How the EHR Makes Clinicians Feel Unsafe

by Lee Milligan, MD | CMIO & VP, Asante Health System

In healthcare, we spend plenty of time talking about patient safety, and rightfully so. But I’ve noticed that oftentimes it is the clinician who walks away from an experience with an EHR feeling “unsafe”.


My Experience

I was working in the ER shortly after our go-live with Epic. In typical emergency-room fashion, we were seeing a lot of patients. It was a busy place, and we covered various degrees of illness. One particular patient had what’s called an unstable angina (USA), which is essentially an almost heart attack that comes and goes. 

The offending clot teeters on the edge of causing blockage. One of the initial treatments for USA is a medicine called Heparin. I had several less critically ill patients in the ER at the same time, including one with an ankle fracture.

While trying to provide care in this busy environment, I had several tabs open in the EHR—each with a different patient chart. I accidentally ordered Heparin, a very powerful blood-thinning medicine, for my ankle-fracture patient. 15 minutes later, the nurse approached me and half-jokingly asked, “Did you really mean to order Heparin on your ankle fracture? Is this a new treatment that I am not aware of?” 

I was floored. I went back to the EHR and pondered, “How could I have done that?” Obviously, I canceled the Heparin and re-ordered it for my unstable angina patient, but the ill feeling in my stomach lingered.

It’s important to note that, at this point in our digital journey, nobody in our ED was more familiar with the use of the EHR than me. Spending months building out our instance of the EHR brought me a level of familiarity that was both hard-earned and authentic. 

Yet despite this level of engagement, I efficiently mis-ordered that potentially dangerous medication without giving it a second thought. If someone with my level of exposure to the EHR can make this type of mistake, is it really any surprise that many clinicians don’t feel safe trusting their actions within the EHR?

This speaks to the criticality of great human relationships. Not only did that nurse recognize my error and feel confident enough in her role to trust her instincts, but she felt comfortable coming up to me and saying, “Get it straight, Lee.” 

At the end of the day, that relationship saved my neck, saved the patient, and got our ED care back on track. That wasn’t the first time—and I doubt it will be the last—that a competent and engaged nurse saved the day.
 

Who’s Responsible?

For the case I described, I own my part in ordering the wrong medicine for the wrong patient. However, current EHR systems don’t do a great job at pointing out this sort of error before it’s too late. Clinical decision support (CDS) has been implemented by nearly every EHR vendor, but for those of us who use these systems, CDS remains rudimentary at best. 

It generally over-fires on most clinically unimportant scenarios and under-fires where you really need it. The result is alert fatigue—the desensitization of clinicians to safety alerts. Because of this, when CDS systems do fire, clinicians frequently ignore them or “click out” of alerts without fully digesting what message they intended to convey. 

For dedicated physicians, all of this contributes to the chaos of the experience. Remember, they spent over 15 years mastering patient care around a specific domain to become experts in their field.

Throughout this span, physicians learn to zero in on three elements with each patient:

  1. Building a relationship with the patient
  2. Diagnosing the problem
  3. Treating the patient with medicine and/or surgery to effect a cure.

Let’s consider a typical ambulatory encounter. From the vast array of maladies that a patient’s symptoms may represent, the clinicians first need to narrow the options down to a small differential diagnosis; without this, they can’t effectively treat the patient. In a typical situation, clinicians are trying to both build that relationship and simultaneously use most of their “cognitive RAM” to think through the diagnostic process.

In the midst of this interpersonal patient-care setting, clinicians sit down in front of a computer packed with icons, navigators, and wording which, in aggregate, look like the cockpit of a commercial airline.   

In the presence of the patient, clinicians then find themselves digging for the proverbial needle in the haystack, and they quickly recognize that they are at risk of appearing incompetent in front of their patient. Thus, the clinicians (a) don’t find what they need, (b) are embarrassed by their perceived incompetence, and (c) project an image which is antithetical to the  confidence they desire.  Is it any wonder that physicians frequently verbalize that they want to punch the screen? To top it off, physicians are rightly concerned that the EHR system could lead to any number of medical errors, either errors of commission or omission

An example of an error of commission would be what I described in the aforementioned case—ordering the wrong medicine for the wrong patient. An EHR error of omission can be just as dangerous. 

These may come in the form of a missed critical lab value buried deep within the chart or a note from another physician who saw the patient and observed something critical but whose insights are lost in the narrative of a bloated copy-paste note. For many clinicians, it feels like a minefield. Focus too much on avoiding errors of commission, and you’re liable to step on an explosive omitted detail. 
 

What’s the Fix?

The ultimate fix may take many forms. Still, one design philosophy will need to reign supreme—simplicity.

We should consider both immediate and long-term solutions. For immediate improvements, a consensus is building that there are three core components. First, providers should insist on an IT governance structure led by physicians and nurses. For example, at my institution, we have a physician advisory council, an ambulatory physician advisory council, and a nursing advisory council. 

These councils enable clinical leadership to rip out the garbage that doesn’t belong within the workflow, leading to a more intuitive EHR navigation design. Without these critical governance structures in place, ITS and operations will default to making decisions on behalf of clinicians. Second, training needs to be overhauled at many institutions—both initial and ongoing training are frequently abysmal. 

There are a lot of great ideas for how to impact training, but at the end of the day, you need to have buy-in from operational and ITS leadership to provide resources for this important effort. Third, along with ongoing training, clinicians should be given ample opportunity to deepen their adoption of EHR personalization tools—something proven to increase both satisfaction and efficiency.

The long-term fix, in my opinion, is taking clinician engagement to the next level: clinicians should fully lead the way on (a) intuitive and simplified EHR design, (b) smart, gamification-based training, and (c) rapid feedback loops. Let’s talk about design. 

The EHR vendor that ultimately wins this electronic health record race is the vendor that can mirror what the physicians see on the screen to their actual clinical workflows, without gratuitous distractions. I liken it to doing your own taxes with TurboTax. Without sounding like an endorsement, TurboTax is pretty straightforward. 

The software has two tabs at the top—the simple tab and the full-view tab. The simple tab streamlines the workflow into a basic question-and-answer format—you input your data, and it spits out your numbers. On the full-view tab, you get a behind-the-scenes look at all the long forms. You’re free to walk through the software using either view. 

Most of the time you would find yourself working within with the simple tab because that corresponds to the main workflow. An EHR designed in a similar fashion could be equally effective. It could be set up in a format where the physicians pick their tab; if a patient comes in for abdominal pain, the system could automatically open the abdominal pain tab, thus presenting the view based on that particular clinical scenario. 

Conversely, if you find yourself in a scenario where you want to check on a nontypical result—say a head CT scan in addition to abdominal pain—you can quickly adjust to a long-form view of the patient record by clicking the full-view tab. 

How close are we to making game-changing improvements in the industry? Although I can’t say for sure, it does feel like we are approaching a tipping point. Provider frustration combined with clear, hard-dollar ROI around improved efficiencies are aligning the planets on this topic. 

Nonetheless, it’s a big lift for vendors and clinicians. Up to this point, this level of investment hasn’t been considered urgent due to cost constraints.  In addition, this level of workflow intuition would require a cadre of clinicians spending hundreds or thousands of hours designing appropriate clinical scenarios. 

On the other hand, can you imagine an EHR scenario where the physicians efficiently find what they need, are confident in their EHR navigation, and do not make errors of commission or omission? Think about how much safer physicians would feel in that environment, navigating the EHR as an augmentation tool to patient care and not a distraction—a vision of an EHR which relieves clinicians of the mundane burdens of healthcare instead of compounding the chaos. 

The time has come for physicians to demand it.