One way to minimize liability risk is to minimize the likelihood of adverse drug reactions. Electronic prescribing systems are supposed to accomplish that by warning of patient allergies or possible interactions with other medications.
But a recent study finds the systems are so hypersensitive that “most alerts may be more of a nuisance than an asset” (Arch Intern Med. 2009;169:305-311) (subscription required). After analyzing 233,537 alerts generated from 3.5 million prescriptions, a team of Harvard internists found clinicians overrode the warnings much more often than not and relied on their own judgments instead.
About 1 in 15 prescriptions (6.6%) generated an alert. Almost all (98.6%) were for potential interactions with medications the patient was already taking, but clinicians accepted only 9.2% of the warnings and either changed or canceled their treatment plans. The remainder were for potential allergies and were rejected in more than three out of four instances (77%).
“Too many alerts are generated for unlikely events, which could lead to alert fatigue,” observes lead author Thomas Isaac, MD, MBA, MPH, an internist at Beth Israel Deaconess Medical Center.
Prescribers accepted 2.2% to 43.1% of high-severity interaction alerts, depending on the classes of medications involved. They were less likely to accept these alerts if the patient had taken the medication before (odds ratio, 0.03).
For example, prescribers overrode 97.8% of high severity alerts concerning topical corticosteroids with anti-infectives and antihyperlipidemic combinations (e.g. topical betamethasone-clotrimazole and ezetimibe-simvastatin). The best high-severity acceptance rate (43.1%) was recorded for noncardioselective beta-blockers and macrolides, (e.g. sotalol and azithromycin).
“We need to find a way to help clinicians separate the proverbial wheat from the chaff,” explains Saul Weingart, MD, PhD, vice president for patient safety at Dana-Farber Cancer Institute and an internist at Beth Israel Deaconess. “Until then, electronic prescribing systems stand to fall short of their promise to enhance patient safety and to generate greater efficiencies and cost savings.”
Based on these findings, the study recommends that systems:
Reclassify severity ratings, especially for alerts that are often overridden;
Allow users to suppress alerts for medications a patient has already received; and
Customize alerts for the prescriber’s specialty.
The researchers also compiled a list of potentially dangerous interactions based on those alerts that most often changed the clinicians’ decision to prescribe.
“Electronic prescribing clearly will improve medication safety, but the full benefit will not be realized without the development and integration of high-quality decision support systems to help clinicians better manage medication safety alerts,” Dr. Weingart comments.