"One of the main arguments proponents of Medicaid expansion make, at least on the fiscal side, is that it would save money as people gaining Medicaid coverage would reduce their use of expensive visits to the Emergency Department (ED). An earlier study from the Oregon Health Insurance Experiment threw some cold water on that theory, as it found that getting Medicaid actually increased the number of ED visits by 40 percent. Some analysts postulated that this increase was only temporary because it was due to either pent-up demand for health care services, or because new enrollees did not have established relationships with doctors. The thinking was that after enrollees became familiar with their coverage or addressed long-gestating health problems, the reductions in ED use and the associated cost savings would materialize.
A new report analyzing a longer time horizon finds that this is not the case, and there is “no evidence that the increase in ED use due to Medicaid coverage is driven by pent-up demand that dissipates over time; the effect on ED use appears to persist over the first two years.” This is another blow to the oft-repeated claim that Medicaid expansion will lead to significant savings from reduced Emergency Department utilization, and the effect actually seems to work in the other direction.
The Oregon case is important because it is one of the few instances of random assignment in health insurance, as the state wanted to expand Medicaid but had funding constraints, so it employed a lottery to determine who would get coverage.
In this new update, the researchers see if there are any time patterns or signs of dissipation when it comes to the impact of Medicaid percent of the population with an ED visit or the number of ED visits per person. They expand upon the earlier utilization study to analyze the two years following the 2008 lottery and break up into six-month segments to see if there are any signs of the effects dissipating.
As they explain, “there is no statistical or substantive evidence of any time pattern in the effect on ED use on either variable.” In the first six-month tranche Medicaid coverage increased the number of ED visits per person by about 65 percent relative to the control group, and the estimates for the following three periods were similar and mostly statistically indistinguishable from each other. They also find that Medicaid increases the probability of an ED visit in the first period by nine percent, and the impact in the subsequent periods does not differ significantly.
Estimated Effect of Medicaid Coverage on Emergency Department Use over Time
Source: Finkelstein et al., New England Journal of Medicine (2016).
A similar analysis for hospital admissions comes to the same result: “no evidence statistically or substantively of a time trend in the impact of Medicaid on having a hospital admission or on the number of days in the hospital.”
They also fail to find evidence that Medicaid coverage makes doctor’s visits and ED use more substitutable, and again, if anything the effect works in the other direction. The authors suggest that this could be due to differences in how people respond to gaining access to Medicaid, or it could be that Medicaid increases the complementarity of these different dimensions of health care.
Whatever the underlying mechanism, there is no evidence here that Medicaid coverage leads to reductions in utilization in other dimensions. In fact, Medicaid coverage makes people more likely to visit the Emergency Department, and increases their number of visits relative to their baseline. This new study confirms that these effects were not temporary and do not dissipate, at least over the two year period they were able to analyze. Expanding Medicaid coverage will not lead to reductions in inefficient, inexpensive Emergency Department visits, and there will be no associated cost savings, undermining one of the common fiscal arguments for expansion."
Saturday, October 22, 2016
Surge in Emergency Department Use Persists in New Oregon Medicaid Study
By Charles Hughes of Cato.