MP23-17 ASSOCIATIONS OF REGIONAL-LEVEL ADOPTION OF PROSTATE MRI AND GENOMIC TESTING AND OBSERVATION FOR PROSTATE CANCER

Abstract

INTRODUCTION AND OBJECTIVE: Prostate MRI and tissuebased gene expression (genomic) testing may facilitate the use of observational management for prostate cancer, however their independent role is unclear. We aimed to evaluate whether geographic regions with more rapid adoption of these tools were more likely to experience increases in the use of observation for prostate cancer. METHODS: The analysis was based on Blue Cross Blue Shield Axis® (BCBS Axis) - the largest collection of commercial insurance claimse through a secure data portal. We included patients aged 40-89 years newly diagnosed with prostate cancer from 7/2012-6/2018, and identified claims for prostate MRI and genomic testing within six months of diagnosis. The dependent variable was the change in the proportion of patients observed for prostate cancer at the hospital referral region (HRR) level between two periods: 7/2012-6/2014 to 7/2016-7/2018. The independent study variables included HRR-level changes in the proportion of men receiving prostate MRI and genomic testing between these periods. We used multivariable linear regression to adjust for proportion of observation and MRI/genomic testing at baseline and HRR-level contextual factors related to prostate cancer incidence, treatment patterns, and access to care. RESULTS: Results: We identified 232,237 men with prostate cancer. The overall proportion of patients who received MRI and genomic testing increased from 7.6% to 15.6% and 1.2% to 11.5%, respectively from 7/2012-6/2014 to 7/2016-7/2018. During this period, the proportion of patients receiving observation increased from 26.0% to 33.4%. Increases in HRR-level use of prostate MRI (r 0.24, p<0.01) and genomic testing (r 0.18, p<0.01) were positively correlated with change in observation. In multivariable linear regression models, HRR-level increases in the use of prostate MRI was associated with change in observation as management (β 0.29, standard error [SE] 0.069, p<0.01). The association between HRR-level change in genomic testing and change in observation did not meet the predefined threshold for statistical significance (β 0.14, SE 0.073, p=0.06). CONCLUSIONS: Conclusion: The adoption of MRI and genomic testing for prostate cancer has varied considerably across regional healthcare marketplaces. Regional-level change in the use of prostate MRI was independently associated with use of observation for prostate cancer. This study is limited by its ecological design and absence of patient-level data.