GARDE: Scalable Clinical Decision Support for Individualized Cancer Risk Management

PROJECT SUMMARY/ABSTRACT Evidence supports individualizing risk-stratified cancer screening, with selective application of specific screening interventions best suited to the individual. Yet, individualizing cancer screening at a population scale requires the implementation of personalized risk assessments which are quite challenging to achieve in

today’s overwhelmed primary care settings. A promising approach to address this challenge is to automate the identification and management of eligible patients using electronic health record (EHR) technologies coupled with advanced clinical decision support (CDS) tools and automated conversational agents (“chatbots”). In

previous research funded by the National Cancer Institute (NCI) Informatics Technology for Cancer Research (ITCR) program, we have enabled GARDE (Genetic Cancer Risk Detector), a standards-based CDS platform for individualized cancer screening. GARDE (i) screens and identifies patients who meet National

Comprehensive Cancer Network (NCCN) criteria for genetic testing based on their family history and other risk factors in the EHR; and (ii) uses automated chatbots offering patient outreach and education, offering access to genetic testing and explanation of test results. GARDE has been integrated with two market-leading

EHR systems (Epic® and Cerner®) using the Fast Healthcare Interoperability Resources (FHIR) and CDS Hooks standards. GARDE has been successfully deployed in clinical settings at two academic medical centers and their respective cancer centers (University of Utah/Huntsman Cancer Institute and New York University) in

support of the BRIDGE trial, funded by the NCI Cancer Moonshot program (U01CA232826 – Kaphingst, PI). The overall objective of the present proposal is to enhance and disseminate GARDE across healthcare systems including high resource medical centers and low resource safety net healthcare systems. Our approach will be

guided by implementation science frameworks that help assess implementation readiness, identify barriers and facilitators, identify needs for adaptation, and develop implementation strategies. Specifically, we will (i) enhance GARDE’s chatbots using open-source technologies; (ii) deploy GARDE at new collaborating sites

(Cornell University, Medical University of South Carolina [MUSC], Beaufort Memorial Hospital); (iii) conduct rapid iterative pilot implementations at these new sites; (iv) iteratively develop and test an implementation toolkit based on experience with the pilot sites; (v) conduct a cost analysis to catalyze further adoption; and

(vi) disseminate GARDE beyond the collaborating sites through the implementation toolkit and direct technical assistance. Through wide dissemination, GARDE has the potential to enable evidence-based individualized cancer screening and reduce cancer burden through a scalable, population-based, and interoperable approach.