The argument against PSA screening rests on the assumption that people will make wrong decisions following a +ve PSA test, and ignores the value of having better health information, even if you don't change treatment decisions. I had a go at this topic a while back https://crookedtimber.org/2012/03/23/danglyparts-and-decision-theory/
Thank you for your thoughts and the link to your very well-written post. My view differs from yours, in that I see three factors undercutting the wisdom of mass screenings for low-prevalence problems, like PSA testing for early prostate cancer detection. Those factors are (1) rarity, (2) uncertainty, and (3) secondary screening harms. In the case of PSA and other medical screenings, with rarity we arguably need to look at the base of rate of death from prostate cancer, not cancer incidence. These deaths are very rare. With uncertainty we need to weigh how the anxiety of a possible positive test result might degrade quality of life, and with secondary screening harms how the costs of possible impotence, incontinence, and even death from unnecessary treatment could affect patients. See, e.g., Gerd Gigerenzer et al's excellent April 2009 *Scientific American* essay "Knowing Your Chances: What Health Stats Really Mean," https://legacy-pt.sciam.com/article/knowing-your-chances.
In your linked post, you say "I don’t know for sure how this translates, say, to mammograms." I treated that case more extensively in this recent post: https://wildetruth.substack.com/p/book-review-overdiagnosed. One of the factors that seems to distinguish mammography from PSA testing is that there are some questions regarding whether mammography itself may incur rare but serious risks, e.g., when compression damages tissue.
Thanks for thoughtful response. Australia has 3000+ prostate cancer deaths annually, all men and mostly old, so it's a much bigger risk for me than, for example, car crashes or (since I don't smoke) lung cancer.
The crucial issue is "how the anxiety of a possible positive test result might degrade quality of life". As I said in my post, I'd rather have more info than less, even if I didn't choose any treatment in response to a +ve test. That's a view that will differ from one person to the next, but not one on which medical professionals have any right to intervene in my choices.
I'll look at the Gigerenzer essay. I'm a big fan of this work.
What I think can be improved in many decisionmaking contexts -- at the policy and individual level -- is the information on which these choices are based. Women may rationally choose on the basis of the best available information to get mammography screening (e.g., because it may lower breast cancer deaths by approximately 1/1000). They may equally rationally choose to forego screening, to avoid the possible harms (e.g., anxiety from a positive result that is almost certainly a false positive, potentially disfiguring/risky follow-up screening and treatment that is also almost likely unnecessary).
Would you be able to please offer some citations for your understanding of the prostate cancer, car crash, and lung cancer death risks for Australians? I'm always interested in body counts, and surprised by this assertion.
Back on the main point, I think the central question here is what you care about. Prasad, Bewley, Welch, and others argue we should care about all-cause mortality benefits. The best available current PSA/mammography/colonoscopy evidence doesn't establish those from mass screenings.
But if you just want to lower your own odds of prostate cancer death a little bit, and accept the costs of big false positive chances, persistent uncertainty, and secondary screening/treatment harms if applicable, that's your choice to make and all the power to you.
This is a place where social institutional responses to cognitive bias and perverse incentives problems have to differ between medicine (where patients need to choose, given good information, for themselves) and security (where policymakers need to choose, given good information, for everyone).
The argument against PSA screening rests on the assumption that people will make wrong decisions following a +ve PSA test, and ignores the value of having better health information, even if you don't change treatment decisions. I had a go at this topic a while back https://crookedtimber.org/2012/03/23/danglyparts-and-decision-theory/
Thank you for your thoughts and the link to your very well-written post. My view differs from yours, in that I see three factors undercutting the wisdom of mass screenings for low-prevalence problems, like PSA testing for early prostate cancer detection. Those factors are (1) rarity, (2) uncertainty, and (3) secondary screening harms. In the case of PSA and other medical screenings, with rarity we arguably need to look at the base of rate of death from prostate cancer, not cancer incidence. These deaths are very rare. With uncertainty we need to weigh how the anxiety of a possible positive test result might degrade quality of life, and with secondary screening harms how the costs of possible impotence, incontinence, and even death from unnecessary treatment could affect patients. See, e.g., Gerd Gigerenzer et al's excellent April 2009 *Scientific American* essay "Knowing Your Chances: What Health Stats Really Mean," https://legacy-pt.sciam.com/article/knowing-your-chances.
