Neurosurgery fellow Brigham and Women's Hospital Boston, MA, US
Introduction: Veterans with GBM may have their initial diagnosis in the VA with follow-up treatment in the community, or have their diagnosis and treatment in the community and then follow-up care within the VA. This fragmented care is inefficient, expensive, and may lead to poor outcomes. We developed GLIOALERT as a data-science algorithm to identify Veterans nationally with a diagnosis of GBM and to coordinate and monitor their management pathway, including enrollment in clinical trials.
Methods: We developed GLIOALERT to detect radiographic diagnosis of glioma from radiology reports. GLIOALERT is designed for patient care navigators to track Veterans identified by the algorithm and coordinate their management pathways. Information captured includes patient demographics, confirmation of diagnosis, radiology impression (i.e. tumor grade, initial vs recurrent), prior treatment (i.e. surgery, radiation, chemotherapy), and the clinical care pathways. Process and quality monitoring information includes timepoints between diagnosis, treatment, and survival. GLIOALERT is designed to integrate with our clinical trial matching application.
Results: Our algorithm identified 49 patients with suspected glioma from radiology reports throughout the VA system between 8/1/2023 and 9/30/2023. After manual review, thirty-two cases had confirmed radiology impression of high-grade glioma, glioblastoma, or low-grade glioma. Seventeen cases had glioma in the report as information from family history, differential diagnosis, or remote history. Record review of the confirmed cases to establish baseline treatment patterns found 31% (10) with molecular testing documented, 50% (16) with surgical treatment, 34% (11) with radiation treatment, 34% (11) with chemotherapy, and 22% (7) with all 3 treatment modalities (i.e., surgery, radiation, chemotherapy) from either care in the community or within the VA.
Conclusion : Here we present preliminary results of our effort to improve care coordination for Veterans with newly diagnosed GBM with the development of a workflow application for patient care navigators. We are continuing to refine our case-finding algorithm with additional natural language processing methods and further improve the usability of the application. Future studies will compare treatment pattern changes in the VA before and after implementation of GLIOALERT.
.jpg)
