Healthcare Analytics, Population Health Management, Healthcare Big Data

Quality & Governance News

Why are EHR Alerts, Clinical Decision Support So Ineffective?

"Clinically inconsequential" EHR alerts are everywhere. How can these clinical decision support features become useful instead of burdensome?

By Jennifer Bresnick

- Backed by sophisticated big data analytics algorithms and advanced clinical know-how, EHR alerts are supposed to be the solution to the deadly problems of adverse drug events and inappropriate prescribing. 

EHR alerts and clinical decision support

Clinical decision support features like warnings for potential drug interactions and dosage reminders ought to produce significant benefits for patient safety, yet they consistently make industry lists of the most hazardous health IT tools on the marketplace.

A new study published in the Annals of Emergency Medicine this month adds to the evidence that EHR alerts and alarms may be doing more harm than good.  Frazzled physicians, besieged by more than one hundred unnecessary and clinically inconsequential alerts, override more than 96 percent of alarms related to opioid prescriptions, most of which would not have prevented an adverse event in the first place.

As EHR dissatisfaction and frustration with mandates like meaningful use continue to reach all-time highs, will developers and providers be able to overcome the workflow challenges that make EHR alarm fatigue such a worryingly common occurrence?

EHR alerts aren’t necessarily problematic in and of themselves.  When used correctly, drug safety dashboards and ADE alarms can improve patient safety and cut the workflow burdens for providers.  The trick is to make these flags and pop-ups intelligent enough to fire only when absolutely necessary to prevent harm.

READ MORE: QPID, Epic EHR Combine for Big Data Analytics at Partners Health

At the Marshfield Clinic, for example, a carefully tuned EHR alert system, bolstered by clinical analytics, helped cut down nearly 80 percent of possible situations where multiple prescriptions may have caused an adverse event.  Using big data analytics and historical EHR data, Marshfield Clinic developed a prioritized system of alerts that only stopped a provider’s workflow in the most serious cases. 

The system delivered less immediate alerts in a daily color-coded report, which the provider could use to educate themselves about potential problems or change a patient’s care plan when necessary.

In contrast, the emergency department systems studied by lead author Emma Genco, MS, of the University of Colorado School of Medicine in Denver, delivered hundreds of alerts that resulted in very little clinical gain.

Physicians routinely dismissed almost all of the opioid dosage and allergy alerts presented to them.  It may seem problematic that providers could be so cavalier about patient safety, but the study also indicated why this behavior may be occurring. 

A staggering 98.9 percent of these EHR alerts did not result in an actual or averted ADE.  Nor could the clinical decision support system have prevented any of the eight opioid-related ADEs that did occur among the more than 14,000 patients in the study sample. 

READ MORE: Smart Big Data is Key to Population Health, Value-Based Care

"Our electronic health record warning system on opioids is overwhelming providers with unnecessary and clinically inconsequential alerts," said Genco. "The danger here is that medical providers may develop 'alert fatigue,' leading to compromised patient safety. It is well established that clinical decision support prevents adverse drug events, but it is essential that alerting systems be refined to highlight only the clinically significant alerts."

Alerts and alarms aren’t the only technology flaws making workflows difficult in the emergency department.  Many ED information systems (EDIS) score poorly on a variety of process and usability problems, with close to 40 percent of providers in a recent Black Book poll stating that they are moderately or highly dissatisfied with the way their software works.

Approximately one in three large hospitals are hoping to replace their EDIS technology within the year, with productivity improvements, usability, and diagnosis enhancements ranking highly among the most-wanted features for new software.

But it is nearly impossible for healthcare organizations to acquire products that meet their needs.  EHRs are often geared towards the inpatient setting and lack specific features that can aid the process of triaging and treating cases in a rapid-fire setting like the ED. 

Despite a newfound emphasis on EHR usability within the developer community, little progress has been made to protect end-users from inadequately designed workflows and unnecessary interruptions.  EHR alert override rates for opioids have increased from 50 percent to 90 percent in the last twenty years, Genco and her team point out.

READ MORE: Lack of Talent, Direction Afflict Healthcare Data Analytics Plans

“For opioid orders, allergy, duplicate therapy, and duplicate drug order alerts make up almost three quarters of drug alerts,” the study says. “Additionally, opioid drug alerts were more likely than non-opioid ones to be overridden across most provider types and all alert types. This high override rate across the board is evidence that alert fatigue is associated with these interruptive alerts.”

Alarm fatigue has become a leading cause of physician burnout, other researchers have stated.  A 2014 article in the American Journal of Managed Care noted that some physicians may even quit their jobs if forced to use poorly optimized EHR alert systems.

“[EHR alerts] likely represent one of the most frustrating components of EHRs for providers,” the study says.  “Compared with paper communication systems, they are perceived to ‘increase the number of work items, inflate the time to process each, and divert work previously done by office staff to them.’”

Alerts are “excessive” for 87 percent of physicians, added a research letter from JAMA Internal Medicine.  Providers were presented with an average of 63 alerts per day. 

Those respondents who said they felt overwhelmed by the constant barrage of pop-ups and flashing lights were also more likely to acknowledge that they sometimes miss vital information and struggle to provide their care teams with critical information at the end of a shift or during a transition.

When physicians received a high number of insignificant alerts, they were more likely to become desensitized to all information of any importance, added lead author Hardeep Singh, MD, MPH.  To prevent “alert blindness” among busy clinicians, notifications should be tailored and prioritized in order to ensure that critical information isn’t being ignored.

"We need to improve the 'signal to noise' ratio of these alerts, especially in the chaotic environment of the emergency department,” Genco stated. “Interruptions are already a significant fact of life in emergency departments, which is why we need to eliminate the meaningless ones."

Developing a prioritization system for EHR alerts is a first step, Genco and her team said.  Creating a notification that is visible but not aggravating can also help to smooth the jagged nerves of struggling end-users.

“To increase selectivity and accuracy, future research should focus on testing a tiered approach to decrease the overall volume of interruptive alerts and increase specificity,” says the study. “This approach moves less clinically significant alerts to a non-interruptive format, and only the most critical, high-severity alerts result in hard stops or interruptive alerts.”

The subtle approach is an effective one, other organizations have found.  When community health clinics in New York City employed an alert system that pushed an unobtrusive immunization reminder in the corner of the computer screen, primary care providers increased their flu vaccination rate by nine percent.

Another study about HPV vaccinations found that giving the provider the option to mark an alert as resolved or otherwise inapplicable – such as when a patient is ineligible for the treatment or discussed but declined the option – helped young patients complete their vaccination series ten times more frequently. 

In the emergency department, where opioid allergy warnings reign supreme, “retaining only exact and base ingredient matches as interruptive alerts could decrease the alert volume without sacrificing sensitivity,” Genco suggests.

“If applied to our data set, this would decrease the volume of interruptive drug-allergy alerts by 85.5 percent without eliminating alerts that resulted in an actual or averted adverse drug event.”

“It is essential to refine alerting systems to highlight clinically significant alerts and eliminate inconsequential alerts, thereby preventing alert fatigue and maintaining patient safety.”

Continue to site...