What if a non-partisan, authoritative entity wrote a robust, evidence-based guideline, but nobody followed it?
That is precisely what’s happening with the USPSTF’s recent revision of their breast cancer screening recommendations. The change most find problematic is their recommendation that women younger than 50 not undergo any breast cancer screening, such as with a mammogram.
Here are their reasons explaining why:
The harms resulting from screening for breast cancer include psychological harms, unnecessary imaging tests and biopsies in women without cancer, and inconvenience due to false-positive screening results. Furthermore, one must also consider the harms associated with treatment of cancer that would not become clinically apparent during a woman’s lifetime (overdiagnosis), as well as the harms of unnecessary earlier treatment of breast cancer that would have become clinically apparent but would not have shortened a woman’s life. Radiation exposure (from radiologic tests), although a minor concern, is also a consideration.
Adequate evidence suggests that the overall harms associated with mammography are moderate for every age group considered, although the main components of the harms shift over time. Although false-positive test results, overdiagnosis, and unnecessary earlier treatment are problems for all age groups, false-positive results are more common for women aged 40 to 49 years, whereas overdiagnosis is a greater concern for women in the older age groups.
They’re completely right, of course. As I previously wrote in a USA Today op-ed on this issue, “For every inspiring story of a person cured from cancer made possible by early detection, there are untold stories of many more who suffer from the side effects of unnecessary invasive procedures stemming from false positive test results.”
But some patients are rebelling against these new guidelines, as noted in this article from The New York Times, New Mammogram Advice Finds a Skeptical Audience. A few doctors, too, are refuting the new recommendations as well. Suspiciously, they are mostly radiologists and oncologists. Cynics may wonder whether there is a financially-motivated basis for their dissent.
What’s fascinating is how mammogram screening has now turned rabidly political, with conservatives making the ridiculous link to “rationing.” And Kathleen Sebelius, the secretary of health and human services, tried to distance herself from the USPSTF, stressing that, “I would be very surprised if any private insurance company changed its mammography coverage decisions as a result of this action,” and that, “our policies remain unchanged.”
What, then, is the point of making any guidelines at all, if our government urges everyone to ignore them?
Progressive reformers, who generally espouse comparative effectiveness data and evidence-based medical practice as a means to control costs, should be very worried about the backlash these guidelines are eliciting.
If recommendations from an entity like the USPSTF – as non-partisan and robust as it gets – gets so much resistance from doctors, patients, and even the government itself, findings from a comparative effectiveness body stand absolutely no chance of changing medical practice.
Related posts:
- Breast cancer screening: Orac’s take
- Will patients accept the new, evidence-based, breast cancer screening guidelines?
- An oncologist takes on the recent breast and cervical cancer screening controversy
- Radiologists don’t agree with the USPSTF breast cancer mammogram screening guidelines
- "The great majority of women in the United States should not be getting MRI scans for breast cancer screening"
- The National Consortium of Breast Centers (NCBC) response to the USPSTF breast cancer screening recommendations
- 15 cancer screening posts you may have missed
{ 72 comments }
← Previous Comments
@Tristan – regarding presence of oncologist, etc. on the panel.
Oncologists see individual patients but they don’t see the overall picture that comes from studies of thousands of people, nor are they particularly qualified to analyze the data. They see a lot of anecdotal data, but like Kevin and others have mentioned many times in the past “plural of anecdotes isn’t data”. This evaluation of data has ALWAYS been the realm of epidemiology. This applies to everything – who do you think evaluates the effectiveness of new drugs? Studies provide information about how many additional diagnoses of cancer were made in screened vs non-screened women; whether or not this increase was accompanied by a corresponding decrease in the number of advanced cases diagnosed later (hint: it was not), differences in breast cancer mortality in screened vs non-screened groups. You need to know statistics to analyze this data a lot more than medicine.
