Insurance and Health Care Expenditures: What's the Real Question?

For more than a quarter century, researchers and policymakers have known that people who must pay out of pocket for health care—because they are uninsured or face high levels of cost sharing—use less care (1, 2). However, uninsured persons differ from insured persons in a variety of measured and unmeasured ways (3). Consequently, the precise effect of insurance on health care utilization and expenditures is difficult to quantify by using observational data.

In this issue, Ward and Franks (4) take a new approach to addressing this question. They used data from the Medical Expenditure Panel Survey, which collects information on respondents over 2 years, to compare health care expenditures for 4 groups: continuously insured persons, continuously uninsured persons, persons who are uninsured during the first year but are insured the next year, and persons who are insured during the first year but are uninsured the next year. After controlling for measured attributes, including demographic characteristics, region of residence, and measures of health status, Ward and Franks found that spending by initially insured persons who lose their insurance is similar to spending by continuously uninsured persons and that spending by initially uninsured persons who gain insurance mirrors spending by continuously insured persons (4). These results suggest that insured persons use predictable amounts of health care regardless of whether they were previously insured or uninsured and that unobserved differences between insured and uninsured persons may not be terribly important. Ward and Franks conclude that their findings are “consistent with the assertion that by increasing access to health care, health insurance may improve health outcomes” and that expansion of health …