Mapping policy effects: The influence of Medicaid expansion on geospatial hotspots of incidence and late-stage breast cancer in Indiana.

e23049 Background: Expansion of Medicaid eligibility has been associated with earlier-stage breast cancer diagnoses, but it remains unclear whether this association holds true uniformly across space or varies between different communities. The objective of this study is to examine changes in the geographic hotspots of incidence of new breast cancer diagnoses and of late-stage disease at the time of presentation following Medicaid expansion in Indiana, using disaggregated estimates and raster-based geospatial visualization of patient residential location. Methods: From the Indiana State Cancer Registry (2010 – 2019), we identified females under 65 years of age who were diagnosed with breast cancer within a 59-month period before and after the Medicaid expansion date of 2/1/2015. Our primary outcomes of interests were breast cancer incidence and late-stage disease at diagnosis. Hotspots were identified using kernel ratio estimation, which compares the actual local patient-to-population ratio with simulated (expected) ratios to determine the statistical significance of the former. To calculate the actual ratio, the block group-level patient location data were disaggregated to points through a restricted and controlled Monte Carlo (RCMC) process. The simulated ratio was generated by an unrestricted but controlled Monte Carlo (UCMC) process. Using p < 0.1, we identified hotspots of outcomes before and after the Medicaid expansion. Results: Of the 28,006 total patients with breast cancer aged 18-64 years old, 14,792 (62.5%) were diagnosed with breast cancer before Medicaid expansion. Late-stage diagnosis occurred in 5,518 (37.3%) of the pre-expansion cohort and in 4,418 (33.4%) of the post-expansion cohort. Hotspots for breast cancer incidence were similar both before and after Medicaid expansion, occurring in the same geographic locations and incorporating approximately the same overall area. However, there was a 41% decreased area of hotspots in late-stage diagnosis after Medicaid expansion, with resolution of many of the pre-expansion rural hotspots of late-stage disease. Geospatial visualization demonstrated persistent hotspots of late-stage presentation in urban areas across the state. Conclusions: The data demonstrate similar hotspots of breast cancer incidence after Medicaid expansion but a reduction in the number and total area of hotspots for late-stage presentation. These persistent hotspots of incidence are consistent with known association between increased breast cancer diagnosis in more affluent populations rather than those most impacted by Medicaid expansion. With decreased but persistent significant hotspots with late-stage diagnosis, our results suggest that additional drivers of inequities beyond insurance coverage alone continue to contribute to geographic disparities late-stage diagnosis with breast cancer.