Detection of adverse drug events using an electronic trigger tool

Area of Science:

• Pharmacovigilance
• Health Informatics
• Clinical Pharmacy

Background:

• Adverse drug events (ADEs) pose a significant risk in healthcare settings.
• Electronic health records (EHRs) offer potential for automated ADE detection.
• Trigger tools are used to identify potential ADEs based on specific events or medications.

Purpose of the Study:

• To implement and refine an integrated electronic trigger tool for detecting ADEs.
• To improve the positive predictive value (PPV) of an EHR-based ADE detection system.
• To evaluate the tool's effectiveness using specific reversal agents: flumazenil, naloxone, phytonadione, and protamine.

Main Methods:

• A three-month prospective study at a large medical center.
• An EHR-based trigger tool alerted pharmacists to the administration of four reversal agents.
• Pharmacists reviewed cases in near real-time to identify actual or potential ADEs.
• Alerting criteria were progressively narrowed to enhance the tool's PPV.

Main Results:

• A total of 87 drug administrations were reviewed, identifying 27 ADEs.
• The PPV varied across study phases: 0.33 (phase 1), 0.21 (phase 2), and 0.36 (phase 3).
• False-positive alerts, particularly for phytonadione, contributed to the overall low PPV (0.31).

Conclusions:

• The electronic trigger tool demonstrated potential for ADE detection but requires further refinement.
• The overall PPV of 0.31 highlights challenges in accurately identifying ADEs using this method.
• Refinement strategies, including narrowing alerting criteria, showed some impact on PPV.