If you are saying that there is no overdiagnosis in mammography only in PSA, feel free to read PDQ summary of evidence on the NIH website as well as USPSTF recommendations and rationale (not just the summary). Feel also free to read this estimate of overdiagnosis from Malmo trial (click on PDF to read the complete article; don’t forget rapid responses as they show an obvious flaw in authors’ math resulting in underestimate). There are other estimates, but this one is based on real data and not just modeling. Given the improvements in technology since the time of Malmo, the numbers are enough to affect the total ratio.

Obviously, if you do more screening (or if the test is more accurate), you’ll detect more cancer. Some of these extra cases will be overdiagnosis. In addition, there are borderline cases which a US pathologist may be more likely to call cancer because of the liability fears. Some of these may also be overdiagnosis.

Certainly comparing survival at the same stage would show if the treatment is better in the US. But the population study would also show the effect of additional screening in the US. Evan is right in that this data must be published somewhere. I don’t have time to look, but if anybody has it, it’d be interesting…

Evan, when we talk about overdiagnosis increasing the ratio of cured to detected, we are increasing both nominator and denominator by the same number (all overdiagnosed cases are cured). So we are comparing (x+a)/(y+a) to x/y where y>x, x,y,a>0. Trivial to prove that former is greater (ay – ax >0 since y>x)

but the cancer stuff has been thoroughly debunked by Gerald Anderson, Jon Cohn et al. It’s intellectaully vacuous to keep repeating it.

Meanwhile the variation in practice and the use of drugs in cancer care in this country is a disgrace for entirely different reasons that a tax-cutting hawk like Stossel would be appalled about if he spent five seconds trying to understand it!

Are you saying there are more false positives in the USA? Are you saying that these false positives are being called cancer?

The patient is not being recorded as having cancer unless there’s a tissue diagnosis of cancer. An abnormal mammogram is not entered into the cancer registry as cancer until a biopsy establishes that it is cancer. I’ve had various imaging studies come back over the years as highly suspicious for cancer (breast, lung, etc.) that came back with a non-cancer tissue diagnosis on biopsy. I do not get calls from the cancer registry about them.

If false positives are a random event, one would think that other countries get the same percentage of false positives we do.

Are you saying that “overdiagnosis” means true cancers that are detected at so early a stage that the patient would likely die of other causes? Like perhaps finding an early prostate cancer in a 95-year-old man?

If you mean by “overdiagnosis” that true positives are being found earlier, making them easier to treat than another country finding the same cancers later, I’m having a hard time understanding why that’s not relevant. If a socialized system can do anything right, one would think it would be able to do primary care and screening right.

And no, I don’t think you need to look at population data necessarily, you look to compare populations with cancer at the same stage. Because I agree, it is not fair to compare survival of one population with advanced cancer against a population with early cancer.

It’s obvious that the fastest way to drop a percentage is to increase the denominator, so you are correct that population-adjusted data is the way to go. I have no idea where this is stored but I’m willing to be it exists somewhere.