Arun Krishnaraj MD, MPH, Presenter: Research support, General Electric Company

Mitchell A. Harris PhD, Abstract Co-Author: Nothing to Disclose

Abraham Lin BS, Abstract Co-Author: Nothing to Disclose

Neeraj Joshi MS, Abstract Co-Author: Nothing to Disclose

Sarita Nair MS, Abstract Co-Author: Nothing to Disclose

Garry Choy MD, Abstract Co-Author: Nothing to Disclose

Michael Ethan Zalis MD, Abstract Co-Author: Research grant, General Electric Company

Utilization of Emergency Department (ED) Imaging is frequently driven by the difficulties of evaluating complex patients under time pressure and is exacerbated by incomplete awareness of past medical history. As a step toward reducing un-necessary imaging in this challenging environment, we developed an application to expedite summarization of relevant past medical history. Based on an ontology-driven EHR search system, known as QPID (Queriable Patient Inference Dossier), the application performs 74 automated structured and natural language searches customized for ED care. In order to validate the accuracy of this tool, we compared the performance of the automated searches against that of manual review by two clinicians.

A cohort of 500 consecutive patients who presented to the ED between January 1, 2010 and December 31, 2010 were selected for review. Automated search for all patients included in the cohort was performed on each of the 74 topics included in the application. Subsequent to this, an untimed manual review was conducted of 30 random patients for each search module utilizing any available electronic resource available within our healthcare enterprise which accessed the EHR. Data that was collected from this manual search was compared to the automated QPID generated data and Sensitivity, Specificity, Positive Predictive Value, and Negative predictive value were calculated with manual review of the EHR serving as the gold standard.

Average search time for the application to complete all 74 modules was 15 seconds (SD +/- 5 sec).  Two categories of results were obtained; laboratory data and free text search data (e.g. “Is there a history of PE in the last 10 years?”). For laboratory results, the automated QPID modules demonstrated a sensitivity of 97% and a specificity of 99% with calculated average PPV and NPV of 99% and 96%, respectively. For free text searches of the EHR, sensitivity was 98% and specificity was 93%, with calculated average PPV and NPV of 90%, and 98%, respectively.

EHR automated searches utilizing QPID are a rapid, accurate means by which ED physicians can gather salient clinical data on patients who present to the ED.

Incorporation of QPID, a validated EHR search tool which summarizes essential past medical history, may help reduce over utilization of imaging in the ED setting.

Krishnaraj, A, Harris, M, Lin, A, Joshi, N, Nair, S, Choy, G, Zalis, M, Optimizing the Use of Emergency Department Imaging: Validation of an Ontology-driven EHR Search System to Summarize Essential Past Medical History.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11004881.html