Abstract
Background
Despite the proven usefulness of appropriate clinical decision support (CDS) alerts,
many CDS systems fire excessive, clinically irrelevant alerts that are often ignored
by clinicians. We have developed a method to suppress false-positive alerts based
on prior drug tolerance but encountered substantial barriers to integrating the method
into widely adopted commercial electronic health record (EHR) systems.
Objectives
This study aimed to describe the challenges faced while attempting to integrate our
method into the CDS infrastructure of two commercial EHR systems and provide recommendations
for future research and CDS design.
Methods
Using a multifaceted approach, we investigated (1) the use of emergent CDS standards
(e.g., CDS Hooks) to create a scalable solution to augment off-the-shelf EHR-based
alerts with patient-specific custom alerts, (2) customize CDS rules of commercial
medication knowledge bases (MKBs) to reduce false-positive alerts, and (3) manually
inactivate allergy documentation in patients with prior drug tolerance.
Results
We were unable to implement the standards-based approach because support for CDS Hooks
was found to be tailored to specific scenarios that involve the creation of new drug
allergy alerts (DAAs) but not the suppression of vendor-supplied DAAs. Likewise, we
were unable to suppress alerts imported from MKBs into the EHR systems investigated
because these systems do not support discrete clinical documentation changes that
drive DAAs. Lastly, we determined that although manually inactivating allergy documentation
in patients with prior drug tolerance is possible, doing so requires the impractical
solution of creating and maintaining individual rules for each drug at the ingredient
level.
Conclusion
We describe the barriers that precluded implementation of a novel method to suppress
clinically irrelevant CDS alerts in two commercial EHR systems. Overcoming these barriers
will require a more flexible CDS infrastructure, as well as collaboration and shared
responsibility across diverse stakeholders.
Keywords
clinical decision support - alert fatigue - alert burden - electronic health records
- interoperability
