Medicaid insures approximately one out of four rural Americans, but we have limited visibility into rural-urban differences in Medicaid utilization. Medicaid participants represent an important subpopulation of rural America, given that Medicaid includes low-income populations, children, and pregnant people, the latter two of are rarely included in Medicare data, which is more commonly used for policy analysis.

The research has two purposes. First: To assess the quality of Transformed Medicaid Statistical Information System (T-MSIS) Analytic Files in conducting analysis of rural-urban differences in utilization among Medicaid eligible. Medicaid programs are run by individual states, and although T-MSIS is designed to standardize the data across multiple systems, the data files are relatively new and unexplored by researchers. For example, a key variable for rural-urban analyses is BENE_ZIP_CD – the beneficiary residence ZIP code – as it identifies who is rural and who is urban. To the extent this variable might be of lower quality (e.g., unstable, inaccurate, or only contain three-digit ZIP codes in some states, for example), the ability to do rural-urban analyses could be considerably hampered. Likewise, it may be important to analyze utilization stratified by eligibility categories (e.g., the limited postpartum eligibility for those enrolled through the Medicaid for Pregnant Women category) and any threats to the quality of the variables that would limit the utility of analyses. Finally, race variables in administrative data are often of questionable quality; a better understanding of the accuracy in these data will be useful in assessing future work in equity using Medicaid data.

The second aim is a scientific aim designed to be an initial step into rural-urban Medicaid analyses of use. To examine key utilization metrics for rural-urban differences in utilization among Medicaid beneficiaries. For this aim, we will analyze common metrics for utilization, including access, cost, and quality, among Medicaid eligibles and assess whether there are rural-urban differences. These analyses will help address key unanswered questions facing rural population including, for example,

• How do costs for Medicaid eligibles vary by rurality?
• Do pregnant Medicaid eligibles have different experiences by their degree of rurality?
• Do rural children insured by Medicaid receive well-child visits as often as urban children?
• How does the race of the eligible intersect with rurality to affect patterns of care?

Answers to questions like these will help inform federal policy, regulations, and programs and empower state and local communities to tailor their approaches to the key differences affecting them.