New methods estimate the costs of covering the uninsured
Research Activities, October 2009
The costs of covering the uninsured are uncertain for two reasons. First, analysts cannot observe all the characteristics of the insured and uninsured, so there is uncertainty about how to extrapolate from the costs of the currently insured to the currently uninsured. Second, estimates rely on survey data, but researchers have found some survey respondents do not accurately report insurance status.
Steven C. Hill, Ph.D., of the Agency for Healthcare Research and Quality, and Brent Kreider, Ph.D., at Iowa State University, used data from the Medical Expenditure Panel Survey (MEPS), a nationally representative household survey, to quantify the impact of both problems on estimating the costs of universal coverage. Their primary analysis considered the impact of extending insurance to the uninsured using a mix of public and private coverage. Related research found a low rate of misreporting private insurance in the MEPS, and they combined this with research on misreporting public insurance to derive statistically conservative assumptions about misreporting overall. They used new econometric methods and transparent assumptions to estimate the maximum amount health care spending could increase.
The researchers estimated that under universal coverage the fraction of the nonelderly population (including the currently insured and currently uninsured) using ambulatory or hospital services would rise no more than 9 percent. Monthly per capita provider visits would rise by no more than 8 percent, and monthly expenditures would rise by no more than 16 percent. These estimates vary only a little under plausible, alternative assumptions about how often insurance status is misreported.
See "Partially identifying treatment effects with an application to covering the uninsured," by Drs. Kreider and Hill, in the 2009 Journal of Human Resources 44(2), pp. 409-449. Reprints (AHRQ Publication No. 09-R058) are available from the AHRQ Publications Clearinghouse.