Then I read some fellow saying the Ken Thorpe is wrong about the problems with the health care system. I don't know if this fellow is right or wrong, I haven't bothered digging lest I discover another conservative ideologue, but he's using breast cancer screening as an example! This month of all months! And it's an irresponsible argument at that! Not to mention that Ken Thorpe didn't make any arguments in his article.
Leave to Health Despairs [sic] to claim that the major reason the US spends more on health care is because we are sicker. Confusing diagnosis (which is a function of being able to see a doctor multiplied by what the doctor can do time what the doctor gets paid for) for prevalence Ken Thorpe et al. claim that at least part of the underlying difference in prevalence is poorer health status. Really? He goes on to estimate that the prevalence of diagnosed cancer was 12.2 percent in the United States but only 5.4 percent in Europe in 2004. A huge difference. Even Thorpe has to write: "Are Americans really more likely to develop malignant tumors, or are they just screened more intensely than Europeans are? Comparisons of breast cancer screening rates and five-year cancer survival rates suggest the latter"Bad idea. It's a little more complicated and this guy doesn't seem to know Thorpe, since anyone who has met the man would be surprised if he ever made a claim that was not supported by the data. I would respectfully suggest the study is an example of how one tests a hypothesis. Once you design the study, you follow the data, wherever it may take you. Authors are permitted to suggest potential explanations for the data, but never state a 'claim', lest their objectivity and intellectual integrity be challenged.
The US is heavily doctored, but access is uneven. The wealthy take care of themselves wherever they may be. The middle class takes the brunt of profit-making, getting access, but not necessarily access to the best. How can you tell if someone is good anyway? How can you tell if you're being encouraged to have an investigation or a procedure just because it is billable?
So the hypothesis that more cancers are being detected earlier is certainly possible, given the data.
But the American health care system is not driven by cost-effectiveness as socialized countries are. There is no point doing a mammogram in a young woman because the test is not sufficiently discriminative until most women get into their 40's. It also detects a lot of abnormalities that are not cancer. PSA's are just not done in low-risk individuals for fear of extra cost for no benefit and sometimes even the potential of harm.
In the US, we have a consumerist society where the patient, now a client, makes the decision themselves. It's a little like picking out a purse at Coach, "I want two preventions and a diagnostic."
So in the middle of breast cancer awareness month, Robert Goldberg is suggesting that the European system rations health care and gives the example of breast cancer and mammography rates. However, in Europe mammography coverage is higher than it is in the US. So they don't need to ration by preventing people from having so-called unnecessary procedures. The procedures have, in fact, become moot. The are sufficient misconceptions about breast cancer, that women don't need someone muddying the waters.
The kicker, I guess, is that the US rations health care more than socialized economies. We ration by money so if the cost is prohibitive the test or procedure simply doesn't get done. Could this have an effect on Dr. Thorpe's data? Why don't we hear about it in the US? Well, the middle class can be vocal, especially when they can't get the handbag -- uh, I mean procedures they want, and still can't afford a better quality of advice... the one that tells them about the tests and procedures they don't need.
The US spends more money on health care because it misplaces its resources. Period.
Ladies, get your mammograms. For this test, there are funds available so that no woman in America should go without. I suspect, when all is said and done, it will be clear that European women who get free mammograms benefit from earlier detection. Thus these countries enjoy a remarkably low cancer rate. In the US, it is not a matter of cost, it is that women don't know they can get free mammograms through their local health departments. Citizenship doesn't matter. And it could save your life.
9 comments:
"Ladies, get your mammograms"
Let's see. If I do mammograms for 10 years, the probability that it'll save my life is around 1/1000 (ok maybe 1/500 if I take an optimistic view of the studies). Granted, if I am this one who beneefits, it is the same of 100% to me.
But... There is price to pay for this benefit. During the same 10 years my risk of having at least one false positive is between 20-50%. Most of these false positives are just for additional films, but about a quarter of them will end in biopsy.
Most importantly - there is a risk of overdiagnosis and consequently unnecessary treatment that can cause complication. The estimates of that are ranging from 5 to 40% with consensus (at least according to NIH) being around 30%. This is pretty huge.
Not on the one hand there is a slightly reduced chance (at least if one uses ARR instead of totally meaningless RRR) of dying from breast cancer. On the other hand there is a higher risk of being diagnosed with breast cancer.
At least to me, the choice is not as obvious as you say.
Thanks for the evaluation of the risks and benefits of breast cancer. It is helpful to hear the analysis, but I would be distressed if it convinced a woman to avoid mammography based on an incomplete assessment of all the data.
The majority of the data certainly does not support much contention about the overall value of annual mammography in women over age 50. I would add concern about the lack of a specified top age, at which mammography may start to lose value, (perhaps more a function of comorbidity, rather than age itself.) So yes, there are limitations that all women must be aware of, but too few women are screened for us to begin adding confusing caveats.
In fact, I am not aware of a single study where women find the decision to have a mammogram as complicated as you do. Would you like to take on PSA? I promise to be very supportive of an articulate argument based on ARR and complication rates of prostate cancer screening.
In the meantime, the recommendation for mammography stands, with the exception of a single European research group who appear to be talented at shaking the cage. Hardly consensus, but worthy of commentary, as long as we do not allow the dissenting voices, so necessary to the scientific method, to confuse an otherwise trusting public.
