Research Update: NBER Studies Look at Medicaid Expansion and Mortality and Impact of Cost-Sharing On Prescription Drug Use

Two recent working papers from the National Bureau of Economic Research (NBER) demonstrate how health policy decisions can affect mortality.  The first examines how the Affordable Care Act’s Medicaid expansion reduces mortality and the second examines the mortality effects of cost-sharing for prescription drugs.

Medicaid and Mortality: New Evidence from Linked Survey and Administrative Data, NBER

In this paper, researchers at the University of Michigan, the US Census Bureau, and New York University use data from the American Community Survey (ACS), the Census Numident file, and CMS administrative records of Medicaid enrollment, to examine changes in mortality among near-elderly adults aged 55-64 in states that have adopted the Affordable Care Act (ACA) Medicaid expansion. They also use data from the Mortality Disparities in American Communities (MDAC) project to examine the causes of death behind changes in mortality.

Key Findings:

  • In each of the four years following Medicaid expansion in 2014, between 49 and 51 percent of adults aged 55-64 gained Medicaid eligibility in expansion states, relative to the year just before expansion. The probability that someone was enrolled in Medicaid also increased by 12.8 percentage points, and individuals spent an average of 43 additional days of the year enrolled compared to similar adults in non-expansion states.
  • Before Medicaid expansions, near-elderly adults in expansion and non-expansion states had similar mortality rates. In the first year after Medicaid expansion, the probability of mortality among adults in expansion states decreased significantly by 7 percent. Moreover, the probability of mortality continued decreasing in expansion states: by the fourth year following expansion, an adult aged 55-64 in an expansion state was 11.9 percent less likely to die that year, even though the sample continued to age.
  • This change in mortality was driven by significant declines in “internal,” disease-related deaths (like cardiovascular disease or diabetes) in the first two years after expansion, as opposed to “external” deaths (like a car accident).
  • Overall, the authors estimate that Medicaid expansion led to 4,800 fewer deaths per year among adults aged 55-64, or about 19,200 fewer total deaths in the four years following expansion

Why This Matters:

  • As of this writing, twelve states have still not adopted the ACA’s Medicaid expansion. The authors estimate the cost of this decision at 15,600 additional deaths in the four-year period since expansions were first implemented in 2014.
  • The unique data linking process the researchers use allows them to track the same individuals through multiple surveys and datasets, providing rich individual-level data on economic characteristics, Medicaid enrollment, and mortality; previous studies have only been able to predict state- or county-level estimates or have relied on death certificate data that does not include socioeconomic or other information about the individual. Linking individuals’ survey records to Medicaid enrollment records also represents an important methodological correction for survey measures that significantly underestimate Medicaid coverage (compared to administrative data), allowing for more accurate estimates of mortality effects.

The Health Costs of Cost-Sharing, NBER

Researchers at Harvard University and the University of California, Berkeley use data on prescription drug cost-sharing and utilization among Medicare Part D beneficiaries who enroll in the year they turn 65 to estimate the impact of cost-sharing on mortality. At the time data were collected, the Medicare Part D drug benefit included a so-called “donut hole” or coverage gap: beneficiaries were required to pay 25 percent co-insurance up to a total drug spending threshold and then were responsible for 100 percent of the cost out-of-pocket (up to a catastrophic limit).  (This donut hole began to be phased out starting in 2011 and was fully closed in 2020.), Medicare eligibility begins in the month a person turns 65, and Part D benefits are not pro-rated or otherwise adjusted for the first year of enrollment to take into account when someone enrolls. This meant that someone turning 65 and enrolling in Medicare in February was likely to enter the donut hole earlier than someone born and enrolled in September and would likely spend more on prescription drugs out-of-pocket over the course of the year. The authors used this cost-sharing variation by eligibility month to estimate impact on mortality, by comparing those beneficiaries to 65-year-old enrollees who are dually eligible for Medicare and Medicaid.  Such enrollees should be enrolled in the Medicare Part D Low-Income Subsidy which requires only nominal cost-sharing.  (Many should also be eligible for the Qualified Medicare Beneficiary program which picks up Medicare cost-sharing charges for other services.)

Key Findings:

  • The authors estimate that a mere $10.40 increase in coinsurance led to a 32.7 percent increase in monthly mortality (measured in December).
  • A large share of this increase in mortality was explained by lower fills of inhalers and prescriptions that lower cholesterol, blood sugar, and blood pressure, which are typically classified as “high-value,” life-saving, drugs. Moreover, the authors used machine-learning tools to predict patients’ risk for health events like heart attacks or strokes, and found that the top one-third of patients at risk of a heart attack were over 280 percent more likely to stop filling cardiovascular prescriptions than the bottom two-thirds were.
  • More people chose to fill no prescriptions when faced with higher out-of-pocket costs, including 18 percent of enrollees overall (regardless of how many drugs they were on initially) and 72.6 percent of enrollees who were on four or more drugs at baseline.

Why This Matters:

  • The rationale behind higher cost-sharing is to discourage individuals from over-consuming health care items and services including unnecessary, low-value care. However, consistent with a wide body of research, this analysis finds that enrollees who cut back on care in response to higher out-of-pocket costs also reduce use of needed, high value care like prescription drugs (and increase the likelihood of death). Increases in cost-sharing or coinsurance rates forces patients to perform complex cost-benefit analyses that can have deadly consequences if, as these results suggest, patients stop filling life-saving prescriptions.
  • Dually-eligible Medicare-Medicaid enrollees are deemed eligible for the Medicare Part D Low-Income Subsidy (LIS) program, which provides Part D premium and cost-sharing assistance. LIS enrollees only had to pay nominal co-payments, even within the coverage gap when it was in effect.
Aubrianna Osorio is a Research Manager at the Georgetown University McCourt School of Public Policy’s Center for Children and Families.