Researchers studied four CDSS malfunctions at Brigham and Women’s Hospital in Boston as case studies and conducted a survey of CMIOs to determine the frequency of malfunctions. The studied malfunctions were an alert monitoring thyroid function stopped working, an alert to screen children for lead stopped working, a software update caused a system to send false alerts and a malfunction in an external drug classification system inappropriately suggested certain drugs.
The study found the alerts to be widespread and occurring more frequently than previously believed. Additionally, researchers found existing detection systems did not detect malfunctions in CDSS alerts before reaching users.
Researchers suggested the malfunctions may be caused by changes in data codes or clinical terminology. Such changes are typically made when analysts workon different elements of an EHR system, and the changes do not get communicated to CDSS authors.
Additionally, certain alerts are dependent upon outside clinical infrastructures or modules. As related to the thyroid CDSS malfunction, when an identifier for a certain drug was changed in one system, it disrupted how the CDSS functioned.
“As CDSSs becomes more complex and widespread and clinicians increase their reliance on them, improved processes and tools for preventing and detecting CDSS malfunctions are essential,” the authors concluded.
More articles on health IT:
The guide to cringeworthy health IT conversations: 10 leaders on the terms they dread most
Dr. Ramanathan Raju speaks out on NYC Health + Hospitals’ Epic transition
To pay or not to pay ransom: A tale of two hospitals
