Meaningful Drug Interaction Alerts
Project Final Report (PDF, 323.31 KB) Disclaimer
Disclaimer
Disclaimer details
This research allows hospital systems to create contextualized drug-drug interaction alerts, which if implemented would result in fewer alerts and a reduction in alert burden with the potential of improving patient safety.
Project Details -
Completed
-
Grant NumberR01 HS025984
-
Funding Mechanism(s)
-
AHRQ Funded Amount$1,549,585
-
Principal Investigator(s)
-
Organization
-
LocationSalt Lake CityUtah
-
Project Dates05/01/2018 - 02/28/2023
-
Technology
-
Care Setting
-
Population
-
Type of Care
-
Health Care Theme
Drug-drug interactions (DDIs) are responsible for up to 14 percent of adverse drug reactions in hospitalized patients, are a major risk factor for hospitalization, and occur in up to 13 percent of elderly ambulatory patients. Clinical decision support (CDS) systems that trigger alerts for drug pairs that potentially interact in a patient’s medication regimen exist, but DDIs persist due to a lack of contextualization such as patient-specific data that would make such an alert clinically irrelevant. As such, the majority of DDI alerts are overridden by prescribers, representing a source of alert fatigue and a potential patient safety issue.
This research developed algorithms for eight key DDIs to inform the building of contextualized DDI alerts; developed and tested three DDI-CDS apps; and disseminated the information and algorithms via the DDI-CDS.org website, webinars, and publications.
The specific aims of the research were as follows:
- Design sharable evidence-based individualized DDI algorithms that capitalize on the wealth of patient data located within electronic health records (EHRs).
- Validate the function of newly designed DDI algorithms using EHR data.
- Conduct a prospective evaluation of DDI algorithms in a variety of healthcare environments, including ambulatory and institutional settings.
The researchers began by identifying DDI combinations to focus on by consulting with experts, conducting a literature review, analyzing EHR data on alerts and overrides, and analyzing data from the Food and Drug Administration’s Federal Adverse Event Reporting System. Eight DDIs were chosen that were overridden frequently by prescribers or had been identified by prescribers as being important. From this work, algorithms for each were developed that allowed for contextual alerting. The decision trees and computable knowledge artifacts were added to DDI-CDS.org, making them available to others for incorporation in their CDS systems. Three DDI-CDS apps were developed: one for the use of NSAIDS with warfarin, one for tizanidine and cytochrome P450 1A2 inhibitors, and the last for colchicine and cytochrome P450 3A4/p-glycoprotein inhibitors. The colchicine-specific interaction app was implemented at Vanderbilt University Medical Center, along with a custom drug interaction editor prototype.
The researchers estimated that using the eight algorithms would result in a 52.4 percent reduction in alerts. They also noted barriers to implementation of such DDI-CDS tools, including a desire for off-the-shelf solutions and solutions included within EHRs making local builds unnecessary; a lack of resources; maintenance concerns; and security. The researchers noted that future efforts should focus on involving vendors to increase adoption and implementation of such tools and creating the ability for their customers to easily modify and maintain DDI alerts.
Disclaimer
Disclaimer details