OBJECTIVES: The effectiveness of clinical decision support (CDS) tools for selecting therapies in precision oncology is dependent on the flow of biomarker testing information through the electronic health record (EHR). The growth of biomarker-specific cancer treatments has contributed to an increasingly complex environment of testing choices and results for providers to navigate. The purpose of this study was to map biomarker information pathways for cancer treatment within an EHR. Specifically, the study used test reports, structured EHR data, and CDS tool records to measure and visualize the flow of biomarker data relative to critical treatment decisions made by providers.

METHODS: Patients initiating treatment for metastatic Non-Small Cell Lung Cancer (mNSCLC) between January 2020 and April 2022 were identified from the iKnowMed oncology EHR database. Events and characteristics of biomarker testing, including dates, laboratories, and results were captured from unstructured and structured sources. The use of an EHR CDS tool and its linkage with structured EHR results was also recorded. The sequence and time intervals between these events were mapped and measured, from initial patient encounter through most recent line of therapy in the study period.

RESULTS: A total of 3,135 treated mNSCLC patients from iKnowMed were identified. All biomarker testing characteristics and results were captured or their absence was noted. 90% (2,818) of patients had at least one test result among 13 biomarkers, including KRAS, EGFR, and NTRK. 74% of patients (2,326) had results saved in structured EHR fields and the CDS tool incorporated these results for 3,133 treatment recommendations across 66% of patients (2,059).

CONCLUSIONS: CDS tools can have a significant impact assisting providers in identifying potential precision oncology treatment options. This study illustrates the flow of testing information and measures the degree to which CDS tools use structured data to support treatment decisions in a community oncology setting.