OBJECTIVES: Many healthcare payers rely on confirmation of treatment guideline concordance when processing prior authorization of costly drugs. The objective of this study was to measure the concordance of prescribed cancer regimens with NCCN Guidelines and NCCN Preferred using Flatiron Assist, a clinical decision support (CDS) tool integrated into the electronic health record (EHR) and deployed at the point-of-care.

METHODS: This study used the de-identified Flatiron Health Research Database for retrospective analysis. We included patients diagnosed with breast, colon, non-small cell lung, prostate, or rectal cancer who had at least one cancer treatment regimen prescribed by a provider who deployed the optional CDS tool across 8 practices between May 2020 and October 2021. The primary outcome measured was the proportion of regimens prescribed that were concordant with relevant NCCN Guidelines. A secondary outcome measured was the proportion of orders that aligned with NCCN Preferred treatment regimens between May 2021 and October 2021 across 7 of the 8 practices with NCCN Preferred recommendations enabled.

RESULTS: The population included 4838 patients and 211 providers with 6500 orders. The three most common cancer types by orders were non-small cell lung (25%), breast (25%), and colon cancer (13%). Most patients had private insurance (60%) or Medicare (38%). Among all regimens ordered through Flatiron Assist, 91% were concordant with NCCN Guidelines. Between May 2021 and October 2021, 85% of 2055 orders across 1763 patients aligned with NCCN Preferred when preferred regimens were available.

CONCLUSIONS: This study assessed the oncology-specific point-of-care CDS tool Flatiron Assist and found high concordance between cancer treatment regimens ordered and both NCCN Guidelines and NCCN Preferred among users of the tool. Future work should explore opportunities for digital authentication of treatment guideline concordance, embedded within EHR workflow, to overcome barriers with manual prior authorization and potentially avoid delays in treatment delivery.