Understanding State Variation In Health Insurance Dynamics Can Help Tailor Enrollment Strategies For ACA Expansion

By John A. Graves and Katherine Swartz
Health Affairs, September 2013 (online)


The number and types of people who become eligible for and enroll in the Affordable Care Act’s (ACA’s) health insurance expansions will depend in part on the factors that cause people to become uninsured for different lengths of time. We used a small-area estimation approach to estimate differences across states in percentages of adults losing health insurance and in lengths of their uninsured spells. We found that nearly 50 percent of the nonelderly adult population in Florida, Nevada, New Mexico, and Texas—but only 18 percent in Massachusetts and 22 percent in Vermont—experienced an uninsured spell between 2009 and 2012. Compared to people who lost private coverage, those with public insurance were more likely to experience an uninsured spell, but their spells of uninsurance were shorter. We categorized states based on estimated incidence of uninsured spells and the spells’ duration. States should tailor their enrollment outreach and retention efforts for the ACA’s coverage expansions to address their own mix of types of coverage lost and durations of uninsured spells.

Policy Implications

The planning for implementing the ACA’s coverage expansions has largely focused on the percentage of people in each state who are uninsured at a certain point in time. In particular, attention has centered on decomposing these state percentages into people eligible for Medicaid coverage and people eligible for premium subsidies if they purchase plans in an exchange. Two states with the same percentage of uninsured people could have quite different proportions of people who are eligible for Medicaid, for premium subsidies, or neither.

Two states with similar uninsurance rates could also have very different uninsured populations in terms of how long adults had been without coverage. Differences in length of time without insurance are just as important as differences in income for policy makers. They must determine whether to focus more on efforts to minimize possible Medicaid churning or on efforts to reduce potential adverse selection—that is, efforts to encourage healthy uninsured people to buy exchange plans, so the plans do not have a disproportionate share of enrollees who require expensive medical care in the near future. States with a high incidence of people losing public coverage, for example, could have more reason to worry about Medicaid churning than states with a high proportion of uninsured spells that last more than two years, where possible adverse selection in their exchanges would be a more likely problem.


This study shows that it is not only the income level but it is also the length of time that an individual has been uninsured that affects the probability of whether the individual would be eligible for Medicaid or for the state insurance exchanges instead. Since there is considerable variation between states, the authors suggest that enrollment strategies be tailored to target the uninsured based not just on the types of coverage lost but also on the duration of being uninsured. What?

The dynamics of public or private insurance eligibility are forever changing. Eligibility depends on employment, income, age, geographical location, level of state participation in Medicaid, immigration status, and other factors that are frequently in flux. The navigator program was established to help individuals sort out the eligibility criteria in order to move them into appropriate health care coverage.

It does not seem logical that states should measure durations of being uninsured and then change enrollment strategies simply because people who previously had been in public programs had shorter periods of being uninsured. Navigators help all uninsured (except the tens of millions not able to enroll in any program) whereas targeted programs are aimed at only selected populations of the uninsured, tacitly acknowledging the difficulties of getting everyone covered and retaining that coverage.

We are making this too complicated. Much of this is due to the fact that we are merely expanding a highly fragmented and dysfunctional system of financing health care that is too costly and doesn’t work very well. It is ridiculous that we should consider looking at all of the fluid variables and then target some individuals while leaving others behind. Instead of trying to sort all of this out we should merely switch to a single payer national health program in which everyone is simply enrolled once, for life.