Expanding health insurance coverage may improve health care access [PDF] and reduce financial stress [PDF].
Ideally, having health insurance and the resultant access to care should improve health outcomes and well-being, although the evidence is complicated and mixed. One thing is sure: expanded insurance coverage typically leads to more utilization – a concern for policymakers and administrators because of effects on budgets, planning, and supply of health care services.
Oregon’s 2008 Medicaid Expansion, the “Oregon Experiment,” has been used to assess how health care utilization changes after people receive Medicaid coverage and how insured and uninsured individuals differ in their use of health care services. Many studies have found that adults who gained coverage used more services including ED, primary care, and specialty care, after becoming insured. Expansions in Wisconsin and Massachusetts have also been studied, with some increases in utilization but some increases that simply fell in line with previous trends. All of these previous studies looked at utilization during the first 12-15 months after individuals gained coverage.
While 12-15 months seems like a long time, it may take substantially longer for individuals, particularly those with complex cases, to establish care and bring health conditions under control. Additionally, other work has suggested that individuals who “churn” on and off coverage may behave differently than individuals who are newly insured. In a recently published Medical Care article, Jean O’Malley and colleagues use data from the Oregon Experiment to compare individuals based on their insurance history for 24 months after gaining coverage.
The authors studied a retrospective cohort of individuals who were continuously covered by Medicaid for 24 months after Oregon’s coverage expansion. They divided the sample into three groups: (1) continuously insured for at least 12 month pre-expansion, (2) returning insured who had coverage at some point in 2005-2006, and (3) newly insured without any coverage in 2005-2006. To summarize and adjust for differences among the groups, the authors used propensity score matching.
Claims data were used to evaluate visit rates for ED, primary care, mental and behavioral health care, and specialist care. Both annual rates and changes in rates using 3-month windows were evaluated. The authors also controlled for a number of characteristics that are associated with health care utilization. So, what happened to utilization?
- Primary care: Both returning insured and newly insured individuals had higher visit rates than the continuously insured for 4-12 months, but not after 12 months.
- ED: Returning insured individuals had much higher visit rates than newly insured and continuously insured individuals, and this continued over time.
- Mental and behavioral health care: No significant differences were seen in the first 12 months, but returning insured and newly insured individuals subsequently had higher visit rates.
- Specialty care: Both returning insured and newly insured individuals had higher visit rates than the continuously insured, particularly in the second year, and this continued over time.
What do these findings suggest? As pointed out by the authors, the higher utilization may point to differences in the severity of illness between the groups that is otherwise unobserved and not dealt with by propensity scoring methods. It may also be the case that individuals who experience churn may have more complex needs, including social and behavioral, that may put them at risk for coverage instability. Some of the increases in utilization seen within 12 months of expansion do not persist over time–so longer follow-up periods may be needed to assess the effects of coverage expansions. Other results showed high utilization of safety-net providers for outpatient visits — about 20% of enrollees, with the percentage increasing over time. Visit rates for new primary care providers (establishing care) remained high for returning and newly insured patients, suggesting that establishing consistent care may take a while for the Medicaid population.
The article by O’Malley and colleagues provides good evidence that the effects of expansion in coverage on utilization are more nuanced than previously thought. Their estimates may help states and health care organizations plan for the increased demand from Medicaid coverage expansion based on the insurance history of their constituents. Given the relatively high utilization among individuals with unstable coverage, further work evaluating how churn can be reduced and the likely consequences for utilization after 2 years will be important in evaluating the outcomes of the ACA.
Jess Williams