Facebook Twitter Tumblr Close Skip to main content
A Project of The Annenberg Public Policy Center

Deaths from a Health Care Bill?


House Minority Leader Nancy Pelosi has said “hundreds of thousands of people will die” if the Senate health care bill becomes law. But what does the research say about the impact of health insurance on mortality rates?

There are several studies backing up the idea that those who lack health insurance have a higher chance of dying prematurely than those with insurance, as we found when we looked at this issue in 2009.

But the research uses terms like “could” and “suggests” and “cannot definitively demonstrate a causal relationship,” not the definitive “will” favored by opponents of the bill. We can’t say whether any specific projection is a correct or valid number.

Democrats have made this a talking point. “We do know that many more people, hundreds of thousands of people, will die if this bill passes,” Pelosi said on “CBS This Morning” on June 26. In a July 2 interview on CNN’s “State of the Union,” Sen. Bernie Sanders said “tens of thousands of people every single year will die” if the Affordable Care Act were repealed entirely, without a replacement.

Later in that interview, Sanders said that “no one knows exactly the number” but that one study said “up to 28,000 people a year … will die” from the increase in the uninsured under the Republican bills. He and Pelosi are referring to an analysis by the left-leaning Center for American Progress.

We’ll explain how that study reached its conclusions, as well as what other research has found.

Pelosi’s Number

The analysis by CAP concluded that if coverage losses in the Senate bill mirrored losses in the House bill, then the legislation “would result in 217,000 additional deaths over the next decade.”

At the time Pelosi made her statement, the Congressional Budget Office hadn’t yet released its analysis of the Senate’s Better Care Reconciliation Act, but it did later in the same day. The CBO concluded that the Senate bill would increase the number of uninsured by 22 million in 2026, compared with current law. The CBO’s estimate for the impact of the House health care bill, the American Health Care Act, had been 23 million more uninsured in 2026, compared with current law.

Using the 23 million figure from CBO and a study on mortality in Massachusetts after it passed sweeping health care changes in 2006, the CAP report extrapolated that “the coverage losses from the Senate bill would result in 27,700 additional deaths in 2026” and 217,000 over 10 years.

The CAP report was written by a professor of social epidemiology and two graduates of the Harvard T.H. Chan School of Public Health, as well as two CAP staffers. (The Harvard researchers later gave updated figures for the Senate bill — 26,500 extra deaths in 2026 and “208,500 unnecessary deaths” over a decade.)

The CAP report has limitations, which the authors acknowledged. And its specific figures are the midpoint of a range. There’s a 95 percent confidence interval associated with the 27,711 excess deaths estimate, meaning, “we are 95 percent confident that the true number of annual excess deaths will be between 9,583 and 46,000,” the authors explain.

What are the limitations to those numbers? The CAP report took a study about Massachusetts and applied it to the entire country, and it took a study about the impact of increasing those with insurance and applied it to the opposite scenario — the number with insurance declining under the GOP health care bills.

The Massachusetts study, published in the Annals of Internal Medicine in May 2014, compared changes in mortality rates for adults age 20 to 64 in the state from 2001 to 2005 and 2007 to 2010 to a control group of counties in other states. Massachusetts passed health care legislation in 2006, often called Romneycare, that was similar in many ways to the federal Affordable Care Act, passed in 2010.

The conclusion of the Massachusetts study: The state’s health care law was associated with a decrease in deaths.

“Changes in Mortality After Massachusetts Health Care Reform: A Quasi-experimental Study,” May 6, 2014: The Massachusetts 2006 health care reform was associated with significant reductions in all-cause mortality over 4 years of follow-up relative to a control group of similar counties in states without reform. Reductions were concentrated in causes of death that were more plausibly amenable to health care and in populations most likely to benefit from expanded access, particularly residents of counties with lower incomes and higher prereform uninsured rates.

The authors found a drop in mortality of 8.2 per 100,000 adults and said that — along with an increase in insurance coverage in the state — implied for “approximately every 830 adults who gained insurance, there was 1 fewer death per year.”

The CAP report simply did the math — applying that finding to the country at large using the CBO’s estimates for the increase in the uninsured under the GOP bills.

The lead author of the Massachusetts study, Benjamin D. Sommers, an associate professor of health policy and economics at Harvard, told us the CAP report was “a reasonable attempt to project mortality effects based on prior research.”

“The authors do mention some of the main limitations of this approach that I also would have pointed out — namely, that Massachusetts is of course not identical to the U.S. as a whole, and withdrawing coverage may not have produced mirror image effects as expanding coverage,” Sommers said.

The Massachusetts study noted that the state differed from the rest of the country in many ways, including having lower mortality, higher income and the highest per-capita number of physicians. “The extent to which our results generalize to the United States as a whole is therefore unclear,” the authors wrote.

“But the ACA was modeled after Massachusetts health reform, so I do think it’s probably the best gauge we have for this projection,” Sommers said. “It’s hard to pin down any specific number with certainty, but I think the CAP projections are at least plausible.”

Sommers said he would use less definitive language than the politicians. He said Pelosi was “taking what is a reasonable projection and stating it as a fact; I would have used the caveats above and been a bit less definitive — but I’m a researcher not a politician!”

There are also limitations to the study on Massachusetts, as with most studies. It’s difficult in research to prove causality — that one thing directly causes another — and that was the case here. The authors noted that “we do not have individual-level insurance information and thus cannot directly link mortality changes to persons gaining insurance coverage.” They said it was possible that the reduction in mortality was caused by other factors in the state, though the study controlled for economic measures and didn’t find that mortality similarly declined for elderly adults in the state — who would have been covered by Medicare and largely unaffected by the health care changes.

