Understanding place-based inequities in lung cancer care: Trends in geographic hotspots of late-stage diagnoses and rates of cancer-directed surgery after Medicaid expansion in Indiana.

Abstract

e23050 Background: While Medicaid expansion has been associated with improved cancer outcomes, current research often relies on aspatial methods and single-point analyses that fail to capture geographic variation and differences in policy associations along the cancer care continuum. This study examines the impact of Medicaid expansion on lung cancer in Indiana by identifying hotspots of late-stage diagnoses and foregoing cancer-directed surgery (CDS) before and after policy implementation. Methods: Using the Indiana State Cancer Registry, we identified all patients between ages of 18 and 65 years of age who were diagnosed with incident lung cancer between 2010 and 2019. Hotspots for late-stage diagnosis and foregoing cancer-directed surgery were identified using kernel ratio estimation, a method that compares the actual local patients-to-population ratio with simulated (expected) ratios. To calculate the actual ratio, the block group-level patient location data were disaggregated to points using a restricted and controlled Monte Carlo (RCMC) process. The simulated ratio was generated by an unrestricted but controlled Monte Carlo (UCMC) process. Both RCMC and UCMC were based on population distribution. This process could determine the statistical significance of the actual ratio at each location. Using p < 0.01, we identified hotspots for periods before (3/1/2010–1/31/2015) and after Medicaid expansion (2/1/2015–12/31/2019). Results: Among 12,126 lung cancer patients who were < 65 years old, 7,931 (65.4%) were diagnosed at a late stage, and 2,243 (46.3%) patients with nonmetastatic disease forewent CDS. The overall number of late-stage diagnoses decreased after Medicaid expansion, with a corresponding 34% reduction in the summed area of hotspots. There was 42% increase in the summed area of hotspots for foregoing CDS. Geographic visualization of foregoing CDS hotspots revealed reductions in hotspot areas surrounding Indianapolis and northern Indiana, while hotspots arose or increased in size within Indianapolis and in smaller cities such as Lafayette, Kokomo, Anderson, and Marion. Conclusions: This study highlights the heterogeneity of Medicaid expansion effects on lung cancer outcomes across the state. A more uniform decrease in hotspots of late-stage diagnosis was observed, suggesting reduced geographic variation in late-stage diagnoses. The observed geographic shifts in hotspots of patients foregoing CDS after expansion was likely driven by multiple spatially-unique factors beyond insurance coverage alone. Incorporating place-based data is essential for understanding the impact of Medicaid expansion and optimizing lung cancer control strategies moving forward.