A Project of The Annenberg Public Policy Center

Cancer Rates and Unjustified Conclusions

A number of opponents of new health care legislation, most recently our old friend Betsy McCaughey on "The Daily Show," have claimed that cancer survival rates are higher in the U.S. than in countries with nationalized health care. They conclude from this that the state of general health and health care quality in the U.S. must therefore be higher. Does the U.S. really have a higher cancer survival rate? And what does that mean about our health care system?

It’s certainly the case that we have higher survival rates than the United Kingdom and other countries with nationalized health care. Across the board, the United States boasts a higher five-year relative survival rate than the European average, according to a 2008 study in the British medical journal Lancet. For breast cancer, for instance, the U.S. survival rate was 83.9 percent, the U.K. rate was 69.7, and the average European rate was 73.1.

But survival rates also differ within the United States, between insured and uninsured populations. The American Cancer Society found that the five-year survival rates for colorectal cancer averaged 63 percent for the privately insured but 49 percent for the uninsured. According to the Lancet study, five-year relative survival rates for colorectal cancer were 59.1 percent in the U.S. and 45.3 percent in Europe. Breast cancer survival rates among the uninsured were also similar to Europe – 85 percent survival for those with private insurance, 75 percent for the uninsured, close to the European average. Rates for people on Medicaid were similar to the uninsured.

So universal insurance is as bad as no insurance, right? Not so fast. For one thing, survival rates in Canada, Japan, Australia and Cuba were all comparable to or higher than U.S. survival rates on all types of cancer that the Lancet study examined, except for prostate cancer. Those countries all have some form of government-provided health care coverage. Prostate cancer often doesn’t require treatment, so the aggressive screening common in the U.S. turns up both early cases and cases that would never need intervention. This leads to an inflated survival rate in the U.S., where asymptomatic patients are more likely to be diagnosed.

Furthermore, simply comparing survival rates isn’t necessarily an accurate measure, and it certainly isn’t a simple reflection of health care quality. For one thing, five-year relative survival rates measure how many people diagnosed with cancer are still alive five years later (compared with how many people in that population you’d expect to be alive if they didn’t have cancer). This means that early detection will always increase survival rates, even if it doesn’t improve outcomes (though oncologists do agree that cancer is less deadly if found early). If two people have exactly the same disease progression, the one who’s diagnosed earlier will be more likely to be alive in five years. Thus, countries with more advance screening will have higher survival rates even if they don’t have better post-diagnosis care. Cancer screening is less widespread in European countries, and people without insurance or who are on Medicaid also are less likely to have access to it. The ACS study showed that the uninsured and Medicaid beneficiaries had lower rates of both mammograms and colorectal cancer screening than the insured.

Lower screening for Medicaid patients, the researchers theorized, came in part from underfunding of the program and differences in implementation of the Breast and Cervical Cancer Prevention and Treatment Act between states. It’s also the case that people who are diagnosed with serious illnesses, including cancer, can become eligible for Medicaid at higher income levels, meaning that some beneficiaries might have entered the program after their diagnosis. According to the study, "[l]ater stage at diagnosis among patients enrolled as a result of diagnosis does not reflect the extent to which Medicaid insurance provides access to care, including prevention and early detection." Some, but not all, studies on Medicaid enrollment and screening have taken the time of enrollment into account. One of those that did found that about a third of the sample enrolled the same month as diagnosis or after diagnosis.

Dr. Marie Diener-West, a professor of biostatistics at Johns Hopkins University Bloomberg School of Public Health, told us that it would be a stretch to draw too many conclusions from comparing survival rates. "Part of the problem with the comparison is that it might not actually be comparable populations," she said. "It could be [one is] an older population, it could be they have more comorbidities [other conditions] that are affecting their survival in addition to cancer, there could be occupational differences. There are many different factors that could be playing a role." Diener-West pointed out that the uninsured, for instance, are generally poorer and may have different diets, different lifestyles and different exposure to tobacco and other drugs than the privately insured. And when you compare across countries, of course, you’re also looking at two different gene pools.

So the survival statistics, while they might be useful for some kinds of comparisons, don’t really present any obvious conclusions when used to compare different populations. They can be interpreted to argue for leaving the U.S. system alone, or for extending coverage to the millions who don’t have it. Our advice: Use with caution.