Medicaid and mortality

A brand new NBER working paper led by Sarah Miller just came out this morning

Our analysis compares changes in mortality for near-elderly adults in states with and without Affordable Care Act Medicaid expansions. We identify adults most likely to benefit using survey information on socioeconomic and citizenship status, and public program participation. We find a 0.13 percentage point decline in annual mortality, a 9.3 percent reduction over the sample mean, associated with Medicaid expansion for this population. The effect is driven by a reduction in disease-related deaths and grows over time. We find no evidence of differential pre-treatment trends in outcomes and no effects among placebo groups.

They follow a cohort of people who were between 55 and 64 in 2014 through 2017. They matched these individuals against death certificates. The decision by the Supreme Court to allow states to choose to expand or not expand Medicaid to the 138% FPL group created variation in geography and timing that the researchers leveraged.

In the first year following the coverage expansion, the probability of mortality declined by about 0.09 percentage points, or 6.4 percent relative to the sample mean. The estimated impact of the expansions increases over time, suggesting that prolonged exposure to Medicaid results in increasing health improvements. By year 4, residents of expansion states have an annual mortality rate that is 0.2 percentage points lower than their non-expansion state counterparts. In our supplemental analysis
using the MDAC data, we find evidence that healthcare amenable and internal causes of death were reduced by the expansions, but no evidence that deaths due to external causes, such as car accidents, fell.

There probably is a dosing effect and a population vulnerability effect. Longer Medicaid Expansion probably saves more lives than a short expansion.  This is important as the Oregon Medicaid study looked at only health outcomes over one or two years. The QJE paper on Oregon only looked at about 14 months worth of coverage instead of the potential for several years of coverage. The other big difference is that this NBER paper is looking at a subgroup (55-64 year olds) that is much older on average than the group that the Oregon study examined (mostly 20-50 year olds (73% of sample)). The NBER paper used the Oregon Health study’s public use data (awesome for replication and validity checks) for the same subset and found large but not statistically significant mortality changes as well. Age is a good predictor of general mortality risk. Higher risk groups should show mortality differences more readily than lower risk groups for a given intervention.

This working paper is showing that Medicaid expansion has significant mortality impacts for a population that is likely to have fairly high mortality rates. They estimate that the decision to not expand Medicaid has led to 15,000 excess deaths for this subgroup.

Medicaid matters. It improves financial wellbeing, improves mental health and lowers mortality rates.

14 replies
  1. 1
    wucheeman says:

    Hi, coming out of long-time lurking to join up and be a tester. While I’m here, I wanted to say a very loud thank you for the amazing amount of work and thought that you put into your commentary!

  2. 2

    @wucheeman: thank you. I get more out of writing than anyone here gets out of reading me, so I think this works out nicely for all concerned.

  3. 3
    jl says:

    Thanks for an important post and link to important research. I think the Oregon study was under powered and the sample frame was not that interesting from a policy or real world impact perspective, and worse, the lack of power was not properly indicated in the paper, so reporters, and pundits with an ax to grind put out misleading stories. The research reported here is an important corrective.

    I think the vast majority of NBER papers haven’t had any peer review, and I haven’t had time to look at it closely. If the results hold up, they are important.

  4. 4
    Duane says:

    Totally agree with the previous commenters. Thanks for informing us on this important study. Medicaid is vital. Every Democratic candidate should pound Republicans on this.

  5. 5
    Fair Economist says:

    Nearly 10% reduction in the number of deaths? That is jawdropping. I can’t think of any medical intervention with a similar effect. And it’s increasing over time! It’s like snapping your finger and making drug addiction disappear.

  6. 6
    Steeplejack says:

    They estimate that the decision to not expand Medicaid has led to 15,000 excess deaths for this subgroup.

    I’m reminded of a recent Twitter screed about corporate malfeasance. The author pointed out that we (rightfully) get enraged about serial killers with five, 10 or 20 victims but are blind to the fact that far worse monsters swim in the depths, killing their victims (not so) indirectly by polluting the environment, mismanaging health care and insurance, trading lives for profit and generally “screwing the little guy.”

  7. 7
    jl says:

    @Fair Economist: When I was doing research on business case ROI for treatment in health care industry, medical directors and execs said (off the record, after meeting were over) that in people in the age group 5 to 10 years before Medicare eligibility tended to be under treated.The buzz phrase was ‘patch ’em up and pass ’em on’ (to Medicare). Both profits and nonprofits did this.It was rationalized as a necessary triage policy driving by market and annual enrollment price competition considerations.

    Also,there is a considerable evidence that people who do not have access to continuity of care because of continuous coverage, have higher costs and worse health outcomes later because of unmet needs. So these results fit with other evidence.

  8. 8
    Tom Levenson says:

    I am so old that I can recall Megan McArdle trumpeting that Oregon study as proof that arglebargle health care who needs it.

  9. 9
    Chuck Betley says:

    @Tom Levenson: I, too, remember the so-called liberal media seizing on a preliminary finding as definitive proof that their ideological biases were correct.

  10. 10
    Kelly says:

    @Steeplejack: Insulin and epi pen price rises comes to mind. Would network news put up a cumulative death count from insulin rationing every night?

  11. 11
    Sam Dobermann says:

    This improvement will expand as time passes. For example diabetes is undiagnosed in many that have it. Late recognition will reduce deaths but not as much as white early treatment & many in their sample were older at the start of the study. The Oregon study was absurdly short to show decreased death rates. Still it did show an improvement in mental health which is absurdly discounted. Mental health is health.

    David, some one asked about what else was in the ACA. Has that been covered in detail but simply? I could do that if it would be seen. I’m usually on here so late that my comments aren’t read or knowing that I skip it but I can do a good explainer.
    I could send it to you to post where you like.

    I’ll check back here this evening.

  12. 12
    jl says:

    This study is also an example of what used to be called ‘the woodwork’ problem, back when universal access to care was considered a radical policy. The idea was that costs would soar because people who were not getting treatment would ‘come out of the woodwork’ and strain capacity. That reasoning, again, was rationalized by market forces, annual price competition for group policies. I’m not sure the bigshots who said this made sense. It was all buzzword CW stuff that they used to rationalize implicit and hidden rationing of care. Maybe the argument had merit in the narrow view of their current business model. But why was it relevant to national policy decisions that necessarily involved changes in the standard business model for insurers and providers? But, there were a lot of people who would simultaneously bemoan that their own health insurance industry business model was slowly falling apart, for both profits and non-profits, and then toss CW buzzword arguments off the top of their heads that inclined them to resist change. It’s like the end of Thelma and Louise, except they are creeping along at two miles an hour, the car is falling apart and might blow up before they reach the cliff.

    Anyway, the results is shown in an updated graph in a recent Big Picture Post. I don’t like using $ comparisons and prefer percent of GDP as the measure of cost of health care, but the pic is the same no matter which you use.

    Health spending and life expectancy

  13. 13
    Steeplejack (phone) says:


    Great examples.

  14. 14
    StringOnAStick says:

    @jl: I’ve been reading Barry Ritholtz for a long, long time. He’s one of the few Wall Street types a liberal can stand to read, meaning he’s not openly touting kill SS and Medicare or praising Ayn Rand.

Comments are closed.