JACR: Advanced EMR search tools could improve radiology services
Automated retrieval of topic-specific and event-specific data according to schedule and care unit census can improve the efficiency of services delivered in radiology and may potentially improve the quality and safety of those services, according to an article in this month's Journal of the American College of Radiology.
Michael E. Zalis, MD, and Mitchell A. Harris, PhD, from the department of radiology at Massachusetts General Hospital (MGH) in Boston, reviewed emerging functions of the EMR to focus on the use of advanced search technologies to automate extraction of information from the EMR to improve the safety and quality of clinical practice.
According to the authors, a simple EMR implementation is not without significant challenges to radiologists. Because radiologists commonly have little prior familiarity with the patient’s clinical situation, time and effort required to retrieve, review and assimilate relevant EMR data for the case at hand has become an important consideration for the use of EMRs in busy clinical practice.
“An EMR system is often hosted in its own hardware and software; hence, radiologists must log on to a completely separate EMR system and then manually review records to obtain the information they seek for the case at hand,” Zalis and Harris wrote. “This manual effort for EMR data retrieval is considerable, in some instances constituting up to 20 percent of the diagnostic effort expended for the interpretation of CT and MRI.” As more data becomes available, the volume of information to sift through becomes ever increasing in a time-limited encounter, the authors added.
“Medical data can be readily categorized into broad classes of data, such as laboratory results and specialty reports, and these broad categories are used in the display of information on all EMR systems. However, even this level of categorization is insufficiently precise to expedite synchronous, rapid search of EMR data," Zalis and Harris noted.
Developed in 2005, MGH's department of radiology designed QPID (Queriable Patient Inference Dossier) to serve as a search engine to MGH’s EMR system to aggregate EMR data across the multi-institution enterprise. It should be noted that QPID does not generate new EMR data but extracts patterns of EMR from separately curated clinical data repositories at MGH.
Since being deployed, QPID has been integrated into MGH's PACS so that with a single icon click (no separate log-on required) from the PACS interface, QPID opens in a pop-up window, presenting context-aware, searchable access to the current patient's entire medical record. QPID also performs rapid and full-text indexing of all the data gathered on a patient (stored in a local, temporary cache) and makes available to both software developer and clinical end user a growing set of natural language processing tools that range from the use Boolean tests and regular expressions to semantic filtration. Zalis and Harris also point out that QPID’s output is directed, as needed, to either a web browser for human review in an EMR portal or directly to external software packages via a web services call.
QPID currently serves 500 registered users at MGH and posts 7,000 to 10,000 pages of medical record data daily. “Rapidly following the introduction of QPID, there was a request by our staff for QPID search functions integrated into PACS and consequent widespread adoption of QPID for a large fraction of medical record search; we are now collecting prospective data to characterize the effect of advanced search on EMR search and examination interpretation times,” the authors wrote.
Zalis and Harris noted that in addition to improved time efficiency, QPID’s automated search modules, which can be activated automatically against each day’s interventional service schedule, can be used to guarantee that each staff member views a set of essential safety data, thereby potentially reducing the number of preventable errors associated with interventional procedures.
“The appropriate integration of search tools into clinical practice will require focus on the validation of these tools, a greater understanding of the full range of their performance characteristics and the continued need for diligence and redundancy in safety mechanisms,” concluded the authors adding that the potential positive impact of advanced EMR search tools is by no means limited to radiology.
Michael E. Zalis, MD, and Mitchell A. Harris, PhD, from the department of radiology at Massachusetts General Hospital (MGH) in Boston, reviewed emerging functions of the EMR to focus on the use of advanced search technologies to automate extraction of information from the EMR to improve the safety and quality of clinical practice.
According to the authors, a simple EMR implementation is not without significant challenges to radiologists. Because radiologists commonly have little prior familiarity with the patient’s clinical situation, time and effort required to retrieve, review and assimilate relevant EMR data for the case at hand has become an important consideration for the use of EMRs in busy clinical practice.
“An EMR system is often hosted in its own hardware and software; hence, radiologists must log on to a completely separate EMR system and then manually review records to obtain the information they seek for the case at hand,” Zalis and Harris wrote. “This manual effort for EMR data retrieval is considerable, in some instances constituting up to 20 percent of the diagnostic effort expended for the interpretation of CT and MRI.” As more data becomes available, the volume of information to sift through becomes ever increasing in a time-limited encounter, the authors added.
“Medical data can be readily categorized into broad classes of data, such as laboratory results and specialty reports, and these broad categories are used in the display of information on all EMR systems. However, even this level of categorization is insufficiently precise to expedite synchronous, rapid search of EMR data," Zalis and Harris noted.
Developed in 2005, MGH's department of radiology designed QPID (Queriable Patient Inference Dossier) to serve as a search engine to MGH’s EMR system to aggregate EMR data across the multi-institution enterprise. It should be noted that QPID does not generate new EMR data but extracts patterns of EMR from separately curated clinical data repositories at MGH.
Since being deployed, QPID has been integrated into MGH's PACS so that with a single icon click (no separate log-on required) from the PACS interface, QPID opens in a pop-up window, presenting context-aware, searchable access to the current patient's entire medical record. QPID also performs rapid and full-text indexing of all the data gathered on a patient (stored in a local, temporary cache) and makes available to both software developer and clinical end user a growing set of natural language processing tools that range from the use Boolean tests and regular expressions to semantic filtration. Zalis and Harris also point out that QPID’s output is directed, as needed, to either a web browser for human review in an EMR portal or directly to external software packages via a web services call.
QPID currently serves 500 registered users at MGH and posts 7,000 to 10,000 pages of medical record data daily. “Rapidly following the introduction of QPID, there was a request by our staff for QPID search functions integrated into PACS and consequent widespread adoption of QPID for a large fraction of medical record search; we are now collecting prospective data to characterize the effect of advanced search on EMR search and examination interpretation times,” the authors wrote.
Zalis and Harris noted that in addition to improved time efficiency, QPID’s automated search modules, which can be activated automatically against each day’s interventional service schedule, can be used to guarantee that each staff member views a set of essential safety data, thereby potentially reducing the number of preventable errors associated with interventional procedures.
“The appropriate integration of search tools into clinical practice will require focus on the validation of these tools, a greater understanding of the full range of their performance characteristics and the continued need for diligence and redundancy in safety mechanisms,” concluded the authors adding that the potential positive impact of advanced EMR search tools is by no means limited to radiology.