Is the data wrong?

Sounds like fuzzy math to me!!!

I just e-mailed your question to my dh (he's using his laptop right now, in our basement/his man-cave that is two floors below my computer room - lol) - he's quite bright with stuff like this. I'll post when he responds to me!

Hi Missy - Here's your post:

The new guidelines are based on the fact that it takes 1900 mammograms to save 1 life.  BUT, 1 in 69 (1.4%) of women between the ages of 40 to 49 will be diagnosed with breast cancer.  Under current recommendations, each women has 10 mammograms (1 per year) over 10 years. The 1900 mammograms do not represent 1900 women.  Shouldn't 1900 be divided by 10 (the number of mammograms for each woman 40 to 49), making it 1 in 190 women who is saved NOT 1 in 1900 mammograms?

----------------------------------------------------------------  

Okay here's dh's response:

Info provided:  

1.   1 mammogram per year for 10 years per individual 

2.   1900 mammograms save 1 life

3.   1 in 69 women between 40 and 49 will be diagnosed with breast cancer  

Conclusion - there is no timeframe provided, however, I assume the 1900 mamograms are given over a very short time period so I would not agree dividing the 1900 by 10.   

For every 1900 mammograms provided (this could be in one day) 1 will show a positive result.

-----------------------------------------------------------

Just his opinion...please don't bash him! lol

I am not a statistician... not even close.  But, here are some things that might help clarify the "1 in 1900" issue.... or not.

The "1 in 1900" number comes from a literature review that was done on behalf of the USPSTF, to provide them with the information they could use in coming up with their recommendations. The literature review involved collection, careful reading, and mathematical analysis of results of clinical trials that tested BC screening methods.  The authors of the literature review focused on work done since the previous USPSTF report was issued in 2002, but they ended up using older studies too.  For those who aren't faint-of-heart, here's a link to the literature review:  http://www.ahrq.gov/clinic/uspstf09/breastcancer/brcanup.htm

Here's what that literature review says about the number of mammograms that would need to be done on women in their 40's to save one of those women from death due to BC: 

"For women aged 39 to 49 years, 8 trials provided data for the meta-analysis, including 6 from the 2002 meta-analysis... [I'm skipping the list of the studies, but they're on the website].  Combining results, the pooled RR [relative risk] for breast cancer mortality for women randomly assigned to mammography screening was 0.85 ... , which indicates a 15% reduction in breast cancer mortality in favor of screening.  This corresponds to a number needed to invite for screening to prevent 1 breast cancer death of 1,904 ... , over several screening rounds that varied by trial (2 to 9 rounds) and 11 to 20 years of follow-up.  ... Results are consistent with the 2002 meta-analysis (RR 0.85 ... ; 7 trials)."

Okay, so what does the "1 in 1,904" really mean, and why does it sound so strange? 

One thing to think about is that, just because "1 in 69 (1.4%) of women between the ages of 40 to 49 will be diagnosed with breast cancer", that doesn't mean 1 in 69 women in that age group will die of breast cancer.  The "1 in 69" is the frequency of diagnosis, right?  Assuming some, if not most, of those women dx'd with BC will be treated successfully, only a small fraction of the 1 in 69 will develop mets and die.  Since the USPSTF analysis and the literature review done on its behalf were focusing on mortality due to BC, and not just diagnosis, we can't compare the "1 in 69" to the "1 in 1,904" or use one of them to calculate the other.  We would need to know how many of the women diagnosed with BC between age 40 and 49 actually died of their diseease, because they're the ones included in the "1 in 1,904 mammograms" calculation.

Also, the "1 in 69" represents the frequency of BC diagnosis in that age range, but it would include all diagnostic methods -- not just mammography -- wouldn't it?  We know that mammography can be a problem in women of that age because their breast tissue tends to be more dense than in older women.  So, some of those tumors were probably detected with other screening methods, such as self-exams, clinical breast exams, ultrasound, etc.  Those diagnoses would be included in the "1 in 69" but not in the USPSTF's "1 in 1,904 mammograms" calculation.

