Here’s how you solve
this kind of question using Bayes Theorem:
1.
Out
of 100 women, we expect 1.2 to actually have breast cancer (12 out of 1000)
2.
However,
only 88% (880 out of 1000) of these will be detected, so 1.056 (i.e. 88% of
1.2) will be detected
3.
In
addition, of the 98.8 women who don’t have breast cancer 9% (90 out of 1000)
will trigger a “false positive” (9% of 98.8 = 8.892)
4.
So
out of 100 women, 9.948 will have positive mammography readings (1.056+9.892)
5.
Of
these 9.948, only 1.056 will actually have breast cancer
6.
10.6% = (1.056/1.056+9.892)
ç This ratio is essentially an
application of Bayes Theorem
Please click on
http://www.thegodofreason.com/bayesintro.pdf
to download a PDF that explains how Bayes Theorem works. Page 3 gives an
example that is very similar to your problem: