Project Details -
Completed
-
Grant NumberR21 HS023390
-
AHRQ Funded Amount$300,000
-
Principal Investigator(s)
-
Organization
-
LocationBaltimoreMaryland
-
Project Dates09/30/2014 - 09/29/2017
-
Population
-
Health Care Theme
Current clinical guidelines for using genomic data are inadequate for physicians to make informed health decisions. For example, physicians agree that genetic variations may influence drug response, but only a small fraction feel adequately informed about pharmacogenomic testing results. The challenges of understanding and interpreting genomic data are compounded by the demands of clinical practice. Clinical decision support (CDS) embedded into electronic health records (EHRs) can facilitate the appropriate use of genomic applications. However, CDS capabilities of existing systems are not robust enough for distilling large quantities of genomic data that is tailored to physician requirements.
To address the need to effectively communicate genomic data, this project leveraged existing clinical practice guidelines, EHRs, and academic partnerships to develop a genomic CDS (gCDS) engine and software application. Researchers aimed to use the gCDS engine and application to streamline the work required by stakeholders to use genomic test results in healthcare decisions.
The specific aims of the project were as follows:
- Employ user-centered approaches to design and develop a prototype CDS for effectively communicating genomic data.
- Formalize genomic knowledge for use by the CDS system.
- Integrate the finalized CDS with the EHR.
Researchers conducted focus groups with stakeholders from two personalized medicine programs at the University of Maryland and used the results to inform the design of a prototype gCDS engine for effectively communicating data to physicians via an EHR. They also translated a Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline into a format that is interpretable by computers and mapped the genomic information to standardized EHR data elements. They designed the gCDS engine to parse files from specialty genomics labs and to associate actionable variants, which contain chromosome and position information, with clinical guidance.
The primary outcome of this effort was translating the CPIC guideline into standardized concepts using a method that may be generalized for other clinical practice guidelines. They also created two new services to enable revising gCDS content based upon changes to the clinical guidelines. First, there was the implementation of the gCDS as a new web service that interfaces with a commercial EHR, with an "Add To Problem List" functionality. The team is also developing a service that populates the content of "Best Practice Alerts" in the EHR. Promising future work includes demonstrating and assessing the performance of an end-to-end system that includes the gCDS engine and the vendor's web services.