In a new Medical Care article published ahead of print, Cheryl R. Clark, MD, ScD, and colleagues, of Brigham and Women’s Hospital and Harvard, provide pre-ACA implementation estimates of income-based disparities in delayed or forgone care due to cost by race/ethnicity, by state-level Medicaid expansion status.
Reforms can be unevenly implemented even if they address the primary causes of disparities. Uneven implementation can increase disparities—undercutting the purpose of the reforms. In this case, geographic differences may be exacerbated by decisions to expand Medicaid. As pointed out by the authors, state decisions about expansion are likely to be associated with other policies and investments in health and healthcare.
Providing insurance coverage under Medicaid expansion as a part of the ACA is intended to increase access to care among low-income adults. In the article, the authors used data collected from the 2012 Behavioral Risk Factor Surveillance System and techniques of small-area estimation to map the distribution of delayed or forgone care by race/ethnicity and to estimate the association among Medicaid expansion, personal income, and delayed or forgone care. Medicaid expansion status was defined as “planned” if the state planned to expand Medicaid as of December 2014. Expansions included “extending traditional Medicaid to low-income groups at 133% of the federal poverty line (plus a 5% income offset), … providing coverage through Section 1115 waivers, or … alternative structures, including using federal funding to provide coverage through exchanges,” the authors wrote.
In the county-level estimates of unadjusted prevalence of delayed or forgone care because of cost, some interesting patterns emerge. For Hispanics, there was more missing data at the county level, but the overall picture was of much higher prevalence of delayed or forgone care across the United States than for non-Hispanic whites, double in many counties (overall range 16.4%-42.9%). For non-Hispanic blacks, there was still more missing counties than for non-Hispanic whites, but with a different geographic pattern than for Hispanics. Prevalence rates of delayed or forgone care for non-Hispanic blacks were also relatively high compared to non-Hispanic whites but with generally more counties with higher rates clustered in the South and the East (overall range 12.9%-35.8%). For non-Hispanic whites, relatively more counties in the South had higher prevalence of delayed or forgone care compared to the North and upper Midwest, but several pockets of high prevalence were evident elsewhere (overall range 5.7%-27.6%). (ed. note: these are unadjusted prevalence rates, but look at those maps — it’s almost like the white folks are living in a different country).
There were also several differences by race/ethnicity in terms of individuals’ states of residence (note: link to Word file), such as non-Hispanic blacks being less likely to live in states that were planning expansion and Hispanics being more likely to live in states planning expansion. As predicted by the authors, the states planning to expand Medicaid were different from states not planning to expand. People living in states with expansion plans (note: link to Word file) had higher pre-ACA Medicaid eligibility thresholds, lower concentrations of poverty, and higher concentrations of primary care physicians, and lower county unemployment.
All these factors might make it relatively less strenuous for the state to implement Medicaid expansion and be evidence that the broader policies and investments in health and healthcare differed between those planning to expand Medicaid and those not planning to expand.
In the multivariable analyses (see article for details), the racial/ethnic differences in the prevalence of delayed or forgone care between non-Hispanic whites and non-Hispanic blacks was completely attenuated once the full set of covariates, including personal household income, were included in the model. The difference between non-Hispanic whites and Hispanics was still statistically significant, but attenuated from the unadjusted model. It should also be noted that there were relatively high levels of missing data, particularly on geography (11%) and income (9%). These results highlight the importance of income as a correlate of forgone or delayed care due to cost (something we would expect). However, the results without income are valuable since cost and income are in some ways inseparable.
As coverage expands, barriers to accessing care in states that expand coverage should decrease. However, the social and environmental factors that influence health and healthcare will only be indirectly affected. Only time will answer the question underlying the article, will differential expansion of Medicaid increase the existing racial/ethnic disparities in delayed or forgone care due to cost?
Jess Williams