Automated adverse event detection collaborative: electronic adverse event identification, classification, and...

Area of Science:

• Pediatric Healthcare Quality Improvement
• Health Informatics
• Patient Safety Research

Background:

• Voluntary incident reporting systems underestimate medical errors.
• Electronic Health Records (EHRs) offer a data-rich source for detecting adverse events.
• Automated adverse event detection is efficient and cost-effective in hospitals.

Purpose of the Study:

• Describe automated adverse event detection processes.
• Report early results from the Automated Adverse Event Detection Collaborative (AAEDC).
• Identify commonalities and differences between two participating organizations.

Main Methods:

• Retrospective observational study comparing automated adverse event detection systems at two academic children's hospitals.
• Utilized EHR data triggers to identify potential adverse events.
• Clinical investigators manually reviewed patient records to confirm events, preventability, and harm level.

Main Results:

• Analyzed data from July 2006 to October 2010.
• Adverse event triggers for opioid/benzodiazepine toxicity and IV infiltration showed high positive predictive values (47%-96%).
• Detected 3,264 adverse events, with 57.3% being preventable; only 15.1% were reported voluntarily.

Conclusions:

• EHR data aggregation and analysis are valuable for detecting and understanding adverse events.
• Comparing and selecting optimal electronic trigger methods aids in recognizing adverse event trends.
• Automated detection facilitates process redesign and quality improvement initiatives.