By Rebecca Mary Myerson, Reginald Tucker-Seeley, Dana Goldman, and Darius N. Lakdawalla
National Bureau of Economic Research, September 2019
Medicare is the largest government insurance program in the United States, providing coverage for over 60 million people in 2018. This paper analyzes the effects of Medicare insurance on health for a group of people in urgent need of medical care – people with cancer. We used a regression discontinuity design to assess impacts of near-universal Medicare insurance at age 65 on cancer detection and outcomes, using population-based cancer registries and vital statistics data. Our analysis focused on the three tumor sites with recommended screening before and after age 65: breast, colorectal, and lung cancer. At age 65, cancer detection increased by 72 per 100,000 population among women and 33 per 100,000 population among men; cancer mortality also decreased by 9 per 100,000 population for women but did not significantly change for men. In a placebo check, we found no comparable changes at age 65 in Canada. This study provides the first evidence to our knowledge that near-universal access to Medicare at age 65 is associated with improvements in population-level cancer mortality, and provides new evidence on the differences in the impact of health insurance by gender.
From Comparison with the literature
Medicare beneficiaries can choose to receive their benefits via traditional Medicare (the publicly administered Medicare plan) or Medicare Advantage (a Medicare plan paid for by the federal government but administered by a private company). After a cancer diagnosis, patients become less likely to leave traditional Medicare for a private Medicare Advantage plan, and become more likely to switch from a private Medicare Advantage plan to traditional Medicare.
We are not aware of any prior study of the effect of Medicare coverage on population-level cancer mortality. The most closely related study to ours focused on the impact of Medicare on post-detection survival, a different outcome from the one we study, using data from a different time period.
Importantly, the outcome of survival after cancer detection employed in these prior studies may be subject to diagnosis bias. Diagnosis bias includes lead and length time bias, which can be explained as follows. First, when people are diagnosed with cancer earlier, they may appear to survive longer after detection simply due to becoming classified as a cancer patient earlier – i.e., “lead-time bias.” Second, the additional tumors detected might be so slow-growing that they would never have killed the patient if left undetected, resulting in overdiagnosis and overtreatment – i.e., “length bias.”
For both these reasons, expansions in cancer detection may improve post-diagnosis survival even when they do not actually improve health or save lives.
As a result, analyzing population-level disease-specific mortality rates is considered a best practice to eliminate diagnosis bias.
Prior studies of the impacts of Medicare rarely stratified the data by gender except when studying gender-specific health care such as mammography. Yet, it is plausible that gender could play an important role in determining the impact of insurance on health. Several studies have found women to be more likely than men to use preventive health care, and less likely than men to delay seeking needed health care. Gaps by gender in socio-economic resources could also play a role, as the impact of Medicare could be larger among patients with less-generous prior insurance coverage and fewer financial resources.
From the Discussion
Insurance coverage rose to 97% at age 65, the age of near-universal eligibility for Medicare. This nearly universal Medicare coverage increased cancer detection by 50 per 100,000 population, a 10% increase compared to people aged 63-64; the majority of additional cancers detected were early-stage cancer. These findings are important for population health because prompt detection improves health for the tumor sites we study here, according to systematic reviews by the United States Preventive Services Task Force.
In vital statistics data, we found that cancer mortality increased by less than expected among women at age 65 by 9 per 100,000 population. The lack of significance among men is consistent with the magnitudes and standard errors we estimated. The increases in cancer detection and early-stage cancer detection at age 65 among men were both less than half the magnitude of the increases at age 65 among women.
Our analysis captures a local average treatment effect of the effect of access to Medicare at age 65, i.e., the effect of having access to Medicare compared to lacking access to Medicare at age 65. Given that access to Medicare is nearly universal at age 65 but severely limited prior to age 65, a research design such as ours which compares a “treated” group aged 65 and older to an “untreated” group younger than 65 has been deemed appropriate in closely related studies. While access to Medicare may be particularly beneficial for the previously uninsured, the previously insured also benefit. Within a few weeks of becoming eligible for Medicare, there is a sharp increase in the proportion of people with multiple forms of coverage, and the proportion of people with only managed care is reduced by half. Thus, Medicare provides the already-insured with access to more generous coverage and a broad network of providers. Generosity of coverage and network breadth are particularly important for cancer patients, who require costly specialty care.
In conclusion, access to Medicare insurance was associated with a significant increase in detection of cancers with recommended screening, as well a decline in mortality from these cancers among women. Our estimates provide new evidence of Medicare’s impact on health outcomes for people in need of medical care.
By Don McCanne, M.D.
This study of breast, colorectal and lung cancers demonstrated that enrollment in Medicare at age 65 is associated with a significant increase in detection of these cancers with recommended screening, as well as a decline in mortality from these cancers among women. Although mortality did not improve in males, there are other benefits of having coverage through Medicare: for instance, health care can improve quality of life in the presence of other infirmities and through preventive measures, and Medicare can reduce financial hardship caused by medical bills.
One interesting observation is that, when diagnosed with cancer, patients become less likely to leave traditional Medicare for a private Medicare Advantage plan, and become more likely to switch from a private Medicare Advantage plan to traditional Medicare. That should make us wonder why we taxpayers are paying more for the private Medicare Advantage plans and why so many politicians and bureaucrats are trying to push us into those programs, especially when most people want more choices, not less, in the experts and institutions that care for cancer patients.
The main point of today’s message is that Medicare for All is the right choice for reform. People do better after they are on Medicare than they did under our existing, fragmented multipayer system. Further, the single payer model of Medicare for All provides better health care benefits and greater financial protection than does the current Medicare program.
How can people say that they want to keep the plans that aren’t working as well? Don’t they understand what Medicare is all about, and, more importantly, what it could be?
Stay informed! Visit www.pnhp.org/qotd to sign up for daily email updates.