On the other hand, oncologists see actual cases. So they see early diagnosis and say “oh, it’s easier to cure early cancer”. But is it easier to cure because the cancer was slow growing to begin with? Maybe a particular cancer would’ve never spread. Lab cannot give answers to these questions. Additionally, one extremely misleading metric oncologists may use is 5-year survival after diagnosis. This number is misleading because of what is known as “lead time bias”. Think about it this way – say someone dies of cancer at the age of 50. The cancer may be diagnosed later – e.g. at the age of 48 or earlier – e.g. 42. The result is the same, but in one case the person survives longer after diagnosis. Only long term data from studies that compare mortality means anything, not survival after diagnosis, and only epidemiologists who have knowledge of medical statistics are qualified to evaluate this data.
@SaraW — overdiagnosis is not an “assumption”. It is a well-known and well-proven, and it has been known for quite a while. People have provided links for you in the past — I remember posting the link to one of the study of overdiagnosis years ago in a similar discussion here. What is new now is more information about the magnitude of this phenomenon. Dave’s post (#23), item 1 contains a link to USPSTF rationale which in itself contains references to study.
It’s not like this information is difficult to find. I found it years ago and I am a lay person too. Google for “overdiagnosis in mammography” and read papers in medical journals. A lot of them can be read in their entirety.
Lisa Lindell:
>>Medicial or scientific evidence be damned, give the people what they want.>>
Exactly!
No matter how many tests one performs on whom, it’s impossible to reduce risk to zero.
“Diagnosis” does not equal “cause of death”.
Consider: if I develop a slow-growing breast tumor at age 47 then die of a heart attack at age 49, the breast cancer (even if undiagnosed and untreated) will be an incidental finding on the autopsy. Breast cancer won’t be my cause of death but I’ll count as “1 in 9″.
Further consider: lung cancer is less common than breast cancer, but far more deadly; many more women survive 5 years with a diagnosis of breast cancer (around 90%) than lung cancer (around 15%). Sure, lung cancer is more common in smokers, but lung cancer isn’t unknown in young non-smokers (remember Dana Reeve, Christopher Reeve’s wife? terribly sad story – she was only 43 at diagnosis and a lifelong non-smoker). Still, physicians don’t routinely order screening chest radiographs. Pancreatic cancer has 5-year survival of only about 5% and we don’t screen for that, either.
To further complicate matters, not all breast cancer is equally deadly – DCIS does NOT have the same prognosis as IBC!
It’s critical to look the raw data (American Cancer Society, CDC, etc.) because it is difficult to determine who is saying what in the current babel of indignation. I work with statistics in the medical field and am infuriated by the rampant twisting of facts in the news media by “professionals” with various agendas. I find it especially repellent that physicians with a financial interest in maintaining current screening levels are so blatantly stoking the fears of women who are terrified of cancer.
I was being sarcastic. Nobody said anything about reducing risk to zero. It’s risk vs benefit. If the risk outweighs the benefit for the general population, then the general population should not contribute to the expense or carry the burden. Many specialists in many fields would foam at the mouth at my absurd suggestions, and find plenty of “data” for pre-screening everything for everybody.
BD:
“I find it especially repellent that physicians with a financial interest in maintaining current screening levels are so blatantly stoking the fears of women who are terrified of cancer.”
Best post yet on the subject (and I would add to ‘physicians,’ numerous non-profit organizations)!
Worst media coverage to date:
Eleanor Clift (who is usually smarter than this) on NPR, arguing in favor of continuing screening for women in their forties:
“I personally knew two women who had breast cancer diagnosed by mammograms in the 40s, and they died!”
Tragic for those women and their families, but what in the world is the logic behind her argument?
Lisa Lindell:
>>I was being sarcastic.
Yeah, I know. Liked your Modest Screening Test Proposal.
>>Many specialists in many fields would foam at the mouth at my absurd suggestions, and find plenty of “data” for pre-screening everything for everybody.>>
I don’t doubt it. Scary, no?
R Watkins:
>>“I personally knew two women who had breast cancer diagnosed by mammograms in the 40s, and they died!”
Tragic for those women and their families, but what in the world is the logic behind her argument?>>
She didn’t prove my point, but she definitely cast doubt on hers!
I’ve spent over 30 years in medicine and I’ve never seen such sham science before. I’m sure the pediatric providers (all 4 of them – 25% of this “robust” expert panel) have a wealth of real-world experience in dealing with the highly evolving area of adult oncology. There wasn’t a single, real expert in breast cancer on this panel.
