IMPORTANCE There has been little rigorous evidence to date comparing public vs private health insurance. With policy makers considering a range of policies to expand coverage, understanding the trade-offs between these coverage types is critical. OBJECTIVE To compare months of coverage, utilization, quality, and costs between low-income adults with Medicaid vs those with subsidized private (Marketplace) insurance.
DESIGN, SETTING, AND PARTICIPANTSThis cross-sectional study used a propensity scorematched sample of adults enrolled in either Medicaid or Marketplace plans at any point between January 1, 2014, and December 31, 2015. The sample was restricted to individuals with incomes narrowly above and below 138% of the federal poverty level (FPL), which represented the eligibility cutoff between the programs. Data were obtained from 3 state agencies merging comprehensive insurance claims with income eligibility data for Colorado Medicaid expansion and Marketplace enrollees. Income data were linked with an all-payer claims database, and generalized linear models were used to adjust for clinical and demographic confounders. Participants included 8182 low-income nonpregnant adults aged 19 to 64 years enrolled in Medicaid or Marketplace coverage during the 2014 to 2015 period, with incomes between 134% and 143% of the FPL. EXPOSURES Health insurance through Colorado Medicaid or Colorado's state-based Marketplace.
MAIN OUTCOMES AND MEASURESThe primary analytical approach was a multivariate regression analysis of the propensity score-matched sample. Primary outcomes were months of coverage in Medicaid or Marketplace insurance, office and emergency department (ED) visits, ambulatory caresensitive hospitalizations, and total costs. For secondary quality outcomes, the propensity scorematched sample was widened to 129% to 148% of the FPL to ensure adequate sample size.Secondary outcomes included prescription drug utilization, types of ED visits, hospitalizations, out-of-pocket costs, and clinical quality measures. Primary data analysis was between September 2018 to July 2019, with revisions finalized in November 2020.
RESULTSThe propensity score-matched narrow-income sample included a total of 8182 participants (4091 Medicaid eligible [50%]: mean [SD] age, 42.8 [13.6] years; 2230 women [54.5%]; 4091 Marketplace eligible [50%]: mean [SD] age, 42.7 [13.9] years; 2229 women [54.5%]).Demographic differences across the 2 groups were well balanced, with all standardized mean differences less than 0.10. Marketplace coverage was associated with fewer ED visits (mean, 0.36 [95% CI, 0.32-0.40] visits vs 0.56 [95% CI, 0.50-0.62] visits; P < .001) and more office (outpatient) visits than Medicaid (mean, 2.22 [95% CI,] visits vs 1.73 [95% CI, 1.64-1.81] visits; P < .001).