Session: Effects of Uninsurance
Room: Hollister 110
Time: Tue 08:30-10:00

Presenter: Lindsay Sabik (Harvard University. Health Policy)

Discussant: Anthony Lo Sasso (University of Illinois at Chicago)

Abstract

Recent studies have suggested that the rate of uninsurance in a community can have negative effects for the insured as well as the uninsured. This paper investigates how variation in the uninsurance rate within a market over time affects access to care for both uninsured and insured individuals within the market. Individual level data from the 1996 and 2003 waves of the Community Tracking Study are used to investigate the effect of the uninsurance rate in a market on access to care for the uninsured and the insured, where the main measure of access is whether an individual reports not having received medical care he or she believed was necessary (“forgoing care”). Models include market fixed effects in order to control for time-invariant area characteristics. Data from other sources, including the Area Resource File and the Bureau of Primary Healthcare Uniform Data System, are included to consider the role of time-varying market-level factors in affecting access for each of the groups. In addition, market level data on adjusted health care premium costs and unemployment levels are used to instrument for the uninsurance rate using a two-stage residual inclusion approach.

Estimates from the base market fixed effects model indicate that increases in the local uninsurance rate lead to statistically significant decreases in access to care for the uninsured but have no effect among the insured, suggesting that increases in the uninsurance rate do not have strong spillover effects on access for the insured. The marginal effect for the average individual in the uninsured sample is a 0.49 percentage point increase in the probability of forgoing care associated with a 1 percentage point increase in the uninsurance rate, with the 95% confidence interval (CI) ranging from 0.20 to 0.78 percentage points. Among the insured sample the marginal effect is a change in the probability of forgoing care of 0.04 percentage points (95% CI -0.07 to 0.14). The IV results are similar but insignificant for both samples.

Estimates from the base models that incorporate market and time effects suggest that increases in the rate of uninsurance in a market lead to poorer access to care for the uninsured. In contrast, there is no evidence of an effect among the insured. These results indicate that previous research finding evidence of negative spillovers of uninsurance on access for the insured may not have adequately controlled for differences between markets associated with both uninsurance and access. Further, this empirical strategy may be more relevant than previous work that relied entirely on cross-sectional variation for assessing the potential effects of policies that reduce the rate of uninsurance. This research suggests that policies that expand coverage to the uninsured may affect access to care for both the previously and continuously uninsured, particularly in communities that see a reduction from high rates of uninsurance. I find no evidence of spillovers to the insured, however, suggesting that such policy changes may have little effect on access to care for those who are already insured.

Key Terms

Uninsurance, Spillovers, Access to Health Care

Authors:

Lindsay Sabik (Harvard University. Health Policy)