A Clinical Case of Electronic Health Record Drug Alert Fatigue: Consequences for Patient Outcome

Department of Biomedical Informatics, Stanford University, Stanford, California
PEDIATRICS (Impact Factor: 5.3). 05/2013; 131(6). DOI: 10.1542/peds.2012-3252
Source: PubMed

ABSTRACT Despite advances in electronic medication order entry systems, it has been well established that clinicians override many drug allergy alerts generated by the electronic health record. The direct clinical consequences of overalerting clinicians in a pediatric setting have not been well demonstrated in the literature. We observed a patient in the PICU who experienced complications as a result of an extended series of non-evidence-based alerts in the electronic health record. Subsequently, evidence-based allergy alerting changes were made to the hospital's system. Incorporating clinical evidence in electronic drug allergy alerting systems remains challenging, especially in pediatric settings.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Reduction of unnecessary head CTs in patients with mild traumatic brain injury (MTBI) was recently endorsed by ACEP in the “Choosing Wisely” campaign. We examined the impact of computerized clinical decision support (CDS) on head CT utilization in MTBI ED visits.Methods We conducted a two-year cohort study at our Level-1 trauma center, and compared our results with the National Hospital Ambulatory Medical Care Survey from 2009–2010. All adult patients discharged from the ED with MTBI-associated diagnoses were included. After a baseline observation period at our institution, real-time CDS was implemented. Based upon the clinical history entered, low utility orders triggered an alert to clinicians, suggesting imaging studies might not adhere to evidence-based guidelines. Clinicians could cancel the order or ignore the alert. Primary outcome was intensity of head CT use in MTBI ED visits. Secondary outcomes included rates of delayed imaging, and delays in diagnosing radiologically significant findings. Chi-square and logistic regression, and process control chart assessed pre- and post-intervention differences.ResultsIn study patients, 58.1% of MTBI-related visits resulted in head CTs pre-intervention vs. 50.3% post-, (13.4% relative decrease, p = 0.005); a change not detected in controls (73.3% vs. 76.9%, p = 0.272). Study cohort patients not receiving a head CT during their index visit were neither more nor less likely to receive one in the subsequent seven days (6.7% pre- vs. 9.4% post-intervention, p = 0.231). Rates of delayed diagnosis of radiologically significant findings were unchanged (0% vs. 0%).Conclusions Evidence-based CDS can reduce low utility imaging for MTBI.
    American Journal of Emergency Medicine 11/2014; 33(3). DOI:10.1016/j.ajem.2014.11.005 · 1.15 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this study was to compare the effectiveness of two types of real-time decision support, an interrupting pop-up alert and a noninterrupting dynamically annotated visualization (DAV), in reducing clinically inappropriate diagnostic imaging orders. Alerts in electronic health record software are frequently disregarded due to high false-alarm rates, interruptions, and uncertainty about what triggered the alert. In other settings, providing visualizations and improving understandability of the guidance has been shown to improve overall decision making. Using a between-subject design, we examined the effect of two forms of decision support, alerts and DAVs, on reducing the proportion of inappropriate diagnostic imaging orders for 11 patients in a simulated environment. Nine attending and 11 resident physicians with experience using an electronic health record were randomly assigned to the form of decision support. Secondary measures were self-reported understandability, algorithm transparency, and clinical relevance. Fewer inappropriate diagnostic imaging tests were ordered with DAVs than with alerts (18% vs. 34%, p < .001). The DAV was rated higher for all three secondary measures (p < .001) for all participants. DAVs were more effective than alerts in reducing inappropriate imaging orders and were preferred for all patient scenarios, especially scenarios where guidance was ambiguous or based on inaccurate information. Creating visualizations that are permanently displayed and vary in the strength of their guidance can mitigate the risk of system performance degradation due to incomplete or incorrect data. This interaction paradigm may be applicable for other settings with high false-alarm rates or where there is a need to reduce interruptions during decision making. © 2015, Human Factors and Ergonomics Society.
    Human Factors The Journal of the Human Factors and Ergonomics Society 05/2015; DOI:10.1177/0018720815585666 · 1.29 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To establish preferred strategies for presenting drug-drug interaction (DDI) clinical decision support alerts. A DDI Clinical Decision Support Conference Series included a workgroup consisting of 24 clinical, usability, and informatics experts representing academia, health information technology (IT) vendors, healthcare organizations, and the Office of the National Coordinator for Health IT. Workgroup members met via web-based meetings 12 times from January 2013 to February 2014, and two in-person meetings to reach consensus on recommendations to improve decision support for DDIs. We addressed three key questions: (1) what, how, where, and when do we display DDI decision support? (2) should presentation of DDI decision support vary by clinicians? and (3) how should effectiveness of DDI decision support be measured? Our recommendations include the consistent use of terminology, visual cues, minimal text, formatting, content, and reporting standards to facilitate usability. All clinicians involved in the medication use process should be able to view DDI alerts and actions by other clinicians. Override rates are common but may not be a good measure of effectiveness. Seven core elements should be included with DDI decision support. DDI information should be presented to all clinicians. Finally, in their current form, override rates have limited capability to evaluate alert effectiveness. DDI clinical decision support alerts need major improvements. We provide recommendations for healthcare organizations and IT vendors to improve the clinician interface of DDI alerts, with the aim of reducing alert fatigue and improving patient safety. © The Author 2015. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved. For Permissions, please email:
    Journal of the American Medical Informatics Association 03/2015; DOI:10.1093/jamia/ocv011 · 3.93 Impact Factor