Kevin, like nearly everyone else, is focusing on the lower bound – the controversial recommendation not to screen women from 40-49 (or 39 to 49 if you go back to the original study). The upper bound is more interesting. The recommendation that women older than age 74 are too old and not worth screening/saving. Economically this makes sense. However, ethically, we have moved into different territory that is unusual in this country – a government-sponsored recommendation not to screen for fatal disease. When it comes to oncology, I’m concerned that “First, do no harm” is quickly becoming “Do not intervene.” We’ve seen it with other cancers, now it comes to breast cancer.
I absolutely support evidence-based science – even if that science calls for scaling back on (or totally eliminating) screening. Let’s not embarass the profession with garbage research and claim it’s wonderful evidence-based medicine. Read it for yourself. This wasn’t it.
This was politics posing as medicine. I’m embarassed for the USPSTF. And I’m shocked at HHS’ funding of such a, frankly, crap study.
@R Watikins–that may be the worst coverage of the story. I think this is one of the fairest:
http://www.msnbc.msn.com/id/34040273/ns/health-cancer/from/ET
Pharm Aid:
“The recommendation that women older than age 74 are too old and not worth screening/saving.”
That is not what the USTFPS said. Your statement is a total falsehood, and discredits everything in your post.
“The upper bound is more interesting. The recommendation that women older than age 74 are too old and not worth screening/saving.”
No, absolutely, that is not what it says. It says mamms haven’t been shown to be helpful in that population. To say we don’t want to save seniors from breast cancer is inflammatory and without any basis in fact.
There seems to be this religious faith in mamms, that they are some amazingly useful tool. They are not. The data has never said they are.
If a tumor is slow growing, as are most breast cancers in that population, finding it on exam is more helpful than finding it on a mammogram. It is just as curable and the risk over overdiagnosis and overtreatment is much lower.
There’s a wide range of functional capacity in seniors. A woman’s doctor can make the best judgment as to likely effectiveness of a screening tool. That’s why the rec is to discuss it with your doctor and not just tell everybody to do something which won’t help the vast majority.
DIora, I think it’s important to distinguish what you mean by “overdiagnosis”. The panel’s assumption that I speak of, is an assumption that cancer will not advance in a significant portion of cases, that cancer is there, perhaps, at age 40, but it is not and will not become clinically significant.
This assumption does not appear to be quantified or well supported according to many critics of the panel’s reccomendation.
I think you mean something else by overdiagnosis – am I correct? You are referring to that percentage of women who have a suspicious mammography, in whom cancer is not detected in follow-up tests.
For me the question is the one previously asked in BC screening controversies: whether there is any mortality benefit for women between ages 40-49 as a result of screening. The goal-posts have been moved. The question for this panel is no longer “is there any mortality benefit at all”, but whether that benefit is worth the “downsides” of screening.
They make a subjective assessment that the pain and discomfort of biopsies and the anxiety a woman might feel over suspicion of cancer, and this isn’t their call.
What this panel is about is to set guidelines for PAYMENT of services. What should the system pay for. There is another underlying assumption – that overtesting is costing the big bucks, money that could be used to fund other things.
I think one way to make this issue rest in its proper place is to change “the system” so that screening and preventive care is a self-pay matter: to be paid for from private or employer funded medical savings accounts that are tax-sheltered.
No, DIsgrace, Diora is referring to the very real issue of diagnosing cancers which will never be clinically significant. The other problem is called “false positives” and it’s an entirely different issue. This is a well documented and supported phenomenon despite what you may have read. That’s not to say it isn’t controversial – it most definitely is – but to completely dismiss it as irrelevant is to ignore a decent volume of evidence. She referred to links that may point you to the research documenting this phenomenon so you can decide for yourself.
And the goalposts for any screening program is, and have always been, mortality benefit relative to “side effects”. This is nothing new.
@R Watkins and @PatriciaMD – read the full recommendations, not the political spin that AHRQ has beeen putting out for the last week. USPSTF has downgraded their recommendation for screening in women age 75+ to just an “inconclusive evidence.” The did not issue the recommendation against screening in ages 75+ because they were worried about the political backlash.
If you’ve spent years telling people to do something and now you don’t say anything, in the mind of a patient, it’s like telling them not to do it. Tell a patient don’t smoke for years. Then one day, tell them, the benefits or harms of smoking are not totally conclusive. See the response.
The guidelines state, “The USPSTF concludes that the current evidence is insufficient to assess the additional benefits and harms of screening mammography in women 75 years or older.”
The guidelines also state “NO RECOMMENDATION” for screening (in the world of Finance with buy, hold, sell – the guidelines have gone from a buy to a hold).
I’m completely in favor of evidence-based medicine, but let’s have evidence. If the evidence in favor of or in opposition to mammography is inconclusive, then hold on a recommendation and gather evidence. This is a step backward. And I’m serious, I’d love to know that the 3 pediatricians and the one pediatric nurse practitioner brought to the USPSTF recommendations on adult breast cancer.
This wasn’t science, this was trying to get the recommendations in before healthcare reform gets finalized.
Disgrace – No you are not correct. I don’t understand why you assume that I don’t know what “overdiagnosis” is and how it is different from false positives. What I mean by “overdiagnosis” is exactly what everyone else who have bothered to read any study of screening (as opposed to media reports) means by “overdiagnosis”. There is only one definition of “overdiagnosis” in epidemiology. “Overdagnosis” refers to early cancers that are diagnosed as cancers and are treated as cancers but that would’ve never progressed had they remained undetected.
Unlike you, I’ve actually followed the mammogram debate for a while and read a number of papers that tried to estimate overdiagnosis. Like for example this one for example. This earlier paperhas lower estimates of overdiagnosis; however, as anybody with any knowledge of math and logic would understand, their math computations were flawed – this is noticed in “rapid responses” to the paper if your math knowledge isn’t good enough to notice the flaw yourself. There have been reports of overdiagnosis for quite a while with estimates ranging from 5% to 50% of all mammographically detected cancers and most researchers putting it around a third. So no, “overdiagnosis” is not “assumption”, it is a well-known phenomenon and its estimates are based on real data such as for example increase in the number of detected early cancer with no corresponding decrease in number of advanced cancers; or comparisons between excess of detected early cancers in screened vs non-screened population and no corresponding increase in non-screened population after screening has ended.
Overdiagnosis is the most important harm of any screening program since unnecessary cancer treatment has real risks. False positives cause less harm, but it’s still more harm that you think – I think you simply don’t realize how high the cumulative probability of having at least one false positive after years of screening is. Keep in mind that, as I mentioned before, no study showed all cause mortality benefit from screening; so if there is more heart disease in screened group because of this extra anxiety, they wouldn’t be shown in the study. This is by the way an interesting paper on the subject of cancer-specific mortality numbers vs all-cause mortality in mammogram trials. Also correlation between mortality reduction and extra number of detected cancers, show in how many cases mammograms had no effect – this correlation shows more than just overdiagnosis: in addition to overdiagnosed cases it includes cancers that are detected later in unscreened group but are still curable as well as cancers that are so agressive that early detection doesn’t make a difference. It is still useful.
Now, as is noted in USPSTF as well as other reports, overdiagnosis applies to women of all ages, not just those 40-49. However, false positives – and if you bother to read the statistics you’d know how common they are – apply to everyone. Biopsies have risks too. As I said, I think Cochrane numbers don’t include all false positives. Other estimates are much higher – between 30 and 50% with the lower numbers in Europe that has less screening and higher numbers corresponding to the US.
Keep in mind that false positives lead to more tests, including more mammograms and more radiation which is not exactly harmless.
I don’t know why you ASSUME that USPSTF’s main reason is cost rather than balance of benefits and harms – in spite of all the evidence to the contrary icluding no political affiliations of the researchers. Mammograms for women of 40-49 has always been a subject of controversy because of a smaller probability of benefit. In fact in 2002, NCI didn’t want to recommend mammograms for women 40-49; USPSTF recommended them in 2002 simply because of a political outcry. This year there simply was more data as to how small the benefit is for women 40-49… IMHO – this USPSTF decision is actually long overdue.
They make a subjective assessment that the pain and discomfort of biopsies and the anxiety a woman might feel over suspicion of cancer, and this isn’t their call..
