Allow me to put some numbers to this:

The PSA test with a cutoff of 4 ng/mL has a sensitivity of just .205, meaning that 20.5% of people who have prostrate cancer will have a positive result. Moreover, it has a specificity of .936, meaning that 93.6% of people who do not have prostate cancer receive a negative result. “But wait...” those of you who can subtract might be saying, “doesn’t that mean that 6.4% of people who take the test will receive a positive result even though they

*don’t have cancer*?” Yep, it does.

But that doesn’t take into account prior probabilities at all. So what does that mean for me if I receive a positive result?

Well, as near as I can find, the prevalence of prostate cancer in those ages 40-44 is 0.000037, or .0037%. The likelihood that I have prostate cancer right now without any further testing are about .0037%. I think that everyone would agree that performing a biopsy based upon this knowledge alone would not be medically justified. That would mean everyone would be given an annual biopsy at huge risk and expense. But if we could add in a positive PSA result, what would be my likelihood of having cancer? Surely it would go up?

Well, yes, of course it does. But the question is

*by how much*? Applying a little Bayesian statistics to the question, we can figure this out. I’ll spare you the math (though I’ll provide it if you want), but the upshot is that my likelihood of having cancer if I get a positive PSA is 0.000118506 or about 0.01%. Yes, I’m three times as likely to have cancer as I was before the test, but the odds that I actually have cancer are still minuscule! Moreover, the likelihood that My result is a false positive are 99.98%!

So, I ask you, do you think such a result really changes the calculus as to whether I should pursue further testing?

And this doesn’t even consider the likelihood that any cancer detected would actually ever become life-threatening (which is low). Nor does it take into account the risks of biopsy or the reduction in quality of life that could come with such treatment. Or the stress such a result would create even if I didn’t pursue further diagnostics.

Somewhere in my 40s m doctor asked me to do one as a baseline for future tests, rather than as a lone diagnostic test. That seemed like a reasonable thing to do.

I had not considered that, and I don’t know anything about the value of such a baseline. My doctor didn’t mention that.