Call it the reverse Midas touch—anything the government touches withers and diminishes in quality. One oft-cited example of this phenomenon is healthcare, where programs such as the Indian Health Service, Veterans Affairs, and Medicaid are scandalized by high costs and lackluster outcomes. But what’s interesting is that there seems to be a notable exception to the rule—Medicare, America’s predominant insurer for citizens aged 65 and older.  However, when looking closer, Medicare isn’t exactly a harbinger of health.

Medicare receives high marks of approval from seniors and seems to avoid low reimbursement rates for healthcare providers and some of the bureaucratic hurdles associated with Medicaid (federal insurance for low-income Americans). But there’s one just one problem: the evidence that Medicare works a whole lot better than flailing programs such as Medicaid just isn’t that strong, despite higher reimbursements and widespread political support. Perhaps it’s time for policymakers to rethink their assumptions about healthcare and strive for solutions that actually keep people healthy.

Ever since large, empirical examinations of healthcare programs have come into vogue, Medicare has proven to be the ugly stepchild of research for economists and policy scholars alike. The federal program has been a staple for America’s seniors since its inception in 1965. This 50-year history, and the lack of a secondary healthcare option for seniors, makes it difficult to compare similar populations with and without access to the federal program. But, a 2019 paper by University of Southern California and University of Wisconsin scholars tries to tease out the effects anyway, examining 15 years of cancer detection and mortality data for Americans aged 59-71.

The scholars examine whether there was a sharp break in outcomes at 65, creating a discontinuity in the age trendlines for cancer rates. In the words of the authors, the results speak for themselves: “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.”

But a look at the regression tables paints a considerably more complicated picture. In Table 1, the researchers parse cancer mortality data by gender, and further break down the evidence by tumor site. Interestingly, significant effects on cancer mortality at age 65 are only found for women. But, when the data is cut into two time periods (to account for the introduction of Medicare Part D in 2006), neither time period shows significant mortality changes for either gender as individuals turn 65. And, of the female population, only lung cancer (not breast nor colorectal) shows statistically significant mortality declines at that age. These findings are particularly strange, given significant, across-the-board findings for increases in cancer detection rates. The scholars note, “at age 65, cancer detection increased by 72 per 100,000 population among women and 33 per 100,000 population among men;” This increase shows up in virtually all the tumor sites studied, except, bizarrely enough, for lung cancer in women!

These results are murky to say the least. Perhaps Medicare indeed works its magic on lung cancer, without doing much for other studied cancers. But one can easily imagine a scenario in which large segments of the workforce quit smoking shortly after retirement (supported in the data), which significantly increases lung health. As for the other cancers, this is just further confirmation that cancer screening on a population level hasn’t really been shown to save lives. For instance, declines in breast cancer mortality in recent years have been just as large in non-screened groups as in screened groups. In 2016, a team of researchers from Maulana Azad Medical College explained the issue: “The problem with the presently available methods of early detection of cancers is that they are not very sensitive and specific and often lead to false positive results which may adversely affect the quality of life of the affected persons, including causing psychological problems.” And if screening detects a late-stage cancer, well, it is often already too late.

It is unwise to generalize conclusions based on the results too much, and screenings for certain cancers undoubtedly save lives. But the disappointing evidence linking health insurance to better health outcomes speaks to a larger need for policymakers to embrace early-life interventions and encourage healthy living before screening and radiation become the last resort. For example, instead of spending billions of dollars in funding greater healthcare utilization, federal policymakers can keep and approve more reduced-risk products on the shelves for cigarette smokers looking for safer alternatives to their deadly habits. When governments use restrictions to curtail access to life-extending foods and products, the reverse Midas touch is difficult to undo.