If this is the case, why have guidelines at all? They tell you what they recommend or not based on balance of benefits and harms. You are welcome to follow or not. For example, earlir guidelines recommended mammograms, but after careful look at studies of both benefits and harms, I choose not to. And I do have insurance. Let’s just have every person decide which tests they want…. a well as which drug. [/sarcasm]
Pharm Aid:
“The did not issue the recommendation against screening in ages 75+ because they were worried about the political backlash.”
Evidence, please.
“The guidelines also state “NO RECOMMENDATION” for screening (in the world of Finance with buy, hold, sell – the guidelines have gone from a buy to a hold).”
Faulty analogy. “Insufficient evidence” is absolutely not the same thing as “hold.” “Hold” is a positive recommendation; “insufficient evidence” is the absence of a recommendation. Very different.
Pharm Aid: Pharm Aid November 20, 2009 at 10:58 pm
I’ve spent over 30 years in medicine and I’ve never seen such sham science before. I’m sure the pediatric providers (all 4 of them – 25% of this “robust” expert panel) have a wealth of real-world experience in dealing with the highly evolving area of adult oncology. There wasn’t a single, real expert in breast cancer on this panel.
Interpreting the research data from studies that involves 1) number of women screened 2) number of early cancers detected 3) numbers of false positives 4) numbers of advanced cancers detected later 5) mortality reduction is not a job for oncology experts, but epidemiologists – researchers with knowledge of medical statistics. Much like pediatricians aren’t the ones evluating effectiveness of drugs from control studies for the pharmaceutical companies – statisticians do, oncologists aren’t the ones who interpret results of studies of screening or cancer drugs. Oncologists look at individual cases, but they don’t interpret the data. It’s not their job. If you really know anything about medicine, you should know this. Pray tell what specific expertise oncologists gain from their job that would enable them to evaluate the data? Oncologists have anecdotal information from their personal practice, but as has been repeated zillion times “plural of anecdotes is not data”.
As to sham or not sham science, I fail to see how your work as a pharm aid gives you any qualifications to make this determination.
Diora:
>>Oncologists look at individual cases, but they don’t interpret the data. It’s not their job. If you really know anything about medicine, you should know this. Pray tell what specific expertise oncologists gain from their job that would enable them to evaluate the data?>>
You are 100% correct. Physicians receive only very basic training in epidemiology in medical school. They are not epidemiologists, a field which requires advanced graduate work in the field of epidemiology (the best-trained usually have a PhD in epidemiology, or at least an MPH or Master’s degree in epidemiology).
i can assure you from personal experience that sitting through a basic epidemiology class comprised of doctors-to-be when one has had ANY training in study design and statistics is excruciating. The basic epidemiology course is designed to help doctors become more discriminating readers of journal articles and leaves one ill-equipped to do much more without further training.
Science is based on hypotheses, predictions and evaluations of predicted results. It begins with observation which the epidemiologists can do, but that is only the starting point. Hence I object to these conclusions of epidemiologists being labeled as scientific – they are making observations which is the starting point of science, not the endpoint. I agree that the data show that there is a problem. However, they don’t show what the problem is or that the problem is mammograms, rather that there is a whole process that needs investigation, but that its time for the observationalists to step down and let medical science take over. What a scientific group would/should do is trace the steps from mammogram, or PSA reading or other screening technique through biopsy, interpretation, treatment or non treatment decision, and evaluate which paths lead to desirable outcomes, which to undesirable. This of course requires serious input for the vast amount of data collected by radiologists, breast surgeons, oncologists, and geneticists. Then, in a scientific approach various control experiments would be performed to see how the number of desirable and undesirable outcomes changes. In any case, I wouldn’t go to an epidemiologist for advice on my health – would you? I recall when my college was having financial difficulties. One wise professor came up with a proposal which would reduce our deficit to zero. It’s simple he said, let’s close down the college. Similarly the epidemiologists have a fairly useless conclusion if the goal is finding the proper approach to breast cancer. They should be thanked for getting their average data together – and now lets let the professionals who deal with real individuals and the substantative components of diagnosis, disease course and treatment take over and utilize there experience and observations of actual people not mass statistics.
