Getty Images/iStockphoto

Users Override Clinical Decision Support Alerts at High Rates

Renal medication clinical decision support alerts were overridden 100 percent of the time, resulting in adverse drug events and clinician burden.

Users overrode all renal clinical decision support alert triggers in a recently implemented EHR system due to the high-level of CDS alert frequency. CDS overrides can lead to adverse drug events (ADEs) and clinician burden, according to a study published in the Journal of the American Medical Informatics Association (JAMIA).

The high volume of overridden CDS alerts proved the need to redesign or optimize medication-related CDS alerts that are associated with renal disease.

ADEs occur roughly 1.5 million times per year, according to the study authors. Some researchers say these injuries account for 5 to 17 percent of hospital admissions. However, the study authors said of those 1.5 million impacted patients, nearly 400,000 adverse drug events are considered preventable.

EHR-based CDS tools could help prevent ADEs. According to the study authors, medication-related CDS tools can reduce up to 81 percent of medication errors. Although most EHRs include CDS alerts, most are vendor-developed and the majority do not address patients with renal insufficiency. Roughly one-third of patients receive incorrect renal function doses.

“Medication-related CDS represents an effective way to reduce errors, and ADEs. However, this impact may be decreased or even extinguished if too many clinically inappropriate alerts are given,” explained the study authors. “This problem represents an important one in informatics today, as EHRs are now broadly implemented, and almost all are vendor-developed.”

From 2017 and 2018, researchers analyzed override rate frequency for all inpatients at Brigham and Women’s Hospital in Boston, Massachusetts who had a renal CDS alert trigger. They also analyzed the quality, appropriateness, and accuracy of alert overrides and the potential harm from overriding alerts in commercial EHR systems.

The research team identified 37,100 “dose change” and 5,095 “avoid medication” CDS alerts during this time period. However, 100 percent of each alert were overridden.

The research team classified 12.5 percent of “dose change” alerts as appropriate and 90.5 percent of overrides as appropriate. The researchers also classified 29.6 percent of “avoid medication” alerts as appropriate and 76.5 percent of overrides as appropriate.

Furthermore, the research team identified five adverse drug events, four of which were a result of an inappropriate overridden alert. To reduce the frequency of medication-related harm, these issues need to be addressed, wrote the study authors.

“The override rate was much higher than previous studies, but the appropriateness of alert overrides was consistent with previous studies,” explained the study authors. “These data demonstrate that there is substantial room for improvement in our current process in several areas, including the monitoring approach, accuracy of alerts, and how suggestions are delivered.”

Since the study, the research team said they shared the results with the health system’s health IT team who has made optimizations and improvements in alerts for the selected medications. The health IT team decreased the CDS alert threshold to reduce the number of triggered alerts.

The research team hypothesized several potential reasons why the current CDS alerts are less effective than if they had been in the homegrown EHR system:

  • No automatic calculation of the level of renal function
  • No automatic dose adjustment
  • No alternative recommendations
  • No patient-specific parameter considerations

“Future studies should focus on application of human factors principles in redesigning and implementing medication-related CDS alerts associated with renal insufficiency, with a goal of reducing the number of inappropriate alerts presented to providers by making alerts more patient-specific and clinically appropriate,” the research team concluded. 

Next Steps

Dig Deeper on Health IT optimization

Cloud Computing
Mobile Computing