JACR: New search tool useful for extracting data from EMR
Researchers at Massachusetts General Hospital in Boston have developed a search tool called the Queriable Patient Inference Dossier (QPID) that helps radiologists and other physicians extract useful data from a patient’s EMR in a timely and efficient manner, according to an article in the August issue of the Journal of the American College of Radiology.

“Even in its simplest implementation, the presence of an EMR system presents considerable challenges to the radiologist,” wrote lead author Michael Zalis, MD, director of QPID informatics at the department of radiology at MGH. “For example, radiologists commonly encounter each patient with little prior familiarity with the patient’s clinical situation. As a result, the time and effort required to retrieve, review and assimilate EMR data relevant for the case at hand becomes an important consideration for use of EMR in busy clinical practice,” said Zalis.

To address this issue, in 2005, researchers at MGH initiated the development of the programmable search system QPID for their institution’s EMR. QPID is a search engine that serves as an adjunct to the hospital’s EMR system; it was developed separately from the EMR and operates in a read-only fashion in relation to it.

“Thus QPID is not a source of new EMR data, but serves as a method to extract useful patterns of EMR data from the separately curated clinical data repositories at our institution,” said Zalis. QPID currently serves 500 registered users at Massachusetts General and posts 7,000-10,000 pages of medical record data daily, according to the article.

“The potential impact of advanced EMR search tools is by no means limited to radiology and in fact many departments in the hospital and outpatient clinic may benefit from these capabilities,” concluded Zalis. “In our own institution, with the QPID search system, we have catalyzed a growing base of enthusiastic users, many of whom have contributed their own insights and content to the system’s catalogue of search modules, each of which is potentially applicable at more than one site.”

For more information on QPID and natural language processing, read Health Imaging & IT’s April article “The Stepping Stone to Structured Reporting: What Does Natural Language Processing Mean for Radiology.”