Medication Alerts in EMRs Are Especially Prone to Creating “Alert Fatigue”

Posted on April 6, 2012 I Written By

Anne Zieger is veteran healthcare consultant and analyst with 20 years of industry experience. Zieger formerly served as editor-in-chief of and her commentaries have appeared in dozens of international business publications, including Forbes, Business Week and Information Week. She has also contributed content to hundreds of healthcare and health IT organizations, including several Fortune 500 companies. Contact her at @ziegerhealth on Twitter or visit her site at Zieger Healthcare.

New research is drawing to attention to a critical subset of  “alert fatigue,” a problem which has dogged medical devices since well before EMRs/EHRs rolled into the picture.  The research, which appears in this month’s issue of the International Journal of Medical Informatics, finds that providers are becoming particularly numb to the high volume of EMR-based medication alerts they get.

To draw this conclusion, researchers followed 30 doctors, nurse practitioners and pharmacists as they treated 146 patients in outpatient settings run by the Indianapolis VA center. The study took place from August 2008 to August 2009.

After watching providers for a year, researchers came up with several different features characteristic of a poorly designed alert:

*  The system generates too many alerts
*  The alert provides far more information than the clinician needs
*  The system produces alerts that appear to go against standard clinical practice
*  The alert doesn’t apply to the patient
*  The alert isn’t needed (such as a warning about a drug a patient has already received and responded to safely)

Another failing found by researchers — which seems particularly acute in my opinion — is that alerts are typically designed to meet the needs and training of pharmacists, even though physicians and nurse practitioners were doing the prescribing.

Yet another issue that stood out was timing. On the one hand, alerts often interrupted the prescribing process, and wouldn’t go away until addressed.  Meanwhile, other alerts disappeared once addressed, and couldn’t be retrieved once they’ve left the screen.

As most readers know, I’m not a clinician, so I’m just reacting to the research data, but the situation sounds pretty dire. If an EMR is guilty of half of these medication alert “sins,” much less all of them, I’d bet it increases the potential for patient harm significantly.

Maybe this is a good place to start addressing the kind of potentially harmful design flaws identified by NIST. After all, virtually every patient takes meds, and  most every clinician touches them. Why not address some of these nagging problems before someone gets hurt?