Date of Award
Dissertation - MUSC Only
Doctor of Philosophy (PhD)
College of Graduate Studies
Chanita Hughes Halbert
Problem: Identifying patients at risk of hereditary cancer based on their family health history (FHx) is a highly nuanced task. Frequently, patients at risk are not referred for genetic counseling because providers lack time and training to collect and assess FHx. Consequently, patients at risk are not receiving the genetic counseling and testing they need to determine the preventive steps they should take to mitigate their hereditary cancer risk. Methods: We combined chatbots, web application programming interfaces (APIs), clinical practice guidelines (CPGs), and ontologies into a web-service-oriented system that can automate FHx collection and assessment. We developed a lightweight, patient-centric, domain ontology using clinical practice guidelines; recruited users with ad campaigns; coded and compared user’s results for complementary guidelines; and regressed concordance with CPGs implemented by the system on type of healthcare professional using surveys and logistic regression. Results: The domain ontology has 758 concepts and encompasses 44 cancers, 144 genes, and 113 clinical practice guideline criteria. We reached 14,140 users in November 2019 through online marketing campaigns (Facebook, Google, and previous ItRunsInMyFamily (ItRuns) users). The final dataset contains 4,915 completed family histories, of which 2,221 met criteria and 2,694 did not. Of the 2,694 probands who did not meet criteria, 90.6% of them reported at least one cancer in their personal or family cancer history. Genetic counselors (GCs) had the best overall concordance with clinical practice guidelines (CPGs) at 82.2%, followed by oncologists with 66.0%, and primary care providers (PCP) with 60.6%. GCs had statistically better concordance with CPGs (p<.001) than non-GCs. All providers had higher concordance with CPGs for family health history (FHx) cases that met criteria than for cases that did not. Conclusions: Our results demonstrate that it is possible to gather FHx information at the population level, with high levels of engagement and interest, and more efficiently identify patients at risk of hereditary cancer using evidence-based guidelines. Earlier and consistent identification of patients at risk of hereditary cancer will lead to more effective screening and preventive actions leading to better long term health outcomes.
Ritchie, Jordon Bryan, "Automating the Identification of Patients at Risk for Hereditary Cancer with Chatbots, Clinical Practice Guidelines, Ontologies, and Web Services" (2021). MUSC Theses and Dissertations. 606.
All rights reserved. Copyright is held by the author.