And, there is this:  The "1 in 1,904" represents sort of an "absolute" benefit of mammography.  It's the number of mammograms that would need to be done to pick up one tumor that would otherwise go undetected and become lethal.  Earlier I quoted a paragraph from the literature review that said mammography in the 40-49 age group had a relative benefit of 15%.  That means women in that age range who were being screened with mammography were 15% less likely to die of breast cancer than women who were not being screened with mammography.  That sounds like a lot.  (I thought it was a lot.)  However, breast cancer really is not very common in that age group (1 diagnosis in 69 women), despite all the young women we know here on the discussion boards.  So, mammography doesn't provide all that much of an advantage, considering its "harms".  Whatever those are.

Finally, I don't think we can figure this out.  The authors of the literature review (the people who came up with the "1 in 1,904" number) calculated that number from data compiled from 8 studies.  Women in those studies did not necessarily have mammograms every year:  the data were included in the analysis if they had at least one mammogram in the previous 2 years.  The number of "rounds" of screening mammograms that were performed and included in the 8 studies over the 10-year age range varied from 2 rounds to 9 rounds.  And, since it takes awhile for BC to develop into mets, the follow-up period is important if we want to know how many women die of cancer.  The follow-up ranged from 11 to 20 years among the 8 studies.

Those are things that happen when someone does a "meta-analysis" and combines huge amounts of data from different studies to try to come up with statistically significant patterns.  I don't think there is any way we can back-calculate from the "1 woman in 69" that are diagnosed with BC, to the "1 in 1,904" mammograms that would be needed before one of them detected a tumor that would otherwise have been lethal. 

If you think all that's confusing, you ought to see the mathematical formulas the authors used in the literature review.  They're in the appendix of the article.  Seriously, please don't be mad at the people who wrote the literature review (the article at the link I provided).  This is the last paragraph of that review:

"Our meta-analysis of mammography screening trials indicates breast cancer mortality benefit for all age groups from 39 to 69 years, wtih insufficient data for older women. False-positive results are common in all age groups and lead to additional imaging and biopsies. Women aged 40 to 49 years experience the highest rate of additional imaging, whereas their biopsy rate is lower than that for older women. Mammography screening at any age is a tradeoff of a continuum of benefits and harms.  The ages at which this tradeoff becomes acceptable to individuals and society are not clearly resolved by the available evidence."

The USPSTF took that review paper and all its findings, and came up with the recommendations we're all riled up about.

I have a headache.  Can I take a nap now?

otter 

Otter - I am more confused than ever.  Not that your response wasn't clear - I just can't seem to digest this information.  But I do want to thank you for taking the time to put all of that together.   I think I am going to take that nap that you suggested!

Well, here's some more fuzzy math.  This is from an article in Time magazine that appeared in my Yahoo News in-box today:  http://www.time.com/time/health/article/0,8599,1941700,00.html

"Overall, based on a review of mammography trials, the [USPSTF] panel found that having a yearly mammogram screening cuts the risk of breast-cancer death by 15% in women ages 40 to 49. That reduction, it should be noted, is relative, not absolute. The absolute risk of breast-cancer death after age 40 is 3% without annual screening, according to the computer models. That means that with routine screening, which leads to a 15% lower risk of death from breast cancer, a woman's absolute risk drops to 2.6%. Small numbers in either case. Put another way, the panel concluded, the benefit of routine mammograms for women in their 40s is one fewer death for every 1,904 women screened annually for up to a decade."

Well, we know that the statement about "1,904 women screened annually for up to a decade" isn't correct, because that's not what the report (or the "Nelson" review article) said.  The frequency of screening was not necessarily annual, and the "cycles" ranged from 2 to 9 over the 10-year period.

