it literally has added work to a busy day.
It has also negatively impacted what I would consider face-to-face time with patients. If you’re a clinician in a room, you’re often not looking into the eyes of your patients, you’re looking into the screen of the computer, and as a consequence it’s impacted the [patient experience] as well. So, it has definitely negatively impacted the workflow and the patient experience as a function of the way the EMR has been designed.
Lee: The EMR has put work that ought to be done in front of us, but it’s created work and it’s distracted clinicians and others from some of the important interactions that lie at the core of health care. So, what’s the solution?
Strongwater: Well, we talked a little bit about some of the solutions. I would refer to this generally as intuitive design, that the designers, who have the ability to watch and monitor physician workflow, begin to understand the dynamics in an outpatient setting of what an office visit is like: what could be done before the visit, what needs to happen during the visit to present information to the clinicians in advance of the clinician having to go through a series of clicks and screens to make that interaction easier, to move more toward an intuitive user interface like your iPad or your smartphone, and to begin to use artificial intelligence and machine learning to anticipate the needs of clinicians, and then to automate as much as possible during that workflow.
One of the things I didn’t talk about are these best practice alerts, which [are] alerts in Epic that you need to do something. Often, there are so many of these best practice alerts that the clinicians just bypass them. Again, I think that has to be looked at from the perspective of better human-factors engineering. I think we’re in version 1.0. I think we also need to add in much better analytics to anticipate who’s sick and who has care gaps and to make that as easy as possible.
And when I talk about machine learning we’ve added so many clicks on the conversion from ICD-9 to ICD-10 coding, there just has to be a better way to do that work. I think about the way Quicken was originally designed, where it sort of learns as you go along. If you’ve seen a thousand patients and you’ve never used a particular ICD-10 code, why should it be on your screen? And you need to maintain updated problem lists as it relates to being sure that you can close all the care gaps and you know what interventions need to be done. It’s a bit of a long-winded answer, but the short version I think is that the user interface and artificial learning would go a great deal toward improving the workflow and reducing burnout.
Lee: Part of the problem is almost surely the electronic records, the software as it were, but part of it is likely to be humanware: how we’re organized, the extent to which we’re organizing teams that can actually trust each other. Can you comment on some of the human-factor design that actually has to go on with humans?
Strongwater: Yeah, it’s a great observation, Tom. I do think we have definitely moved into team-based care. If you could filter away work, triage it to the top-of-license person — whether that’s a nurse or a medical assistant or a pharmacist or a social worker — that would reduce the workload for the clinician. I think it has to be done in such a way that there’s confidence in the team that that work will be done, because what tends to happen, at least with our primary care docs, is they’re insecure that we have higher liability, that is that things happen 100% of the time, and so they want to check on it.
So, if I as an example say, “Look, I can take all your normal labs out of your inbox, would that work?” They’ll say, “No, I want to take a look at them.” In and of itself, checking those normal labs doesn’t take that much time — until you add 150 patients, and then it adds work in that results review. So, if we could get the team to work at a high level of reliability and people had trust and confidence, I think you could do a great deal to reduce the demand time on the clinicians. I think that is absolutely the case.
Lee: One other thing that runs through my mind is customization of electronic medical records. I wonder whether we sometimes shoot ourselves in the foot. I recently visited an organization who said they can’t upgrade to the new version of their electronic medical record because they’ve done so much customization of their current edition that it would be too complicated. Could we be making life more complicated than it needs to be, and that it would be better for us all to get used to vanilla, for example?
Strongwater: It’s a great question, and I would want to reframe the question and somewhat the answer. When we started out with electronic medical records, every organization wanted to customize it to their own needs. In part, that’s because there wasn’t a good enough product out there. I would argue that we need to allow the best product available to evolve and then make that more broadly available.
In the case of Epic, when you have one instance of Epic, you have one instance of Epic. You get sort of a vanilla shell, and then you customize it and there are no two instances that are exactly the same. But I would argue that if you had a mechanism like a library of apps that you could choose from and then import into your baseline system, that you would still be able to upgrade much as you upgrade your iPhone and have best of breed available in the practice. It hasn’t happened that way.
When I refer to more artificial intelligence, nothing would be better than to have a wrapper that would provide pretty current upgrades of content — medical content, workflow content — that marries up to a basic EMR program. A lot of this has to do with whether the electronic medical record is going to be able to do all these things all at once for everyone.
I have a feeling that over time, we’re going to see the IBM Watsons of the world providing supplemental content and workflow information that complement the available tools inside an embedded electronic medical record like Epic. And when that happens, it’ll be possible to go with vanilla because you could pick from an outside library of enhancements that you could pull in seamlessly into your ecosystem, which would be wonderful.
Lee: I know that better days do lie ahead. I just don’t know how far ahead, and I also know that you and your team at Atrius are doing some innovative things to hold your own feet to the fire to hasten the arrival of those better days. Could you comment on that?
Strongwater: Well, we have an initiative modeled after the ABIM [American Board of Internal Medicine] called Returning Joy to the Practice of Medicine. In order to make that real, our C-suite and senior administrators’ incentive plans are in alignment with that, really tracking the time spent on Epic — particularly that pajama time — and we expect that we should be able to reduce that time out of work, meaning out of the office, materially. And that is built into our C-suite incentive plan.
We hope that the administrators and the physicians come into alignment. Certainly the docs would love that. We need to have our administrators understand that workflow matters, quality of life matters, returning joy to the practice of medicine — not only for the physicians but for the rest of the practice — is really important. And we hope that joy returns really quickly.
Lee: Well, I know that your physicians are rooting for you to hit those incentive targets, and I hope we can check back in a year and see how it went.
Strongwater: We look forward to it. I hope we’ll have great news to report.
This interview originally appeared in NEJM Catalyst on October 14, 2016.