Electronic Health Reporter

George Robinson, RPh, senior product manager, First Databank.

Approximately $20 billion is lost annually in the United States because of medication errors, with the average hospitalized patient subject to at least one medication mistake per day.Alert fatigue is often cited as a reason for these errors—even though alerts generated by clinical decision support (CDS) systems call attention to important information (such as potential drug interactions), excessive alerts wear clinicians down, resulting in boy-who-cries-wolf scenarios. The result: clinicians instinctively override the alerts instead of implementing an override monitoring plan.

Consider the following:

• In 2009, researchers at the Boston-based Beth Israel Deaconess Medical Center and the Dana-Farber Cancer Institute looked at the safety alerts generated by 2,872 clinicians through 3.5 million electronic prescriptions over a nine-month period. Of the 233,537 alerts, 98 percent were drug-drug interaction issues, and more than 90 percent were overridden.

• A more recent 2013 study, published in the Journal of the American Medical Informatics Association, showed improved override rates with only about half of alerts overridden by providers, with half of those overrides classified as appropriate. Authors concluded that further refinement of these alerts could improve relevance and reduce alert fatigue.

A Driver in Need of a Clearer View

The afore-mentioned studies conclude that clinicians are indeed overriding medication alerts at alarming rates. Although the industry has made significant progress in addressing alert fatigue during the time the data from these studies was being analyzed, these studies clearly support what most healthcare professionals already suspect: The practice of ignoring and overriding medication alerts is widespread and can potentially lead to undesirable consequences.

When I think about the “alert fatigue” issue, I believe we are so distracted by the onslaught of alerts that we miss the significant problems that appear right in front of us. I liken it to the experience of driving during a blinding snowstorm on a dark winter’s night, with thousands of swirling snowflakes illuminated by the car’s headlights. This can create a hypnotizing effect on the driver, distracting him so much that he can’t easily see the hazards of the road. Growing up in Minnesota, I learned to turn on the “low beam” headlights to dim the brightness of the snowflakes, giving me a clearer view of the traveled road.

So, how is this metaphor connected to alert fatigue? By increasing the specificity of alerting and improving the user interface display of alerts, medication alert systems become akin to low beam headlights that diminish the snowflake distraction, giving the driver a clear line of sight to the risks presented during the journey home.

The mission is to make CDS work as it is intended to—but making alerts less of a nuisance and more of a necessity is far from simple. The big hurdle for healthcare organizations: How can we implement CDS that offers just the right level of alerts? It’s a not-too-much, not-too-little challenge that stretches far beyond the anxiety that was experienced by a porridge-seeking Goldilocks.

Healthcare leaders need to start examining why it is so difficult to find that “just right” level of alerts—and then, perhaps more importantly, find a way to implement a system that optimally leverages CDS at the point of care while providing warnings that actually mean something to individual clinicians in specific care settings. An easier-said-than-done proposition if ever there was one.

Three-pronged Approach to Clearing View

There are layers of complexity with alert management, from differences between patients, to varieties of care settings and the role of individual clinicians. To remedy the spellbinding effect that the snowstorm of alert fatigue, the answer lies in how we are able to more clearly illuminate the journey. Clarity comes in the form of deployment of CDS in the way it is intended. By processing additional clinical context and configuring the presentation of alerting content in a manner specific to the unique role of the clinician, the effectiveness of alerting can be improved in a manner that more clearly displays medication related risks to the patient.

The issue of alert fatigue can be addressed by a three-pronged strategy:

Regularly fine-tune drug knowledge—For drug-drug interactions, enhanced data models should identify potential clinical consequences associated to the drug therapy with links to labs, drugs and demographic information that serve as risk factors. For example, within drug allergy screening, the processing of a coded reaction (i.e., nausea, rash, shortness of breath) may be included as a variable for triggering broad hypersensitivity based allergen group and cross-sensitivity alerts for related medications in contrast to simply warning against the repeat use of a medication for which the reported reaction is considered an “intolerance.”

Make interruptive alerts more meaningful—By implementing additional parameters and filters in drug knowledge, and only firing alerts relevant to the individual clinician treating a single patient at a unique moment in time, alerts can become customized and less likely to be overlooked. These first two prongs require both the drug content provider and the EHR system vendor to work in concert together to make this work seamlessly.

Enable clinicians to customize medication alert display based on local experience—As industry experts and healthcare organizations work together to alleviate alert fatigue, it’s clear that the answer is to create systems that take human behavior and supplemental patient data into account when writing rules that decide when and why an alert fires. Coupled with alert configuration tools and advanced user interface design (i.e., tiered alerts, dashboards), the presentation of relevant, meaningful medication management alerts can indeed be improved, and the patient’s health care journey can be more safely illuminated.