'I am not aware of a single study where women find the decision to have a mammogram as complicated as you do. "
Most women overestimate the benefit, and are almost completely unaware of even the concept of overdiagnosis. Guidelines are fine because they are based on the existance of the overall mortality benefit for the population. But an individual may still thing differently of what is right for her. Doesn't an otherwise trusting public have a right to know the true probability of an individual benefit as well as the potential risks? I'd imagine the vast majority of women would want to follow the guidelines and be completely uninterested in details. Some individuals, however, may want to be informed of both the magnitude of benefit and the risks.
Given no evidence of benefit in PSA screening (you cannot really talk about ARR of death if there is no evidence of any risk reduction?). In terms of articulate argument, have you read the book of H. Welch "Should I be tested for cancer"? He tackles the subject of screening much more articulately than I ever could.
There is a benefit to PSA, and a risk reduction to be had by screening. The problem is when you factor in the complication rates.
The frequently quoted rate of complications from investigating abnormal PSA can push the balance of benefit and risk form positive to negative.
Both the accuracy and complications of PSA are a moving target. As with most technologies, there is constant improvement, which means we must all be ready to re-examine the data frequently.
As such, the guidelines state that PSA should be OFFERED to all males over 50 IN ADDITION TO annual digital rectal examination. In addition, annual PSA is RECOMMENDED for high-risk groups such as African-Americans and people with a family history of prostate cancer.
The logic is, although we will miss a bunch of prostate cancers, we can minimize the risk of unnecessary procedures and false alarms.
What's the big deal? The only people I can think of that wouldn't want a woman to have a mammogram are men.
Oh I don't know.
You could get a mammogram, and end up with DCIS - a non-life threatening event. By broad accounts between 20-40% of this condition will turn into an invasive life-threatening cancer.
Unfortunately our "superior" science has no way of telling which 60% will never be a problem -- so it's all treated. Oftentimes with mastectomy, and if the woman goes nuclear about it a bilateral mastectomy.
I think sometimes we screen better than treat.
"There is a benefit to PSA, and a risk reduction to be had by screening."
Would you care to show a reference to a single RCT that shows reduced prostate cancer mortality with PSA screening? Only studies that show mortalithy reduction per number of people qualify, anything else is subject to lead-time and length- biases. USPSTF recommends that you discuss pros and cons with patients and clearly states insufficient evidence.
In terms of overdiagnosis in mammography - check out BMJ publication from last year about the rate of overdiagnosis based on Malmo data. There are other estimates, but this one is based on real data rather than modeling. They estimate it at 10% of all detected cancers, but read rapid responses to the article. Their math is flawed as it wasn't adjusted for dilution (when denominator is greater, the ratio is less). When corrected, the number would be at 25% of all detected cases - this is pretty big. Especially if you consider the actual probability of an individual benefitting. USPSTF estimates over 1200 women screened for 10 years to save one life. Even if you look at more optimistic benefit of 500 / 10 years, it is still small. Also keep in mind that studies only show reduction in breast cancer mortality. Yes, I understand that it would be very difficult to have a study that would show statistically significant effect on all-cause mortality because breast cancer deaths represent only a small percentage of all deaths. But the fact remains that nobody really knows if the number of women saved by screening is really greater than the number of women who'd have, for example, heart desease because of radiation for overdiagnosed cases.
By the way, here is an interesting paper on breast-cancer mortality vs all-cause mortality in mammogram trials. Pretty interesting data.
By the way, I do recommend Welch book "Should I be tested for cancer. Maybe not and here is why". It is written for lay people, but some doctors should read it too. I believe the reviews of this book in medical journals were fairly positive.
I may be approaching it the wrong way, but I am trying to say that there is an improvement in outcome from PSA screening that is not sufficiently balanced by adverse events.
Perhaps you are implying that studies of PSA screening are still not even good enough to convincingly overcome lead time bias. I have not reviewed the data in such depth in the past five years, so I'll take your word for it. Suffice it to say, i am not a fan of UNIVERSAL PSA screening.
Your points on the Malmo data are well taken.
I would like to point out that the level of cost and benefit at which we decide to recommend a screening procedure is a political decision, not a scientific one. AS LONG AS PEOPLE UNDERSTAND THE DATA, THE LIMITATIONS AND THE SHORTCOMINGS.
The points we are discussing are the methodological fine points of adding risk and benefit. But given what I know and what I have lived through, 1 life per 1200 women is a pretty darned good reason to do mammograms. It's a lot better place to put scarce resources than 400 NNT for cholesterol, in my opinion. But that's politics, not data.
"1 life per 1200 women is a pretty darned good reason to do mammograms"
I would agree that it is a good reason to make sure it is covered by insurance and for making it available for every woman including the uninsured (provided that you are ready to cover the cost of treatment as well). Especially if individuals understand the cost in terms of overtreatment and false positives. My problem is that most women aren't told about these risks. I would also second Welch in arguing that a decision of an individual woman to forgo mammograms is a reasonable choice given the risks and the uncertainty about if this 1/1200 in desease-specific mortality translates into all-cause mortality.
Agreeed on cholesterol as well - especially as a healthy slim female with 10-year heart attack risk of under 1% whose doctor wanted to put me on statins for mildly elevated LDL.
Post a Comment