An electronic health record's clinical decision support (CDS) tool in a hospital's emergency department issued so many unnecessary and clinically inconsequential alerts relating to opioids that providers found them "overwhelming" and contributing to alert fatigue, according to a new study in the Annals of Emergency Medicine.
The researchers, from the University of Colorado School of Medicine and elsewhere, studied the characteristics of CDS alerts triggered when opioids were prescribed in the emergency department of a large academic medical center that uses a commercial EHR. They found that the alerts were highly sensitive but low on specificity, and triggered so often that they interrupted workflow.
Providers sorted through 4,692 alerts to avert 38 adverse drug events. In other words, to prevent one adverse drug event, the providers had to deal with more than 123 unnecessary alerts. Moreover, 98.9 percent of the opioid alerts did not result in an actual or averted adverse drug event; and 96.3 percent of opioid alerts were overridden, a higher percentage than overrides for non-opioids, indicating that many providers determined that the override for opioids would be low risk.
While 14 of the 4,581 patient records reviewed had an adverse drug event, eight due to opioids, none were prevented by the CDS alert in the EHR.
The researchers suggested that alerts be redesigned to take a more tiered approach, so that less clinically significant ones would be more non-intrusive, and only the most specific and critical ones require a hard stop.
"We need to improve the 'signal to noise' ratio of these alerts, especially in the chaotic environment of the emergency department," Emma Genco, the lead study author, said in a statement. "Interruptions are already a significant fact of life in emergency departments, which is why we need to eliminate the meaningless ones."
Electronic CDS tools have been found to greatly improve safety in the emergency department and elsewhere. However, they backfire when they're so intrusive that they overload providers. The alerts appear to work best when they are customized to be more useful.
To learn more:
- read the announcement
- here's the study