Electronic health records may be used to capture a broad range of information--including clinical notes, lab test results, diagnostic studies, and demographic data--but according to a recent study in the Journal of the American College of Radiology, much of this data can be difficult to review and access efficiently.
This can be problematic in the emergency department, where an incomplete awareness of medical histories, lab and imaging tests can result in an overutilization of imaging because of time pressures involved in treating patients.
In fact, according to the authors, this is one of the reasons why the percentage of patients who visit EDs and undergo imaging has grown dramatically over the last several decades, particularly in relationship to other settings such as outpatient facilities, inpatient facilities and private offices.
ED doctors understand there is a problem with imaging overutilization in their departments. Last fall, FierceMedicalImaging reported on a survey presented at the American College of Emergency Physicians Research Forum in which surveyed emergency physicians agreed that imaging is being overutilized in EDs. The physicians agreed that they would welcome clinical decision support as a way to reduce inappropriate imaging.
In response, according to the authors--led by Arun Krishnaraj of the University of Virginia's department of radiology and medical imaging--clinicians and software developers at the institution have turned to the Queriable Patient Inference Dossier (QPID), created in 2005 at Massachusetts General Hospital.
QPID is a programmable, ontology-driven semantic search application that extracts data from the EHR and indexes the information for a searcher. If, for example, a user wants to associate the term "abscess" with other EHR data and text such as "fever" or "white blood cell count," he can execute an "abscess" query from the systems query library.
The QPID emergency department application included 74 query topics important for ED screening and management of patients. This study included 500 patients with clinical documents containing both structured and unstructured data in the EHR.
The researchers executed these 74 queries on each of the patients.
They found that the application was able to complete all 74 searches on each patient in a mean search time of 15 seconds. The queries for structured data demonstrated a positive predictive value of 87 percent and a negative predictive value of 86 percent, and a positive predictive value of 75 percent and a negative predictive value of 88 percent for unstructured data.
"We hope to demonstrate in future studies that this tool will influence the rate and appropriateness of imaging utilization as well as other healthcare resources," the authors wrote.
To learn more:
- see the study in the Journal of the American College of Radiology