 | The 131st Annual Meeting (November 15-19, 2003) of APHA |
|
| The 131st Annual Meeting (November 15-19, 2003) of APHA |
Amélie Quesnel-Vallée, MSc MA, Department of Sociology, Duke University, 268 Soc/Psych Building, Durham, NC 27708-0088, 919-660-5604, aq@duke.edu
Background: Among the general U.S. population of working age, privately insured individuals enjoy better general health than the uninsured. However, the same finding does not seem to hold for public insurance, as most studies find a negative association of public insurance with general health that is stronger than that of being uninsured. Such findings are generally attributed to the strong selection bias of unhealthy and low socioeconomic status (SES) individuals into public insurance coverage. Objective: To determine the extent to which different sources of health insurance (including public health insurance) have an impact on self-reported general health, using fixed effects regression with sibling clusters to lessen the selection bias in health insurance coverage. Study Design: The data are taken from the 1979 National Longitudinal Survey of Youth, an ongoing longitudinal panel survey that has been following since 1979 a national probability sample of American youth. The sample was limited to those 18 to 22 years old at baseline (1979) and was constituted of 221 clusters of siblings, or 461 individual respondents. Analytically, fixed effects models with sibling clusters will be compared with the results of a conventional OLS regression. Fixed effects models postulate that, by using one sibling virtually as a “control” for another one, it is possible to estimate the effects of current factors net of common family background (SES) and genetic predispositions. These are factors that can affect the selection of individuals into different types of health insurance coverage, but that have typically been unavailable from health surveys. Results: Public insurance has in the conventional regression a strong negative association with health that declines in magnitude by 34.7% and becomes non significant in the fixed effects regression with sibling clusters. Conversely, the effects of years uninsured, which were non significant and half those of years publicly insured in the OLS regression, increase in magnitude by 64.9% and become highly significant and double those of years publicly insured in the fixed effects regression. Substantively, this suggests that each year uninsured is associated with a decrease in general health equivalent to the benefits of two years of education. Conclusions: These results suggest that being uninsured has a strong negative association with health, while public and private insurance may have no impact beyond that of adult SES. However, this relationship can only be observed with methods that specifically address the selection bias in health insurance allocation.
Learning Objectives:
Keywords: Adult Health, Medicare/Medicaid
Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.