I am not a medical person, and I tend to try to avoid medical care as much as possible although I do go for my annual mammo as part of my annual maintenance visit. I thought it was great that less screening for younger ages was recommended and was a little taken aback by the blowback. The release of the study at the height of the healthcare debate might not have been the best thing. I think the ACS and maybe the imaging industry has really oversold the mammo thing and the effectiveness of lots of tests and treatments that really aren’t that wonderful. We all view medical care as more is better, but it isn’t and we need to start talking about that. And I am not in favor of government involvement in healthcare. I am in favor of the healthcare we used to have back in the 60s…personalized and accessible.
Pat,
Then you want a Health Care Savings Account. Pay cash, and you will get what you want, and you don’t have to worry about the government rationing it, as the insurance companies do now.
Lee Smith:
>>It begins with observation which the epidemiologists can do>>
Do you know what an epidemiologist is?
>>Hence I object to these conclusions of epidemiologists being labeled as scientific – they are making observations which is the starting point of science, not the endpoint.>>
The endpoint in this case was mortality. Not sure how one could use a more final endpoint.
>>I agree that the data show that there is a problem. However, they don’t show what the problem is or that the problem is mammograms>>
What the data showed is that in a large group of women who had mammograms, doing so did not significantly impact mortality in younger age groups but did generate significant rates of false positives and false negatives in this age group. Morbidity rates are more complex but I can expand if you’d like.
>>its time for the observationalists to step down and let medical science take over>>
The irony is that much of what you call “medical science” IS observation and anecdote. That is why so much of “medical science” is not actually very scientific. Sad but true.
Individual clinicians generate anecdotes (case studies or case series), also known as qualitative data, which is only as good as the individual clinician. Large groups of individual clinicians generate quantitative data, the quality of which may also vary or be impacted by confounding factors, some obvious and some not.
>>What a scientific group would/should do is… evaluate which paths lead to desirable outcomes>>
That’s what was done here. Desirable = high sensitivity and specificity of test detecting a condition, treatment of which results in statistically significant lower mortality and non-treatment of which would not result in significantly lower mortality.
This was not a study comparing treatment protocols, which may be where your confusion lies.
>>Then, in a scientific approach various control experiments would be performed to see how the number of desirable and undesirable outcomes changes.>>
Okay. So you’re suggesting that a large number of people should be recruited as subjects for a randomized study. How would this group differ from the group studied by the task force?
>>In any case, I wouldn’t go to an epidemiologist for advice on my health – would you?>>
Sure, if the epidemiologist was also an MD. The epidemiologists I know are smart people.
>>Similarly the epidemiologists have a fairly useless conclusion if the goal is finding the proper approach to breast cancer. They should be thanked for getting their average data together…>>
Well, it may be just a little more complicated than looking at “average data”. Incidentally, most epidemiologists frown upon meta-analysis.
Question: do you think men should be screened for breast cancer? If not, why not?
>>…and now lets let the professionals who deal with real individuals and the substantative components of diagnosis, disease course and treatment take over and utilize there experience and observations of actual people not mass statistics.>>
Okay. How does this differ from the task force’s recommendations?
>>Desirable = high sensitivity and specificity of test detecting a condition, treatment of which results in statistically significant lower mortality and non-treatment of which would not result in significantly lower mortality.>>
That should read: “non-treatment of which would not result in significantly higher mortality”.
I apologize for any confusion.
BD: Do you know what the scientific method is? In any case, epidemiologists should certainly present their approach for consideration by the public. The problem is that this is being labeled “science” when it doesn’t seem to me to be science and having an apparent governmental “seal of approval”. And yes I have a bit of knowledge of the scientific method having a PhD in biomedical science. I admit to being biased by the unbelievably poor recent and highly contaminated “studies” of PSA screening. Talk about confusing the public! I have read a significant amount about DCIS and about prostate cancer (since these things directly impacted on my family) and realize there is a lot of confusion out there but would certainly recommend to anyone I knowto not forgo PSA testing or mammograms and to begin relatively early, following the recommendations of consensus groups of experts in the disease being considered such as NCCN. The flow charts that they develop are much more important, in my opinion, than anything that USPSTF generates.
← Previous Comments
Comments on this entry are closed.