Developing algorithms to determine the cancer patients with metastatic recurrence using real-world data.
A better understanding of metastatic breast cancer recurrence (MBCR) is essential to understand long-term survival outcomes as well as to implement personalized treatments to reduce related mortality. Although many clinical trials have been performed on the treatment of MBCR patients, they are not straightforward to reflect real-world outcomes due to limited trial-size. In contrast, evidence derived from the large population-based real-world health data will be able to provide timely and accurate information to detect outcomes, and to optimize healthcare performance by revealing therapies to fit individual patients. Administrative data is a widely used data source for high-volume, population-based, multi-institutional research. However, it is inadequate for reliably addressing many important cancer outcomes, cancer recurrences, which are not explicitly documented in administrative data. Complementary to administrative data, EMRs (physicians’ progress text notes, radiology and pathology reports, and so on) contain large amounts of detailed information collected during routine healthcare delivery, which potentially can be used to identify the presence and timing of metastatic cancer recurrence. However, it has been challenging to extract information from unstructured free-text EMRs. The traditional manual chart review method is time-consuming and infeasible for big data. In the present study, we intend to develop and validate natural language processing (NLP) -based case-finding algorithms for timely and accurate detection of MBCR by integrating EMR and administrative data.
Principal Investigators: Yuan Xu
Co-Investigators: May Lynn Quan, Winson Cheung, Darren Brenne, Sasha Lupichik
Funded by: CCS data transformation grant ($125,000)
SPHERE | Strategies for Precision Health in Breast Cancer
University of Calgary
Powered by GoDaddy