In July, as the Senate was about to vote on one of its Affordable Care Act repeal bills, the Wall Street Journal editorial pages published a commentary by editorial writer Allysia Finley questioning the effectiveness of state Medicaid programs. On September 23, days before an expected Senate vote on the Graham-Cassidy repeal bill, Finley wrote that data on opioid overdoses and hospital admissions suggested that the states that expanded Medicaid were hit hardest by the opioid epidemic. Finley's thesis was that the Medicaid expansion fueled the epidemic. "Perhaps a rollback is what the doctor ordered," she wrote. People who want to reduce federal funding for Medicaid have seized on such claims.
In August, Andrew Goodman-Bacon, an economist at Vanderbilt University, and Emma Sandoe, a health policy PhD candidate at Harvard University, examined the Medicaid-opioid claim on the Health Affairs Blog in an article titled "Did Medicaid Expansion Cause the Opioid Epidemic? There's Little Evidence That It Did." I asked Sandoe, a former spokesperson for the US Centers for Medicare & Medicaid Services, about Finley's assertions. This interview has been condensed and lightly edited for clarity.
Q: Why did you tackle the issue of Medicaid and opioids on the Health Affairs Blog?
A: This argument that Medicaid could be a cause of the opioid epidemic has been lurking out there in the ether for a while, and I knew that the analysis needed some clarification. Andrew and I decided to act after a US senator alluded to it in a floor speech back in July during debate over repeal of the ACA. The argument relied on correlations that weren't entirely accurate. And now here it is again. It wasn't terribly surprising that that claim is still out there, and I guess I was expecting questions like yours.
Q: The Wall Street Journal piece echoed some of the same arguments. What issues do you have with the thinking behind them?
A: A lot of the arguments I saw on the internet relied on assumptions I knew to be untrue. For example, they said the Medicaid expansion went into effect in 2010 and suggested that many states implemented it then. The Medicaid expansion took effect January 1, 2014. The rise in opioid deaths did start to trend upward in 2010 across the country, but since the Medicaid expansion happened later, the expansion cannot be evidence for the rise in opioid deaths.
Q: What about the treatment timeline?
A: States have made uneven progress in treatment programs. Many states only started ramping up treatment in 2015 or 2016 and haven't fully implemented it, so most of the efforts around addiction treatment haven't been fully captured by the data yet. It's also important to remember that gaining coverage doesn't equal immediate treatment and help. It takes people a while to start seeing a doctor.
Q: The Journal looked at statewide death rates both pre- and post-Medicaid expansion. You took a different approach.
A: In addition to comparing raw death rates across states, we looked at county-level insurance rates and how they changed after the ACA took effect in 2014. We found that counties with the greatest increases in coverage had smaller increases in the death rate for opioid overdose. Counties with smaller increases in the coverage rate saw larger increases in the overdose death rate. In addition, for states that had the most severe opioid problems, the epidemic itself might have been a contributing factor in their decision to expand Medicaid. Because they were dealing with the problem and had residents who couldn't get treatment, officials may have tilted toward Medicaid expansion to give greater access to treatment covered by the program.
Q: The Journal repeatedly compared the Medicaid population to people with private insurance, which seems reasonable. You disagree with that approach. Why?
A: It's a false comparison. The Medicaid population is very different than the privately insured population. Medicaid covers a large, heterogeneous group, including children, pregnant women, seniors, and people with disabilities. It's not uncommon that people with disabilities also have chronic pain conditions that would mean they have been prescribed opioids. Taking prescription opioids over the long term puts people at higher risk for opioid addiction. It is inaccurate to compare a larger group of people with disabilities to a more generally healthy group.
Q: Has Medicaid played any role in intensifying the epidemic?
A: Much more research is needed, but we know that it's really hard to isolate Medicaid as a single contributing factor because there is so much going on in the opioid epidemic. Teasing out individual factors would take significant time to do in a rigorous way. Linking Medicaid to the opioid epidemic continues to be a political tool more than anything — a justification by some for larger cuts to the program. It's important to avoid adopting conclusions like that without rock-solid scientific evidence, and we certainly haven't seen data that bring us anywhere near that.
It's fair to say now, with the benefit of hindsight, that everyone in health care should have done more in the 1990s and 2000s to curb the epidemic. If we had a time machine and could go back, it would be helpful and benefit everyone if back then we could have started the kinds of treatment and controls that are being implemented now.