Chua K, Sommers BD. Changes in Health and Medical Spending Among Young Adults Under Health Reform. JAMA. 2014;311(23):2437–2439. doi:10.1001/jama.2014.2202
Beginning September 23, 2010, the Affordable Care Act allowed young adults to be covered under their parents’ plans until 26 years of age. This dependent coverage provision increased insurance coverage and access among young adults.1,2 However, the association between implementation of the provision and medical spending, health care use, and overall health is unknown.
Our sample included adults aged 19 to 34 years in the 2002-2011 Medical Expenditure Panel Survey, an annual household survey of the US civilian population conducted by the Agency for Healthcare Research and Quality.3 The study used deidentified, publicly available data and was exempted from institutional board review. Based on previous research,1 we conducted a differences-in-differences analysis, defining the “intervention” group as adults aged 19 to 25 years and the control group as adults aged 26 to 34 years. We excluded 2010 as a washout period; 2011 was the postimplementation period.
Binary outcomes were having health insurance; having any outpatient visit, primary care physician visit, emergency department visit, hospitalization, or prescription medicine fill within the past 12 months; and reporting excellent physical and mental health. Continuous outcomes were inflation-adjusted annual health care expenditures, annual out-of-pocket expenditures, and percentage of expenditures paid out-of-pocket.
We fitted models predicting outcomes as a function of intervention group status, postimplementation year status, and their interaction (the population-level differences-in-differences estimate). We used linear models for binary outcomes and percentage of expenditures paid out-of-pocket. For dollar-value expenditure outcomes, we used a 2-part model: a linear model predicting the probability of any expenditures and a linear model predicting log-transformed expenditures among individuals with any expenditures.
We tested for diverging preimplementation trends in outcomes between groups. Regressions controlled for sex, self-reported race/ethnicity, marital status, Census region, and urban residence. We used SAS version 9.3 (SAS Institute Inc), sampling weights, and robust design-based variance estimators. We considered 2-sided P<.05 to indicate statistical significance.
The sample included 26 453 individuals in the intervention group and 34 052 in the control group. Overall, the sample was 46.6% male and 73.9% white. Group demographics were similar, except fewer adults in the intervention group were married (17.6% vs 56.1% in the control group).
Compared with the control group, the dependent coverage provision was associated with an increase of 7.2 (95% CI, 4.2-10.2) percentage points in the probability of insurance coverage among adults aged 19 to 25 years (P < .001), no statistically significant changes in health care use, an increase of 6.2 (95% CI, 3.2-9.3) percentage points in the probability of reporting excellent physical health (P < .001), and an increase of 4.0 (95% CI, 0.6-7.5) percentage points in the probability of reporting excellent mental health (P = .02) (Table 1).
Compared with the control group, implementation of the provision was associated with a decrease of 3.7 (95% CI, 0.9-6.4) percentage points in the percentage of expenditures paid out-of-pocket among adults aged 19 to 25 years with any expenditures (P = .009; Table 2). Annual out-of-pocket expenditures declined by approximately 18% (95% CI, 5%-31%) in the intervention group (from an unadjusted mean of $546.11), relative to the control group (P = .006). Results were similar after additionally controlling for household income, education, and employment. Preimplementation trends did not differ significantly between groups.
The dependent coverage provision was associated with improved self-reported health and protection against medical costs among adults aged 19 to 25 years compared with older adults unaffected by the law. One recent study indicated that the provision was associated with improved protection against emergency care costs; we found this financial protection extended to overall medical expenditures.4 Previous research documented rapid improvements in self-reported health among low-income and elderly adults gaining Medicaid and Medicare coverage, respectively.5,6 In one study, these gains occurred before any changes in health care use, suggesting that insurance may improve peace of mind and perceptions of health.5
We did not detect significant changes in health care use; however, only 1 year of postimplementation data was available for our study, which limited statistical power and prevented examination of longer-term changes. Another limitation is that other factors during the postimplementation period could have differentially affected outcomes between groups.
Corresponding Author: Kao-Ping Chua, MD, Harvard PhD Program in Health Policy, 14 Story St, Cambridge, MA 02138 (firstname.lastname@example.org).
Author Contributions: Dr Chua 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.
Study concept and design: Both authors.
Acquisition, analysis, or interpretation of data: Both authors.
Drafting of the manuscript: Chua.
Critical revision of the manuscript for important intellectual content: Both authors.
Statistical analysis: Chua.
Administrative, technical, or material support: Sommers.
Study supervision: Sommers.
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Sommers currently serves part-time as an advisor in the Office of the Assistant Secretary for Planning and Evaluation at the US Department of Health and Human Services. No other disclosures were reported.
Funding/Support: Dr Chua was supported by funding from the Harvard PhD Program in Health Policy.
Role of the Sponsor: The Harvard PhD Program in Health Policy had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: This was written entirely in Dr Sommers’ capacity as a Harvard employee and does not represent the views of US Department of Health and Human Services.