Medicaid Expansion versus the Private Option: Changes in Access to Care, Medical Spending, and Health

Under the Affordable Care Act, 30 states and the District of Columbia have expanded Medicaid, with several states using private insurance to expand (the so-called “private option”).  Despite vigorous debate in several states about whether to expand coverage to low-income adults with private insurance versus traditional Medicaid, there is little evidence on the relative merits of the two approaches. 

We compared the preliminary impacts of a traditional Medicaid expansion, the private option, and non-expansion by conducting a repeat cross-sectional telephone survey of low-income adults in Kentucky, Arkansas, and Texas (n=5,665) prior to and one year after expansion (November-December 2013 versus November-December 2014).   We used a differences-in-differences analysis, adjusting for sociodemographic covariates, to compare the impact of these alternative expansion approaches in three Southern states with high baseline poverty and uninsured rates.  We conducted two sets of analyses: first, pooling the two expansion states together to assess the overall effect of coverage expansion vs. non-expansion, using Texas as the control group.  Then, we compared the changes in the traditional Medicaid expansion (Ketucky) vs. those in the private option (Arkansas).

We found that the uninsured rate declined significantly in the two expansion states compared to Texas, by 14.0 percentage points [ppt] (p<0.001). Coverage expansion resulted in a greater likelihood of having a personal doctor (+7.9 ppt, p=0.07) and a check-up in the past year (+6.9 ppt, p=0.07), and a reduced likelihood of using the Emergency Department as a usual source of care (-5.1 ppt, p=0.06).  In terms of affordability, skipping medications due to cost (-9.9 ppt, p=0.002) and trouble paying medical bills both declined significantly after expansion (-8.9 ppt, p=0.003), and average out-of-pocket medical spending declined by 24% from a baseline mean of $434 per year.  The share of individuals with chronic conditions obtaining regular care for those conditions increased substantially (+11.6 ppt, p=0.002).  There were no significant changes in overall utilization or self-reported health in the first year of the expansion. 

Comparisons between Kentucky and Arkansas revealed only a few differences across the outcomes assessed in the survey. First, as expected, coverage gains were primarily via Medicaid in Kentucky and private insurance in Arkansas.  Second, trouble paying medical bills dropped by 12.9 ppt in Kentucky versus just 4.8 ppt in Arkansas (between-group difference p=0.05), both relative to Texas. Otherwise, there were no significant differences for any outcomes between Kentucky and Arkansas. 

Overall, coverage expansions in Kentucky and Arkansas both produced significant improvements in access to care and affordability, compared to non-expansion, but with only modest differences between the Medicaid and private option models – namely improved financial protection for beneficiaries in Medicaid.  This suggests that both the private option and traditional Medicaid expansion can produce similar benefits for low-income adults.