InSPECting Alerts to Improve Patient Care

Published December 18, 2012 by Sarah Kelly


Photo of Allison McCoy, Ph.D.Allison McCoy, Ph.D., assistant professor of biomedical informatics at The University of Texas School of Biomedical Informatics at Houston, received KL2 funding for Young Clinical and Translational Investigators from the UTHealth Center for Clinical and Translational Sciences to evaluate and improve alerts in hospital and ambulatory settings.

Her project seeks to improve patient outcomes by effectively evaluating the appropriateness of clinical decision support alerts and responses and ultimately creating an alert evaluation dashboard called InSPECt: Interactive Surveillance Portal for Evaluating Clinical Decision Support.

“Many of the alerts that I am evaluating deal with drug interactions of some sort. For instance, a patient’s kidney function may be worsening, and some drugs could decrease kidney function even further,” said McCoy. “Physicians can order a medication, and if there’s any interaction an alert will pop up.”

The most commonly implemented computerized alerts in clinical decision support systems provide dosing guidance or prompt clinicians about drug-allergy, drug-drug, and drug-disease warnings.

“Physicians know about these interactions, but when they’re being inundated with information and tasks, they may miss a new lab result or other information. That’s why it’s vital to incorporate effective alerts,” said McCoy.

Evaluation and appropriateness of alerts and clinical responses are not new concepts. However, traditional non-computerized methods are time and labor intensive and consistency across institutions is challenging.

“Clinical decision support systems are being implemented all over the U.S. to meet meaningful use requirements, save money and improve patient care,” said McCoy. “Sending alerts through this system is more efficient than previous methods of critical information delivery. However, issues like alert overrides need to be evaluated.”

Alert overrides can be an issue when trying to improve patient outcomes. Physicians may override or ignore alerts for many reasons, including the transmission of repeated, irrelevant alerts, which can result in alert fatigue. If over time physicians habitually override alerts, they may be ignoring relevant alerts that warn of potential patient safety issues.

Sometimes overriding an alert is appropriate like determining whether the benefits of administering a needed drug to a patient outweigh its costs.

“With this grant, I’m aiming to identify factors contributing to alert inappropriateness and develop more specific alerting systems through semi-automated evaluation methods,” said McCoy. “This research can transform alert evaluation processes across healthcare settings, leading to improved clinical decision support, reduced alert fatigue and better improved patient care.”