Mammograms have an almost magical reputation among many women. At pink ribbon events, breast cancer survivors credit the screening tool as having saved their lives, and the U.S. government believes in their power so strongly that coverage for it is mandated by the Affordable Care Act for newer insurance plans.

Much of this faith in mammograms comes from a pivotal paper based on tests conducted in the 1960s and '70s in Sweden that showed that screening could reduce breast cancer mortality by an amazing 20 percent to 25 percent.

Now a new review of that study, published in the Royal Society of Medicine on Tuesday, argues that the method of statistical analysis was flawed and that the reduction in deaths is actually closer to 10 percent.

The findings support a growing body of evidence in North America, Europe and Australia that show that rates of advanced breast cancer have not declined in countries where most women regularly get screened, said Philippe Autier, the study's lead author and a researcher at the University Strathclyde Institute of Global Public Health.

A 2011 paper in Archives of Internal Medicine titled "Likelihood That a Woman With Screen-Detected Breast Cancer Has Had Her 'Life Saved' by That Screening," for instance, concluded that "most women with screen-detected breast cancer have not had their life saved by screening."

In fact, the authors wrote, "they are instead either diagnosed early (with no effect on their mortality) or overdiagnosed."

Earlier this week, a study in JAMA Internal Medicine that looked at data from 16 million women in 547 U.S. counties in 2000 suggested that there was "widespread overdiagnosis" from mammograms.

The debate over the utility of mammograms is far from settled, however. The authors of the JAMA study themselves point out that their study is "inherently vulnerable to ecological biases," meaning that it's impossible to tell whether the people who received the intervention -- in this case breast cancer screening -- were the same ones who developed the disease or whether there are other explanations for the link.

This post has been updated.

Read more: