Heterogenous Effects of Medicaid Expansion: The Moderating Role of Geographic Access to Oncology Care

• In Event: Revisiting Medicaid Expansion: Uneven Effects Across Patients, Providers, and Other Insurance Markets

Abstract

This study examines how the effects of Medicaid expansion on cancer outcomes vary by geographic access to oncology care. While prior research has evaluated the overall impact of Medicaid expansion, less is known about how these effects depend on patients’ proximity to oncology centers. This study addresses that gap by assessing whether geographic access acts as a key moderating mechanism influencing the effectiveness of expanded insurance coverage.

Leveraging cross-state variation in Medicaid expansion, this study employs a staggered difference-in-differences design comparing cancer outcomes before and after expansion, with non-expansion states serving as controls. To measure spatial access, a geospatial dataset of cancer care facilities was constructed using OpenStreetMap (OSM) and refined through a Large Language Model (LLM) assisted validation pipeline. Facility locations are geocoded and linked to patients’ county of diagnosis to derive travel time measures, which are then used to classify observations into subgroups reflecting different levels of geographic accessibility across the Surveillance, Epidemiology, and End Results (SEER) registry regions.

Linked SEER-Medicaid data, a nationally representative dataset maintained by the National Cancer Institute and covering cancer cases from 1999 to 2020 (N > 8.9M), was utilized for the analysis. Key outcomes include stage at diagnosis, survival months, and all-cause mortality across five major cancer sites: breast, lung, colorectal, prostate, and pancreatic.

Results indicate that Medicaid expansion is associated with improved cancer outcomes, including reductions in mortality and increases in survival months, with some evidence of earlier-stage diagnosis. However, these effects vary systematically by travel time to oncology care. In expansion states, improvements are observed across all levels of geographic access, including in areas with longer travel time, although gains are more evident among populations with shorter travel times to oncology centers.

In contrast, non-expansion states exhibit more limited and heterogeneous patterns. Even in counties with relatively short travel times to oncology facilities, improvements in survival outcomes are modest and less consistent. This suggests that geographic proximity alone is insufficient to generate meaningful gains in the absence of expanded insurance coverage.

Overall, the findings indicate that Medicaid expansion and geographic access jointly shape cancer outcomes, with the most consistent and substantial improvements observed where expanded coverage is paired with shorter travel times to oncology care. These patterns persist across SEER regions and cancer types and remain robust to a regression discontinuity design based on county-level comparisons along expansion and non-expansion state borders. These results underscore the need for policy approaches that simultaneously address insurance coverage and the spatial distribution of oncology care.

Author

• Kristine Jenny Hahm, University of Texas, Dallas
  Presenting Author