In your linked post, you say "I don’t know for sure how this translates, say, to mammograms." I treated that case more extensively in this recent post: https://wildetruth.substack.com/p/book-review-overdiagnosed. One of the factors that seems to distinguish mammography from PSA testing is that there are some questions regarding whether mammography itself may incur rare but serious risks, e.g., when compression damages tissue.
Thanks for thoughtful response. Australia has 3000+ prostate cancer deaths annually, all men and mostly old, so it's a much bigger risk for me than, for example, car crashes or (since I don't smoke) lung cancer.
The crucial issue is "how the anxiety of a possible positive test result might degrade quality of life". As I said in my post, I'd rather have more info than less, even if I didn't choose any treatment in response to a +ve test. That's a view that will differ from one person to the next, but not one on which medical professionals have any right to intervene in my choices.
I'll look at the Gigerenzer essay. I'm a big fan of this work.
Absolutely, personal choice is paramount.
What I think can be improved in many decisionmaking contexts -- at the policy and individual level -- is the information on which these choices are based. Women may rationally choose on the basis of the best available information to get mammography screening (e.g., because it may lower breast cancer deaths by approximately 1/1000). They may equally rationally choose to forego screening, to avoid the possible harms (e.g., anxiety from a positive result that is almost certainly a false positive, potentially disfiguring/risky follow-up screening and treatment that is also almost likely unnecessary).
It was my understanding that the evidence of likely net harm from PSA testing was even stronger than that for mammography. See https://www.hardingcenter.de/en/transfer-and-impact/fact-boxes/early-detection-of-cancer/early-detection-of-prostate-cancer-with-psa-testing. (Gerd Gigerenzer is the Director of the Harding Center.)
Would you be able to please offer some citations for your understanding of the prostate cancer, car crash, and lung cancer death risks for Australians? I'm always interested in body counts, and surprised by this assertion.
Here's prostate cancer 3000+
https://www.canceraustralia.gov.au/cancer-types/prostate-cancer/statistics
Road deaths
1000+ of who about 150 were males 65+
https://app.powerbi.com/view?r=eyJrIjoiZjU4YjhiMGYtNTljYS00ZjEyLTg5NTItN2YyOWExNDcwNTdmIiwidCI6ImFhMjFiNjQwLWJhYzItNDU2ZC04NTA1LWYyY2MwN2Y1MTc4NCJ9&pageName=ReportSection0872ca58aca00519e8e1
Lung cancer 8000 male deaths, but only 15 per cent of male cases are non-smokers (and most of those occupational exposure, I think)
https://www.canceraustralia.gov.au/publications-and-resources/cancer-australia-publications/risk-factors-lung-cancer-overview-evidence
https://www.abc.net.au/news/health/2021-06-21/lung-cancer-in-nsw-never-smokers-under-the-microscope/100129264
Thanks for these links. Somewhat tangentially, air pollution's possible role in lung cancer risk was recently in the news: https://www.nature.com/articles/d41586-023-00929-x.
Back on the main point, I think the central question here is what you care about. Prasad, Bewley, Welch, and others argue we should care about all-cause mortality benefits. The best available current PSA/mammography/colonoscopy evidence doesn't establish those from mass screenings.
But if you just want to lower your own odds of prostate cancer death a little bit, and accept the costs of big false positive chances, persistent uncertainty, and secondary screening/treatment harms if applicable, that's your choice to make and all the power to you.
This is a place where social institutional responses to cognitive bias and perverse incentives problems have to differ between medicine (where patients need to choose, given good information, for themselves) and security (where policymakers need to choose, given good information, for everyone).