Room: Room 207
Purpose: To implement the usage of a largely automated software-driven approach that provides rapidly available reject analysis metrics and a flexible means of showing historical trends and performance
Methods: Usage of developmental software was implemented with the aim of providing x-ray reject rates for users of conventional radiography systems within a hospital network. Two virtual machines (VM, data ingestion and analytics) were established to aggregate, analyze, and present results via a web-based interface. Sixteen fixed and portable systems (GE XR656, XR656+, XR220) at three different hospitals and clinics were connected to the ingestion VM. From each system, standard vendor-generated reports are pulled, and the user uploads a custom-generated report from the Radiology Information System. Using aggregated data from these two reports, the reject rate for an individual technologist was calculated as the ratio of the number of rejected images to the number of total images acquired. Time performance and practical utility was benchmarked against previously quarterly process (~7 hours), which relied on manual fetch of reports from each system and analysis using spreadsheets.
Results: Post-implementation, data generation takes under 15 minutes. The process is less susceptible to data loss/corruption, provides improved data integrity, and offers exceptional insights for targeted technologist training. For example, a new employee decreased reject rate from 20.6% to 6.9% after improving AP Chest view. Feedback reports are now provided monthly to each technologist, but all users have the ability to interrogate their own recent performance at any time. Forthcoming work includes integration of other conventional radiography vendors and the collection of various dose indices as a means of analyzing dose performance.
Conclusion: Rapid and user-specific reject rates calculated using readily available standardized reports is achievable with the end goal of performing targeted training to improve the quality and speed with which medical images are produced.
Funding Support, Disclosures, and Conflict of Interest: Access to the software used for data aggregation, analytics, and presentation was made possible through a Beta agreement between the University of Washington and GE Healthcare.