Shannon Brownlee and Jeanne Lenzer teamed to deliver one of the best pieces on cancer screening that I’ve seen.

“Can Cancer Ever Be Ignored?” appears online today but will be in print in next Sunday’s New York Times magazine.

The piece includes great perspectives from:

• Dr. Otis Brawley of the American Cancer Society (“I’m not against prostate cancer screening. I’m against lying to men. I’m against exaggerating the evidence to get men to get screened.”)

• Dr. Barry Kramer of the National Institutes of Health/National Cancer Institute (“Men may be trading one cause of death for another.”)

• Dr. Richard Ablin, who discovered a prostate-specific antigen and now calls the PSA test “a public health disaster”

• Darryl Mitteldorf, executive direction of Malecare, a cancer patient support group, who says it is not uncommon for men to regret their decision to be tested and treated for prostate cancer.

• Dr. David Newman of Mount Sinai School of Medicine in NY, who says researchers must look not just at the number of deaths from prostate cancer but also at the number of deaths caused by treatment.

• Dartmouth’s Dr. Gil Welch who says “The European trial says 50 men have to be treated for a cancer that was never going to bother them to reduce one death. Fifty men. That’s huge. To me, prostate screening feels like an incredibly bad deal.”

The piece also confirms that politics did enter into the decision to cancel a 2010 meeting of the US Preventive Services Task Force at which the group was to consider a proposal to give routine PSA testing a “D” rating – “D” as in don’t do it – for any man of any age. Apparently the heat from the USPSTF’s recommendation on mammography the previous year was still too great.

Former USPSTF staffer Dr. Kenny Lin writes about that clash between science and politics on his blog. Excerpt:

“I hope that the New York Times story sheds some much-needed light on the shadowy politics surrounding prostate cancer screening, and in so doing, allows the current Task Force to re-assert its recently curtailed independence and unfettered ability to make science-based recommendations for the good of the public, rather than the agenda of any politician or political party.”

Last month was prostate cancer awareness month. I only wish the New York Times had published this in the midst of all that hoopla.

You might also like


Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Katen moore

October 5, 2011 at 11:07 am

Great review look forward to the magazine–where’s the breast ca crit?

Tim Bartik

October 14, 2011 at 12:07 pm

Mr. Schwitzer:
I think that the Brownlee piece, and some of those quoted in it, as well as the Task Force report, misinterpret some of the evidence.
First, there is no good evidence that prostate cancer treatment causes deaths from non-prostate cancer related causes that offset the reduced risk of death from prostate cancer. The existing studies and really any feasible study simply have inadequate sample size to accurately measure the prostate cancer effect on overall mortality.
If overall mortality is 10% or 20%, and the screening effect after 9 years is to reduce prostate cancer mortality by 0.07% (the European study estimate), there is no way that a sample size of 80,000 or so each in the screening group and control group will be able to detect such an effect. And in fact, if you look at the confidence intervals for the overall mortality effects in the European study, they clearly are much too large to tell anything at all about effects on overall mortality.
You would need a study with a sample size of over 1.5 million in the screening group, and over 1.5 million in the control group. to detect effects on overall mortality.
Detecting effects on prostate cancer specific mortality is easier, because it is a rarer event.
So, Dr. Newman’s statement that prostate cancer treatment kills as many as it saves is not based on any empirical evidence that this is the case. And it seems unlikely. The European study already included deaths from prostate cancer surgery itself in its definition of prostate cancer related mortality.
Second, the 50 to 1 figure cited by Dr. Welch presumably comes from the European study’s estimate of a NNT of 48. But this is after an average of 9 years of followup. The survival probabilities only diverge at 7 years, and then widen over time, as one would expect from the biology of the disease.
More recent research suggests that the NNT at 12 years, the max you can go with the European data, is 18.
Now, you can look at this and decry the 17 men treated unnecessarily. But this treatment is “unnecessary” only in some possible future world of perfect knowledge.
In this world, we need to think probabilistically about it. The 18 figure for NNT means that diagnoses via PSA screening, and possible subsequent treatments, reduces the risk of dying from prostate cancer by 12 years by a little over 5%. (Note that this is 12 years after SCREENING has begun, and is 5% for everyone DIAGNOSED even if they do not undergo surgery. So the true risk reduction will be even greater for actual treatment and will be realized at somewhat less than 12 years after TREATMENT.) The tradeoff for this is what the Task Force estimates are major side-effects at a probability of 20 to 30%.
So, on the one hand, the chances of major side effects exceed the risk reduction in death. On the other hand, many men, and perhaps most men, would view death as a worse outcome than these side-effects.
I don’t think that the screening decision is at all clear-cut. It involves a complex weighting of probabilities, and difficult tradeoffs. The Task Force and some of the news coverage has been tempted to try to minimize the possible risk reduction due to prostate cancer treatment, and treat it as small. It is NOT small, at least not for many men, relative to the major cost of screening, which are the side-effects of treatment.
By way of disclosure, I am not a urologist. I am an economist who has recently had to face all these decisions about screening and treatment. I decided to use my econometrics training to see what the medical statistics said. I think some of the interpretations that are being widely used by many participants in this debate are mistaken.