Skip to main content

Why We Overrate the Lifesaving Power of Cancer Tests

Science journalist Gary Taubes weighs in on PSA tests and mammograms

Many men think positive PSA tests saved them from prostate cancer (visible in top right corner of this micrograph), but odds are greater are that they received surgery or radiation that they did not need.

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


My previous post presented data showing that the PSA test for prostate cancer harms more men than it helps. According to a recent analysis by the U.S. Preventive Services Task Force, you are 240-120 times more likely to be misdiagnosed as a result of a positive PSA test and 80-40 times more likely to get unnecessary surgery or radiation than you are to have your life saved.

I mentioned that my physician, trying to persuade me to get the PSA test, said he knew someone whose life it had saved. Many readers also insist that the PSA test saved them or someone they know, and they dismiss the massive evidence of overdiagnosis and overtreatment as irrelevant. See for example comments on Scientific American’s Facebook page. I have gotten similar responses to my posts on mammograms.

Celebrities, aided by mass media, have reinforced the public’s belief in the life-saving power of cancer tests. Actor Ben Stiller said last year on the Howard Stern Show that the PSA test “saved my life,” and ABC reporter Amy Robach asserted in 2013 that a mammogram saved her.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


I understand why many people adamantly reject data exposing the flaws of PSA tests and mammograms. No one wants to think they've endured surgery, radiation or chemotherapy unnecessarily. But for our broken health care system to be fixed, consumers must take a more rational approach to health.

I received an email on this issue from science journalist Gary Taubes, renowned for his writings on diet, including the recent bestseller The Case Against Sugar, and other medical topics. I routinely assign my science writing students his 2007 article on the limits of epidemiology. Here are Taubes’s comments, in italics:

A friend just sent me your post on the PSA test, which I thought was great. The only thing that kept nagging at me was your doctor’s statement that “he knew someone whose life had just been saved by the test.” My question is whether such a thing is possible. I don’t think there’s any way he can know that. He can know someone who had a PSA test that was positive, discovered he had prostate cancer, got treatment and is still alive, but he’ll never know if that guy might have been alive today and thriving even if he hadn’t had the PSA test and hadn’t got treated (yet). Maybe the technology is such now that they genotyped this guy’s cancer and realized it was fast-growing and virulent, and maybe they know he has a family history of early death from prostate cancer, and so your doctor can conclude with high probability that they really did catch it just in the nick of time — thank God for the PSA. But without that kind of evidence and revelation, as you note, this might have been a slow-growing thing that he could die with thirty years from now when his heart finally goes kaput in his assisted-living facility. So I wish you had asked your MD, “How do you know the test saved his life?”

I thought about this considerably when I was reporting and writing on the very similar mammogram story 20-odd years ago. All the cognitive biases — the “what-you-see-is-all-there-is” effects as Daniel Kahneman would call it, or WYSIATI — work to make you think the tests are valuable. A woman gets a mammogram, learns she has a tumor, has treatment and lives. Both she and her MD naturally assume that the mammogram saved her life, although, who knows. Maybe the cancer never would have killed her. Maybe she would have been fine waiting five or ten years before she felt a lump. If the woman gets a mammogram, learns she has a tumor, has treatment, and dies anyway, both she and her MD assume that had they only gotten the mammogram earlier, she would have lived. Although, who knows, maybe her cancer was such, as oncologists now believe, that no matter when the mammogram detected it, it would have been too late. 

The scenario, no-mammogram->tumor->death, leads to the assumption that she should have gotten a mammogram and so been saved. The scenario no-mammogram->tumor->life leads to no conclusions one way or the other. Perhaps only in the rare instances of false negatives — negative-mammogram->actual tumor->death — will either the woman or the MD come to the conclusion that somehow the mammogram was a mistake, although even that takes a leap of the imagination that we humans are unlikely to make. After all, had she not had the mammogram, she probably (critical caveat) would have remained unaware of the tumor and it still would killed her. One of my uncles had what appears to have been a false negative on a colonoscopy. When he did finally feel symptoms of illness, he ignored them for months assuming it was nothing serious. The colon cancer then killed him. But it might have killed hm anyway and none of us — even me — assumed the colonoscopy was somehow a mistake. And we don’t know for sure he had the cancer when he had the colonoscopy.  

So once you create the screening technology and make it available, these WYSIATI assumptions will then work to convince everyone closely involved with its use, both patients and doctors, that it is beneficial, even if it’s ultimately doing far more harm than good. 

 Or so I think. Am I right? The question is always, Is there a way to determine that such a screening test — PSA, mammogram, colonoscopy, whatever — actually saved a life in an anecdotal situation? Or is it always an assumption based on the chain of events — screening-> tumor -> successful treatment, and no-screening -> tumor -> death and the variations? Are clinical trials always necessary, and even then the answer will always be a probability?

 My conflict of interest statement: I haven’t had a PSA test of any kind other then the digital rectal method in check-ups in the distant past. My wife is still in her forties and had her first mammogram a few years ago. As I remember it (always an important caveat), it was “irregular”, or some such terminology, thus prompting her physician to ask her to return for a follow-up, which was not irregular. Now her physician suggested she should return for another follow-up in six months to assure no irregularities return, as her physician now seemed to think (all too naturally if completely due to cognitive bias) that my wife is somehow at higher risk. And the physician’s belief that this could be the case could easily infect my wife, making her also think she's at higher risk, thus bestowing a lifetime anxiety. I like to think I function an immune system against such unfounded worries, but I’m certainly not an effective one. And so it goes.  When the first irregular mammogram came in, I asked some of my female friends whether they had false positives with mammograms, thus prompting anxieties and follow-ups. Five of five said yes and told me their stories.

I recognize that, should my life be ended prematurely by any cancer, it will be assumed that my attitude toward these tests was assuredly responsible. In other words, if only I had gotten _______ test, I’d still be alive. And it could be true, but there’s no way to tell. (If the killer is a heart attack, it will be blamed on my diet.) 

Further Reading:

Why I Won't Get a PSA Test for Prostate Cancer

Meta-Post: John Horgan Posts on Cancer