What Other Research Says

Brendan Saloner, an assistant professor of health policy and management at the Johns Hopkins Bloomberg School of Public Health, told us that “no single study is as good as looking at the full body of research.” And while not all studies have found health improvement related to insurance coverage, the “preponderance of studies, especially the well-done studies, find that gaining insurance coverage, especially for low-income people, improves health and reduces mortality risks.”

Let’s look at some of the research:

  • A 2017 review of studies over the past decade — published in the New England Journal of Medicine with Sommers as the lead author — concluded: “Insurance coverage increases access to care and improves a wide range of health outcomes. Arguing that health insurance coverage doesn’t improve health is simply inconsistent with the evidence.” On the specific issue of whether coverage saves lives, the review described three recent studies, including the Massachusetts study. We’ll describe the other two, on Medicaid expansion, below.
  • 2012 study published in the New England Journal of Medicine, also with Sommers as the lead author, compared three states that substantially expanded Medicaid (before the Affordable Care Act) to neighboring states that did not expand Medicaid. It concluded, “State Medicaid expansions to cover low-income adults were significantly associated with reduced mortality as well as improved coverage, access to care, and self-reported health.” The study looked at five years before and after expansions of Medicaid eligibility in New York, Maine and Arizona. Among the limitations: “We examined three expansion states, and the results are largely driven by the largest (New York), so our results may not be generalizable to other states.”
  • The Oregon Health Insurance Experiment, published in the New England Journal of Medicine in 2013, provided mixed results. The study, whose lead author, Harvard’s Katherine Baicker was a coauthor on the two studies above, compared data from 6,387 adults who were able to apply for Medicaid coverage in Oregon through a lottery drawing to 5,842 adults who were not selected. The authors concluded: “This randomized, controlled study showed that Medicaid coverage generated no significant improvements in measured physical health outcomes in the first 2 years, but it did increase use of health care services, raise rates of diabetes detection and management, lower rates of depression, and reduce financial strain.”
    The study measured three health indicators — blood pressure, cholesterol and glycated hemoglobin levels (which measure diabetic blood sugar control) — over a two-year period. The time period studied was among its limitations; it said “the effects of insurance in the longer run may differ.” The study didn’t address mortality rates; an earlier working paper on the first year of the Oregon experiment said “not surprisingly” researchers didn’t see “any statistically significant improvement in survival probability” during the short time frame with few deaths.
  • A 2014 study published in the Journal of Clinical Oncology looked at cancer diagnoses among 39,447 people age 20 to 40 between 2007 and 2009 and concluded that “young adults who are uninsured may be more likely to present with metastatic disease, be undertreated, and die after a diagnosis of cancer relative to those who are insured.”
  • A 2017 study in the journal Medical Care determined that the provision in the ACA allowing young adults to stay on their parents’ plan until age 26 “was associated with a 6.1% decline in monthly disease-related mortality” and no significant difference in external-trauma mortality, such as from accidents. The study compared mortality in 19- to 25-year-olds to that of a control group of 26- to 30-year-olds in 2008-2013.
  • Among older research: A 2004 study published in Health Affairs found that the uninsured among those age 55 to 64 had a 3 percent higher risk of dying over an eight-year period, calculating that 13,000 yearly deaths “may be attributable to the present lack of insurance coverage among the near-elderly.” It used data from the Health and Retirement Study, sponsored by the National Institute on Aging, and controlled for socioeconomic factors.
  • In March 2009, Dr. John Z. Ayanian of the National Academies’ Institute of Medicine testified to Congress on the review by the Committee on Health Insurance Status and Its Consequences of nearly 100 studies released since 2002. He described the uninsured as being more likely to delay or forgo preventive care, medications and treatment. His written testimony said: “Uninsured adults are also more likely to be diagnosed with later-stage cancers compared to their insured peers. If hospitalized for a serious acute condition, such as a heart attack, stroke, or major trauma, uninsured adults are more likely to die after admission to a hospital. Uninsured adults are 25 percent more likely to die prematurely than insured adults overall, and with serious conditions such as heart disease, diabetes or cancer, their risk of premature death can be 40 to 50 percent higher. Fortunately, our Committee also found good news to report: when uninsured people acquire health insurance they can experience both immediate and long-term improvements in their health.”
  • In 2002, the Institute of Medicine, drawing on previous research that found a higher mortality risk for the uninsured, determined that 18,000 nonelderly adults died because they lacked health insurance in 2000. A 2009 study, published in the journal Health Services Research and authored by Richard Kronick, who later headed the Agency for Healthcare Research and Quality, said that finding was “almost certainly incorrect.” Kronick’s study found no difference in mortality between the uninsured and those with employer-sponsored insurance after adjusting for demographic, health status and health behavior factors.

This body of research looked at the impact of already being uninsured, or the impact of gaining coverage. The CAP report, as we noted, examined the opposite situation — declining insurance coverage.

That’s a challenge in applying past research to the proposed health care legislation. “We’ve never seen large scale reductions in insurance coverage” such as those projected by the CBO, Johns Hopkins’ Saloner notes.

The 2017 review of studies in the New England Journal of Medicine noted that critics of the ACA have contended that benefits from insurance gains depend on the type of coverage — that Medicaid coverage isn’t beneficial, while private insurance is. (That’s one argument in a February paper published by the conservative Manhattan Institute that said the “best statistical estimate” for lives saved by the ACA was “zero.”)

“But there is no large quasi-experimental or randomized trial demonstrating unique health benefits of private insurance,” the NEJM review said. It said that “further research is needed to assess the relative effects of various insurance providers and plan designs.”