Using Voogle to Search Within Patient Records in the VA Corporate Data Warehouse

Digitalization of patient-specific information over the past 2 decades has dramatically altered health care delivery. Nonetheless, this technology has yet to live up to its promise of improving patient outcomes, in part due to data storage challenges as well as the emphasis on data entry to support administrative and financial goals of the institution.^1-4 Substantially less emphasis has been placed on the retrieval of information required for accurate diagnosis.

A new search engine, Voogle, is now available through Microsoft Internet Explorer (Redmond, WA) to all providers in the US Department of Veterans Affairs (VA) on any intranet-enabled computer behind the VA firewall. Voogle facilitates rapid query-based search and retrieval of patient-specific data in the VA Corporate Data Warehouse (CDW).

Case Example

A veteran presented requesting consideration for implantation of a new device for obstructive sleep apnea. Guidelines for implantation of the new device specify a narrow therapeutic window, so determination of his apnea-hypopnea index (AHI) was critical. The patient had received care at more than 20 VA facilities and knew the approximate year the test had been performed at a non-VA facility.

A health care provider (HCP) using Voogle from his VA computer indexed all Veterans Information Systems and Technology Architecture (VistA) notes for the desired date range. The indexing of > 200 notes was completed in seconds. The HCP opened the indexed records with Voogle and entered a query for “sleep apnea,” which displayed multiple instances of the term within the patient record notes. A VA HCP had previously entered the data from the outside sleep study into a note shortly after the study.

This information was found immediately by sorting the indexed notes by date. The total time required by Voogle to find and display the critical information from the sleep study entered at a different VA more than a dozen years earlier was about 1 minute. These data provided the information needed for decision making at the time of the current patient encounter, without which repeat (and unnecessary) testing would have been required.

Information Overload

Electronic health records (EHRs) such as VistA, upload, store, collate, and present data in near real-time across multiple locations. Although the availability of these data can potentially reduce the risk of error due to missing critical information, its sheer volume limits its utility for point-of-care decision making. Much patient-specific text data found in clinical notes are recorded for administrative, financial, and business purposes rather than to support patient care decision making.^1-3 The majority of data documents processes of care rather than HCP observations, assessment of current status, or plans for care. Much of this text is inserted into templates, consists of imported structured data elements, and may contain repeated copy-and-paste free text.

Data uploaded to the CDW are aggregated from multiple hospitals, each with its own “instance” of VistA. Often the CDW contains thousands of text notes for a single patient. This volume of text may conceal critical historical information needed for patient care mixed with a plethora of duplicated or extraneous text entered to satisfy administrative requirements. The effects of information overload and poor system usability have been studied extensively in other disciplines, but this science has largely not been incorporated into EHR design.^1,3,4

A position paper published recently by the American College of Physicians notes that physician cognitive work is adversely impacted by the incorporation of nonclinical information into the EHR for use by other administrative and financial functions.²