Jan 1, 1997
Can We Get to Universal Coverage?
The goal of expanding public health insurance programs is to provide coverage to most of the nation's uninsured. But policymakers face a number of challenges in determining whether this goal can be reached. For example, how much will it cost? How should new public programs be financed? How will benefits be distributed? Can expanding insurance be left up to the states?
These and related issues are explored in a series of studies by economists Stephen Long and Susan Marquis. Their work draws on data from the Robert Wood Johnson Foundation (RWJF) Family Health Insurance and Employer surveys, conducted in 1993–1994 in Colorado, Florida, Minnesota, New Mexico, New York, North Dakota, Oklahoma, Oregon, Vermont, and Washington. Collectively, these states are similar to all states in their health care systems and population characteristics, and span the variation observed in all 50 states in important population and health policy characteristics. The surveys, which Long and Marquis designed in collaboration with leading survey organizations, compiled extensive insurance, utilization, health status, and demographic information.
Among key study findings to date:
Many states have proposed or implemented programs to provide insurance to low-income, uninsured residents. How much will these programs cost?
Long and Marquis estimated eligibility and costs for three programs illustrative of those enacted or under consideration. The number of people who would be eligible for the new programs ranges from 6 to 10 percent of the 10 RWJF survey states' combined population, depending on the specific program parameters. The cost of the expanded coverage in the 10 states combined ranges from about $4.3 to $7.9 billion, depending on program features. This represents 3 to 5 percent of total personal health care spending in the 10 states.
But aggregate estimates mask substantial variation. Indeed, health problems cluster within states: Residents of states with the highest percentage of uninsured are more likely to be in ill health and to have more severe problems with access to care.
The percentage of the nonelderly population without insurance coverage varies substantially, ranging from 27 percent in New Mexico to 10 percent in Minnesota. However, across all 10 of the RWJF survey states, the percentage of the population with public coverage varies little. Thus, the interstate spread of the noninsured stems from variations in the rate of private coverage.
Long and Marquis took a detailed look at the interactions among insurance coverage, health status, and access to care. They grouped the three states with the highest percentage of uninsured persons (Florida, New Mexico, and Oklahoma) and the three having the lowest (Minnesota, North Dakota, and Vermont). They used these groupings to characterize access and health status in states with similar uninsured rates. The following profiles emerge.
Health status: Persons living in states with a higher percentage of uninsured are about twice as likely to be reported in fair or poor health as those living in states with a lower percentage.
Access to care: Residents in states with a higher percentage of uninsured have less access to care. Figure 1 shows variation in several measures of access separately for children and adults. The first two measures (no usual source of care and no emergency care when needed) are about two to three times higher in the high-uninsured states.
Because of differences such as those described above, policies designed to expand health care coverage will have different effects in different states. One example:
Long and Marquis investigated a prototype plan similar to the State Children's Health Insurance Program (CHIP), a federal-state partnership intended to extend health care coverage to a significant proportion of the nation's uninsured children.
Expanding public insurance would substantially improve access for low-income uninsured children. On average, across all 10 RWJF states, a CHIP-like plan would increase physician contact from 2.3 to 4.6 visits per year. But the increase in visit rates for uninsured children would vary significantly, ranging from lows of 41 percent in Minnesota and 50 percent in New York to highs of 135 percent in New Mexico and Vermont and 189 percent in Oregon.
A state's safety net capacity, assessed by measures such as public hospital beds as a percentage of total hospital beds and emergency room visits per low-income person, plays an important role in this variation. Predicted access gains for the three states ranking lowest in safety net capacity are 150 percent, whereas the gains are 80 percent for the three states with the highest capacity.
This analysis suggests that a CHIP-like program is likely to boost the number of low-income children who will be newly insured, substantially increase their access to physician services, and do so across the country. But the magnitude of the effects will vary greatly from one state to another. The biggest potential improvements in access to care are in states that have traditionally provided the scantiest health safety nets.
Can independent actions by states, taken collectively, substantially reduce the nation's uninsured? Probably not. Here's why.
States with a high percentage of uninsured face a significant challenge in expanding health insurance coverage. They will have to spend more per capita than other states to attain equivalent outcomes. But they lack the tax capacity to do so.
Long and Marquis used the additional federal income taxes that a family would pay to finance an illustrative national program of subsidized health insurance for low-income persons as a measure of a state's capacity to finance health reform. They compared this measure of tax capacity with the additional state income taxes the family would be required to pay if each state introduced the same program.
They grouped the 48 states of the continental United States into four groups of 12 states, ordered by the uninsured rate in the state. Figure 2 shows the percentage of uninsured in each group, ranging from 10 to 21 percent. It also shows that the illustrative program Long and Marquis considered would extend coverage to most of the uninsured in all of the states, no matter how many uninsured residents a state had initially.
How much an extended insurance program will cost is directly related to how many people it will cover; however, the costs can be distributed in different ways. In a national program, they are distributed according to the distribution of family incomes among the states. As a result, per capita taxes are higher in the 12 states with the lowest uninsured rates—about $188—because these states have higher-income populations. In the 12 states with the highest rates of uninsured, the per capita tax increase is about $154; tax increases are lower in these states because they have lower-income populations.
In contrast, under the state-financed plan, costs are distributed according to where the newly insured live. Thus, costs will be higher in states that have more uninsured residents. The estimated per capita tax increase for such states is $230. For residents of the states with the lowest uninsured rates, the average per capita tax increase would be about $130 (in 1993 dollars).
The Long and Marquis analysis illustrates that very unequal programs among the states would result if each state financed a program with a budget limited to its capacity to finance health reform. As Figure 3 shows, the 24 states with the smaller percentage of uninsured would essentially be able to finance the full insurance reform plan because they are also the states with the highest tax capacity. The 24 states with the larger percentage of uninsured would not be able to cover all of their low-income uninsured population with a budget limited to their estimated capacity to finance health system reform.
In sum: the states that most need to expand insurance coverage have the smallest capacity to do so. As a consequence, a strategy relying on incremental, state-by-state action is likely to leave the nation with significant lingering gaps in health care coverage. Some states may need targeted federal assistance—for example, a program like CHIP, which provides federal matching funds to help states implement expanded coverage.
The research by Long and Marquis is helping individual states develop and implement changes in health care financing and delivery that will lead to improved access for the uninsured. But the researchers emphasize that other factors will affect program costs and cost-effectiveness.
For example, on the one hand, not all the people who are eligible enroll in public programs, and participation rates may vary among states because of differences in program implementation. As a result, estimates of program effects may overstate both the number of uninsured who will actually be covered and program costs.
On the other hand, the public program may "crowd out" private insurance. Some families may shift from employer plans to a public program because the latter is cheaper. Or families may lose the opportunity to purchase private insurance—for example, if some employers stop offering insurance because they know that employees can be covered by the public program.
This kind of crowd-out could increase program costs. Moreover, the program would not be reaching the target population—the uninsured.
Long and Marquis will use data from before and after the expanded programs to cast light on these, and other, critical issues.