MGH Goes Epic: What I Learned in the Trenches

Xconomy Boston ^[1] — 

On April 2 at Massachusetts General Hospital (MGH), the green charts went away. Every morning around 6:30 am since starting residency, my co-interns and I had frantically scrambled around the halls searching for these plastic binders, which contained the sole record of our patients’ vital signs. It was a ritual that blended equal parts anxiety, frustration, and disbelief that “Man’s Greatest Hospital” still had not managed to bring vital signs into the digital age.

That all changed when the old electronic medical record (EMR) system was shut off and Epic^[2]—the juggernaut EMR that cost MGH’s parent company, Partners Healthcare, over $1 billion—was turned on. Suddenly, vital signs were available with a single click.

The addition of vital signs into the EMR was an unequivocal improvement. Residents celebrated with a video dramatizing the longed-for demise of the green chart, followed by a ceremonial burial. A few of the old guard stood around, bewildered, like shell-shocked soldiers at the end of a war, unsure what to do now that the enemy was finally gone.

I have been surprised by how easy the transition to Epic has been. In the first few days, resident and nurse work areas were flooded with “super users” to answer our questions. The super user on my floor was an intern at Brigham and Women’s Hospital. He had already used Epic for most of a year and knew the answers to all my questions. Now, a month into Epic, I feel not only comfortable with most tasks but find that the program saves me time.

That said, as with any major software launch, there have been some challenges. I’ve learned a few lessons over the past five weeks of combat with a new EMR. Here are three key lessons.

The first lesson is that simplicity is a virtue. After seeing vital signs projected onto my computer screen for the first time, the next thing I noticed was that there are about seven different ways to access the vital signs in Epic, and each way provides slightly different information. The same goes for laboratory and imaging results, notes, medication records, and orders. There are overviews and detail views, lists, tables, and charts, and physician views and nurse views. A co-resident and I can be looking at the same patient’s results and see very different things.

Some of this complexity is an inherent feature of Epic, but part of it was our own creation. Partners negotiated for a great degree of customization in its Epic build. This has resulted in added bells and whistles. Some of these features may be helpful, but many features turned out to be problematic. Partners is now working to scale back some of that customization to simplify the system.

Lesson two is that an EMR should be flexible. Epic often requires hard-coded responses and forces users into standardized pathways. This can be a useful safety measure to prevent errors and ensures adherence to protocol. Hard-coded responses also allow administrators to collect and track practice patterns with much more clarity, permitting more focused interventions to improve quality of care.

However, this approach falls apart when a situation is atypical, requiring an atypical management approach. Several times, I have been forced to order something a standard way when I really wanted to make a customization that would have been better for the patient. We need to recognize that not every situation is typical and there should be a way to work outside standard operating procedure.

The third lesson is that frontline providers need an ongoing voice in EMR design. A few features of the Partners version of Epic clearly reflect gaps in understanding of clinical practice. Some orders have defaults that would be funny if not so dangerous (e.g., it’s easier to order twelve days of intravenous fluids than twelve hours’ worth). Other orders flash nonsensical warnings, requiring several clicks to dismiss. These warnings slow down my workflow and desensitize me to serious alarms about potentially dangerous interactions or allergies. Impractical features like these have already attracted the attention of MGH administrators, who are seeking out the perspective of residents and other providers to guide improvements.

This approach is key: while glitches and obstacles are inevitable in an EMR rollout of this size, MGH is taking the right step in gathering the voices of users to make improvements based on practical experience.

The Epic rollout at MGH has gone well. For other EMR launches, leaders should prioritize simplicity, flexibility, and listening to the voices of frontline users.

Alex Harding is a resident physician in the primary care track of Internal Medicine at Massachusetts General Hospital. He played a minor role in preparing for the Epic rollout at MGH. He has no financial interests to disclose. Follow @alexharding7