The US Affordable Care Act (ACA) Health Insurance Marketplaces have been associated with improvements in access to care,1 but evidence on primary care is limited.2 There is evidence that suggests restricted hospital networks as well as low primary care physician participation in the Marketplaces3,4 may limit access to primary care services. Therefore, we examined whether the ACA Marketplace was associated with changes in primary care spending and use.
We used nationally representative data from the 2010-2017 Medical Expenditure Panel Survey. We included adults 26 to 64 years of age with family incomes between 138% and 400% of the federal poverty level in our sample. The intervention group was defined as persons with private nongroup insurance coverage who were eligible for Marketplace coverage with premium subsidies and/or cost-sharing reductions. Persons with employer-sponsored insurance were defined as the control group because a negligible association between the ACA and employer-sponsored insurance was found.5 Outcome measures included annual primary care spending, out-of-pocket spending for primary care visits, any primary care visit, and number of primary care visits. Spending was inflation-adjusted to 2017 US dollars. We defined the pre-ACA period as being from 2010 to 2013 and the post-ACA period as being from 2014 to 2017. Control variables included age, sex, race/ethnicity, marital status, educational level, family income, metropolitan residence, self-reported health status, number of chronic conditions, and survey year.
We used difference-in-difference multivariate linear regression (annual primary care spending, out of pocket spending for primary care, and number of primary care visits) and linear probability (any primary care visit) models that were adjusted for survey weights and robust standard errors. We conducted sensitivity analysis by limiting the intervention group in the post-ACA period to those who gained Marketplace coverage. We also conducted a sensitivity analysis by using a 2-part model for annual primary care spending and out-of-pocket spending for primary care visits, a negative binomial model for primary care visits, and a logistic model for any primary care visit. Because the distribution of health care spending is skewed, we also conducted a sensitivity analysis by reexamining the linear regression model with log-transformed spending. This study used deidentified, publicly available data and therefore was considered nonhuman subjects research. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. All analyses were conducted using Stata statistical software version 16.0 (StataCorp). All P values were from 2-sided tests, and the results were deemed statistically significant at P < .05.
We included 39 415 adults (19 274 women [48.9%]; mean [SD] age, 42.4 [11.8] years). Our analysis showed no significant changes in annual primary care spending, out-of-pocket spending for primary care visits, or any primary care visits between the intervention and control groups (Table 1). We found a statistically significant, but modest, increase in the number of primary care visits among the intervention group relative to the control group (differences-in-differences adjusted estimate, 0.24 [95% CI, 0.11-0.38]; P < .001). Our findings were robust to sensitivity analyses (Table 2). The parallel trends assumption for the difference-in-difference analysis was met for all outcomes.
There were no or little changes in primary care spending and use after the implementation of the ACA Health Insurance Marketplaces in the US. The null findings might be explained by the poor incentives for primary care associated with the low dollar value of cost-sharing subsidies; newly enrolled individuals, especially those with low health literacy,6 might be unaware of the new system; or low participation by primary care physicians’ in Marketplace plans may not be increasing perceived access to care.4 Our findings suggest that the provisions in the ACA to improve access to primary care may not be sufficient. The interpretation of this study should be viewed within the limitations of the data, which includes that our study used cross-sectional data and that there may be no perfect control group for the ACA Marketplace.5
Accepted for Publication: April 6, 2020.
Published: June 10, 2020. doi:10.1001/jamanetworkopen.2020.7442
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Park S et al. JAMA Network Open.
Corresponding Author: Sungchul Park, PhD, Department of Health Management and Policy, Dornsife School of Public Health, Drexel University, 3125 Market St, Nesbitt Hall 3rd Floor, Philadelphia, PA 19104 (smp462@drexel.edu).
Author Contributions: Dr Park had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: Park.
Drafting of the manuscript: Park.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Park.
Administrative, technical, or material support: Park.
Supervision: Park, Stimpson.
Conflict of Interest Disclosures: None reported.