Implementation of DDInteract: A Shared Decision Making Tool for Anticoagulant Drug-Drug Interactions

Patients receiving anticoagulant medications often have additional comorbidities requiring other medications, which increases the risk of bleeding due to drug-drug interactions (DDIs). Existing electronic health record (EHR) alert systems can provide warnings about potential DDIs; however, these systems are often text-based and do not use contextual, patient-specific data, which results in clinically irrelevant alerts that are overridden most of the time.

The researchers have previously developed and validated a shared decision making (SDM) tool called DDInteract that brings context and relevancy to anticoagulant DDIs alerts. DDInteract includes a visual display of risk, a patient-specific risk profile, patient education information, and a treatment decision tree. It was developed using SMART on FHIR standards, making it vendor-agnostic, shareable, and interoperable.

This current research will implement DDInteract at three university-based healthcare systems: University of Utah, University of Colorado, and Vanderbilt University, who all use the same commercial EHR but differ in geography, patient demographics, staff, and workflows. Facilitators and barriers will be noted during implementation that will inform an implementation guide.

The specific aims of the research are as follows:

• Examine contextual factors and define relevant implementation strategies to facilitate integration of DDInteract within three University-based healthcare systems.
• Evaluate DDInteract in production and real-world situations to identify barriers and facilitators to adoption and use of DDInteract and develop solutions to overcome those barriers.
• Conduct a cluster randomized, multisystem trial to evaluate the effectiveness of DDInteract to mitigate exposure to drug interactions involving oral anticoagulants.

For each of the three organizations, an implementation plan will be developed that is specific to that healthcare organization in terms of their workflow, information needs, and EHR governance. A series of cognitive work analyses will be done to determine where in the clinician workflow DDInteract should be placed. Options include launching the tool upon chart open during a visit, before a visit, or integrated as an interruptive or non interruptive alert. Implementation options also exist around when patients may access the tool to improve SDM, such as only during the visit, through the patient portal, or after a visit, allowing patients to reconsider risks and alternatives following the visit. A randomized implementation trial will then be conducted to measure the effect of DDInteract on exposure to DDIs between oral anticoagulants and risk of bleeding.

Throughout implementation planning and rollout, stakeholders, including providers and patients, will be engaged via interviews and focus groups. The research will also include an expert panel with recognized expertise in implementation science, drug interactions, medication informatics, and clinical decision support to review the design and technical documentation, review results, and help guide implementation. DDInteract and its implementation guide are currently available at CDS Connect; at the conclusion of the research, these artifacts will be updated with the information gleaned from the research.