Alert Fatigue: Risks & Solutions
- Nearly three-quarters of opioid allergy alerts may be unnecessary, according to new data, prompting calls for improved decision support tools adn better opioid management.
- The multi-institutional study, conducted at University of Colorado Health (UCHealth) and Mass General Brigham (MGB), reviewed over 700,000 alerts generated over a year.
- rachel Wasserman, lead author from Brigham and Women’s Hospital, noted that most overridden alerts stemmed from symptoms like nausea or itching, rarely indicative of a true immune response....
Clinicians are overriding opioid allergy alerts at an alarming rate—nearly 72%—due to a perilous phenomenon: alert fatigue. This News Directory 3 report dives into a recent study revealing the risks of excessive alerts in electronic health records. Many alerts are triggered by mild reactions, not true allergies, thus potentially endangering patient safety. Discover how redesigning clinical decision support tools and integrating patient-specific data can reduce the number of unnecessary interruptions. Actionable insights include the importance of distinguishing immune versus non-immune responses and tailoring alerts accordingly. Learn why the study’s authors are urging healthcare professionals to build more adaptable alert systems to improve opioid management and overall patient well-being. Discover what’s next …
High Override Rate for Opioid Allergy Alerts Signals Alert fatigue
Updated May 29, 2025
Nearly three-quarters of opioid allergy alerts may be unnecessary, according to new data, prompting calls for improved decision support tools adn better opioid management. A study published in BMJ Health & Care Informatics revealed that clinicians overrode 71.8% of opioid drug allergy alerts, suggesting an excess of interruptive alerts in electronic health records (EHRs).
The multi-institutional study, conducted at University of Colorado Health (UCHealth) and Mass General Brigham (MGB), reviewed over 700,000 alerts generated over a year. Researchers found that many bypassed alerts involved reactions considered mild or not genuinely allergic.
Dr. rachel Wasserman, lead author from Brigham and Women’s Hospital, noted that most overridden alerts stemmed from symptoms like nausea or itching, rarely indicative of a true immune response. This contributes to alert fatigue, posing a threat to patient safety.
The study indicated that nearly three-quarters of overridden alerts involved low or medium reaction severity. Only 29.3% were linked to genuine immune-mediated hypersensitivity. Codeine and morphine were the most frequently listed allergens, often overridden when ordering oxycodone or hydromorphone.
Dr. david Bates, senior author and clinical informatics expert, warned that frequent exposure to meaningless alerts increases the likelihood of missing critical ones.
“Most of these overridden alerts were triggered by reactions such as nausea or itching—symptoms that are rarely indicative of a true immune response,” said lead author Rachel Wasserman, MD, of Brigham and Women’s Hospital. “This creates a situation where alert fatigue becomes a real threat to patient safety.”
Researchers suggest redesigning clinical decision support (CDS) tools to account for reaction severity and type, potentially cutting interruptive alerts by 46.4%.They propose a tiered model distinguishing between immune and non-immune responses,emphasizing interruption only when truly warranted.
The study also highlighted the potential of integrating patient-specific context, such as prior medication tolerance, into alert logic. A notable number of overridden alerts were linked to vague or non-specific reactions.
“If a patient has tolerated the drug before, or if their reaction was nausea, an interruptive alert may not be clinically necessary,” said co-author Kimberly Blumenthal, MD. “There’s an possibility here to make CDS safer and smarter.”
What’s next
the principles outlined in the report could serve as a model for rethinking alert structures in other high-volume drug categories. The authors urge vendors and EHR developers to provide more versatility in configuring alert behaviors and to support learning-based systems that adapt over time, ultimately improving opioid management and patient safety.