Breast cancer detection bill passes Legislature
Original post made on Sep 13, 2011
Read the full story here Web Link posted Tuesday, September 13, 2011, 1:50 PM
on Sep 13, 2011 at 2:47 pm
OK, I admit that this is a step in the right direction.
But, it is a tiny, pidling shuffling step!
Insurance only covers mamograms, despite the fact that (according to this article) are only useful for about half the women. Despite the fact that medical science is continuing to develop MUCH BETTER screening methods, these are rarely covered by insurance. So, while this bill is a good thing, I'm afraid that in my case it is just going to add to the fear & uncertainty in my life by not knowing if the mamogram is accurate or not.
on Sep 13, 2011 at 4:36 pm
The headline struck me as odd; why not "mammogram disclosure bill"? The relevant numbers in the story tell us that cancer detection via mammogram is unlikely for over half of women, those with dense breast tissue (DBT). Their readings should be labelled as "inconclusive" given that the false-negative rate is an overwhelming 75%. Clearly all the statistics are different for DBT women. A little algebra teases them apart from the fortunate few (or less than half) non-DBT women. Approximating the incidence of DBT as 1/2, the 1 in 8 overall BC rate would break down to only 1 in 24 non-DBT, along with 5 in 24 (over 1 in 5) DBT women, having cancer at some time. I'd like to apply Bayes' formula to the subgroup of women with dense breast tissue (despite the fact they probably shouldn't trust their mammograms) to figure out their actual probability of cancer given a "positive" mammography verdict: With DBT, prob of a positive mammogram given (previously undetected) cancer is (only) M|C = 1/4 and Bayes tells us C|M = (M|C) C / M = (1/4) * (5/24) / M where M is prob of a positive mammogram, which is unfortunately not given (for either subgroup) but is no more than 1, so the probability of cancer given a positive, dense-tissue mammogram is at least 5/96 or 1 in 19. We already knew their "prior" probability was 5 in 24. So if anything, a mammogram, for most women (those with DBT), would create a false sense of assurance-- hence the need for a clear, specific disclosure! What's most clearly needed is a test for early-stage breast cancer that produces meaningful results for the vast majority of women. Maybe a biochemical assay.
on Sep 13, 2011 at 5:32 pm
...and/or ultrasound...? By the way, the cancer incidence rates (5/24 with DBT and 1/24 without) I figured were lifetime totals, so my result for C|M (probability of cancer given a positive mammogram, ever) for those with DBT would also be cumulative. Depending on how the clinician calibrates her "positive" rate M (which should be higher for DBT), then C|M, the probability of a patient with DBT ever having breast cancer, given that she ever has a positive mammogram, could be between 25% (the "ideal" case, for M = 5/24)and 42% (the "naive" case, for M normalized to equal the total cancer rate of 1/8). My point is, it's really hard for a patient with DBT to figure her actual odds.
At least, it's hard for me. Which could help explain my unemployment.