4 ms·
Gerd Gigerenzer talks about medical testing, and the scary numbers of health professionals who are unable to interpret the results. https://plus.maths.org/cont
by DanBC 11y ago
Gerd Gigerenzer talks about medical testing, and the scary numbers of health professionals who are unable to interpret the results.
https://plus.maths.org/content/reckoning-risk https://plus.maths.org/content/reckoning-risk
This is what most people struggle with:
> The probability that one of these women [asymptomatic, aged 40 to 50, from a particular region, participating in mammography screening] has breast cancer is 0.8 percent. If a woman has breast cancer, the probability is 90 percent that she will have a positive mammogram. If a woman does not have breast cancer, the probability is 7 percent that she will still have a positive mammogram. Imagine a woman who has a positive mammogram. What is the probability that she actually has breast cancer?
That can be reworded to something much easier:
> Eight out of every 1,000 women have breast cancer. Of these 8 women with breast cancer, 7 will have a positive mammogram. Of the remaining 992 women who don't have breast cancer, some 70 will still have a positive mammogram. Imagine a sample of women who have positive mammograms in screening. How many of these women actually have breast cancer?
It's a good book and covers cases of misdiagnosed HIV.
http://www.amazon.co.uk/gp/aw/d/0140297863 http://www.amazon.co.uk/gp/aw/d/0140297863
http://www.amazon.com/gp/aw/d/0140297863 http://www.amazon.com/gp/aw/d/0140297863