How Brigham and Women’s Hospital is Relieving EHR Alert Fatigue

Picture life on the internet before pop-up blockers. An individual would log onto the internet, and unrelated or obnoxious pop-ups would fill the screen. EHR alerts are just like those web browser pop-ups, except these alerts are not only frustrating, but an excessive amount could also carry significant consequences.

“An EHR alert appears and clinicians develop a reflex to immediately close it,” David Bates, MD, chief of the Division of General Internal Medicine at Brigham and Women’s Hospital, said in an interview with EHRIntelligence.

“That happens especially when the warnings are mostly not that clinically useful. Clinicians are very motivated to get their work done and they're typically pressed for time. One of the key things about alerting is that it's critical to do so, fairly judiciously. In other words, only when there's a reasonable chance that the clinician should change what they're doing.”

Bates and his team of researchers found that Brigham and Women’s clinicians were getting roughly one alert for every two medication orders, and clinicians were overriding approximately 98 percent of the alerts.

Although warnings can offer providers practical suggestions and updates, EHR alert fatigue has been an issue for clinicians already struggling with EHR usability problems. Low-value EHR alerts can disrupt patient care and contribute to clinician burnout.

“One of the big issues is that many of the clinical systems that are in routine use today, alert too frequently,” said Bates, who also doubles as a professor of medicine at Harvard Medical School. “When clinicians are overriding that high a proportion of alerts, clinicians get very used to closing the alert, and sometimes they aren’t fully processing what the alerts are saying and they tend to stop paying attention to the important alerts.”

Unsatisfied with how their EHR vendor fired off alerts, Bates and his health IT team tapped Seegnal eHealth to leverage its EHR alert solution and conduct an EHR alert study at the Boston-based hospital. 

The health IT professionals compared Brigham’s current EHR solution alerts from inpatient and outpatient settings to the vendor’s alerts from a de-identified data set from the two settings.

“We looked at how often the alerts were valid and we also looked at how the clinicians responded to them,” Bates explained. 

The researchers continued to run the same number of patients through its new alert system to see how many alerts would crop up.

“The biggest finding was that moving to this approach would eliminate 93 percent of the medication-related alerts that clinicians see now, which would be a big change.”

Following a quick EHR integration, the alert vendor solution evaluated the EHR vendor alerts to try and improve the quality of each alert. It also presented the alert in a way that the clinician could immediately tell how crucial the alert is and what the next step should be in the clinical process.  

For example, if a patient was currently on one drug and the prescriber wanted to start a second drug, such as an antidepressant. There are a lot of options for antidepressants, but Bates said some of them could create an adverse reaction.

Ultimately, the clinician wants to know, “what are the antidepressant choices that would not cause that problem?”

“The solution is anticipating what the clinician's next move might want to be,” Bates explained. “It doesn't tell the clinician which antidepressant to prescribe, but it does tell the user of the available ones, which ones are not going to cause a problem given the medications that the patient's already on.”

“It also gives patient-specific recommendations,” he added. “It takes into account the entire context of the patient, what their prior allergies are, what their lab tests show, what other drugs they're on, and so on. The recommendations are very much personalized for that individual.”

Following the study results, Bates said he is excited to fully integrate the alert solution into the EHR system, anticipating approval from the Brigham in the near future.

“The way myself and other leaders in the alert space think about it is, EHR alerts should fire when the clinician needs to be shown something,” Bates said. “Having the clinician change what they do at least a third and ideally a half the time. If I'm going to show you something and there's less than a one in three chance that you're going to change what you do, it's probably not worth interrupting you.”

Less is sometimes more, and this is especially true with EHR alerts.

“You might want to show something non-interruptive for some things that are less likely to change things than that. But that's the consensus. There's no absolute right answer to how each provider should integrate or optimize alerts,” Bates concluded.