Abstract P147: Telestroke Software Offers Improved and Novel Methods of Studying Physician Decision Making

Introduction: Data quality in stroke registries is typically dependent upon some form of chart review and manual data abstraction. The retrospective nature of this process is inherently prone to incomplete and inaccurate data collection with limited insight into the process of physician decision making. Hypothesis: New software packages accompanying telestroke systems will dramatically improve the quality of data by automating the abstraction process and providing real-time access to electronic databases. Methods: Telestroke systems provide web-based programs that record various levels of data. InTouch Technologies, Inc. currently provides StrokeRESPOND v3.0, a web-based program that facilitates telestroke consultation by organizing elements of the physician-patient encounter, including history, vitals, physical exam, laboratory results, and radiographs, and by generating a consultation note. Many data elements captured in the user interface mirror traditional metrics of acute stroke care research and can be de-identified and then directly transferred into an electronic database. The “forced choice” (aka hard-stop) design of data entry and elimination of secondhand abstraction can minimize data corruption and loss. Further, because each point of data entry and manipulation is time-stamped, powerful metadata_“data about data”_can be explored. By analyzing the sequence and patterns of clinical information entry and utilization, the actual thought process of the physician user can be investigated and provide new insights into stroke treatment. Optimization of acute stroke management, a complicated protocol, can be driven by identification of physician decision making patterns associated with multiple outcomes, including higher rates of treatment and faster treatment times. Conclusions: Specialized software programs will improve registry data collection, completeness and accuracy. The generation of metadata offers exciting, new avenues of research. Prospective stroke research using this methodology will require the collaboration of multiple academic institutions and industry partners.