But, other than that, the information in that paragraph might be useful.  Remember the "1 in 69" number -- the number of women who will be diagnosed with BC in their 40's?  From the Yahoo/Time article, we see that the risk of dying of BC after age 40 is 3% among women who are not regularly screened with mammograms.  I haven't been able to force myself to read the article from which the writer got that number.  It's a paper commissioned by the USPSTF that contains results of mathematical modeling (computer simulations).  The statistical models tested the effect of various hypothetical mammogram schedules on mortality from BC.  I can't read that stuff, so for now I'll just trust the "3%" number.  I honestly don't know if the number is from all women older than 40 (which would be irrelevant), or from women in their 40's.

So, if the mortality rate is 3% without regular mammograms, and mammograms provide a 15% relative benefit, that means the mortality rate in women who get regular mammograms would be (by calculation, not by trial results) 15% x 0.85 = 2.6%.  I guess the argument is that a 0.4% absolute decrease in the mortality rate from BC among women in their 40's is small.  It does seem a lot smaller than the 15% we have been talking about.  Do we know how many women in the U.S. are in their 40's?  It looks like we would lose 0.4% of those women to breast cancer if none of them had regular, screening mammograms.

Or is my math fuzzy, too?  [Please realize that I'm not passing judgment on whether that's a significant number.  I know some of those women, too. I'm just juggling numbers here.]

otter 

Yeah Laura! LOL..now they can start the study of "what happened to all those women in their 40's that we decided to ignore. How many died? How many ended up with stage III or IV BC because they were "statistically insignificant?"

peppi1 - You are so right! I was "statistically insignificant" when diagnosed at 45 yrs.old. From now on...everytime I introduce myself to someone, while shaking their hand, I am going to say, "Hi, I'm Laura, I'm statistically insignificant, what's your status?"

otter - et al - I will read that article. But even before I do, I have to comment about my concerns. All of us here, unfortunately due to our dx's, are very well informed about breast cancer. We are survivors, patients, warriors, victims, etc. And due to our dx, are very knowledgeable about everything pertaining to bc, a bc diagnosis and all that takes place prior to the actual diagnosis (suspicious mammo, discovery of an unusual lump, a breast indentation, breast redness, nipple invertion, etc.). We have done the research, we know the lingo, we have learned about all the treatments, medications, diagnostic equipment, prognosis, etc. We have read thousands of "stories" here about thousands of women dx'd w/breast cancer. 

My concern is that a huge percentage of the female population in our country know nothing about the above. They are like most of us were pre-diagnosis...like good soldiers. They listen to their Dr's and they hear and read the media reports. At 40 we started getting mammos and we heard and read about the importance of BSE's. Some of our Drs and nurses even took the time to show us how to do them. And now, women who don't know all that we do, are hearing and reading about the new guidelines. Because they don't know what we know, they will be believe and trust what they hear and read. And some, perhaps many, will simply shrug their shoulders and comply...like good soldiers.

For the past two years, I have been educating my nieces, ages 17 and 19 about BSE's. I explained to them that I found a lump when I was 16 YEARS OLD! I was fortunate...after a biopsy, it was found to be a B-9 cyst. No one told me to do BSE's in fact, at that time I had not ever heard of the words BSE. So for the following 25 years - I simply did not do them. And then years later, when the media broadcasted the importance of doing them, I complied. And four years later, I found a lump. Last week, my 19 year old niece called and told me she heard that BSE's wouldn't make a difference and she questioned whether she should even "bother doing them". It's not that she's unintelligent. It's not that she's lazy. It's more that she believes and trusts what she hears and reads. I asked her to listen to me...a woman who knows far more than she would like to know about bc and the consequences involved with ignorance about early detection.

As others have said...instead of moving ahead, women will be taking steps backward. No doubt this is causing confusion for millions of women of all ages. 

The message the latest guidelines is sending is archaic and lethal.

It's not 1 out of 190 women.  It's 1 of 1904 women (not mammograms) screened for an unspecified course ranging from 2 to 8 mammograms beginning at age 40, and followed for a length of time ranging from 11 to 20 years.  That's what it says in the paper.  I think it's crazy for them to specify "1 of 1904 women" while NOT specifying the number of mammograms and length of follow-up, but that's exactly what they said (see meta-analysis paper).