What the November, 2009, Breast Cancer Screening Argument Obscured

In November, 2009, a rancorous argument about screening mammography for women aged 40-49 was touched off by the publication of updated guidelines(1), supported by a systematic literature review(2) by the US Preventive Services Task Force (USPSTF).  The guidelines suggested that yearly mammographic screening for women in that age group should not be automatic, but a decision made for individual patients after discussion between the patients and their doctors.  This was based on a critical review of the best available data which suggested that the benefits of screening acrue to only a few patients.  1904 women would have to start screening and continue for multiple rounds to prevent one cancer death over 11-20 years of follow-up.  These benefits had to be balanced against a number of potential risks, including the risks of treatment of cancers that might never behave malignantly, and the at least theoretical risk of generating new cancers through radiation exposure from mammography. 

These seemingly reasonable recommendations generated heated responses.  The debate, to be charitable, seemed to be at its core about how one should weigh benefits and harms in making individual and health policy decisions.  Since different people value different outcomes differently, I was not sure at the time how to make a meaningful contribution to this debate, or whether the debate had to do with the issues we usually discuss on Health Care Renewal.

I should note that the USPSTF guidelines never said "do not screen" women under age 50, or that the government should not pay for such screening.  They did say "the decision to start regular, biennial screening mammography before the age of 50 years should be an individual one and take pateint context into account, including the patient's values regarding specific benefits and harms."  It is hard to see how anyone could argue with that as an expression empowering patients' choices and values.  (For further discussion about how the recommendations were actually modest and reasonable, see Partridge and Winer.)(3)

Now that the dust has settled, it may be useful to reflect further on this.  Doing so suggests that the rancorous US debate mainly obscured rather than illuminated the major issues regarding mammography screening, particularly about our lack of clear evidence from clinical research needed to make the best individual and  policy decisions about mammography. 

It seems to me that the main questions one must answer to make an individual or policy decision about screening mammography are:

Does it improve longevity?

This is not the same as asking whether mammography reduces mortality from breast cancer.  It is theoretically possible that while decreasing mortality due to breast cancer, screening and its downstream consequences increases mortality from other causes.  At least in theory, screening may detect small tumors that would never grow or metastasize.  Treating such tumors could sometimes lead to premature death due to complications of surgery, radiation, or chemotherapy.  Furthermore, screening also involves periodically exposing large numbers of women to radiation, which may sometimes cause new tumors.  So reducing breast cancer mortality does not automatically mean that overall longevity would be improved.

There have been eight major trials of breast cancer screening which included women younger than 50. (See reference 2.)  None demonstrated a statistically significant increase in overall survival (that is, an increase unlikely to have been due to chance alone) due to screening.

Does it reduce suffering, improve functioning or generally improve quality of life?

To my knowledge, no major trial attempted to answer this question.  No such data is mentioned in the USPSTF systematic review.

Do the above benefits outweigh all its potential harms and risks?

So we cannot answer this question, because the benefits that might be most meaningful to patients (overall survival, symptom reduction, functional improvement, overall quality of life improvement) have not been clearly measured.

A Lack of Relevant Evidence

So the USPSTF guidelines, like other relevant guidelines, were based on the evidence that is available.  Since the evidence did not directly answer the most important questions, the guideline writers were left doing the best they could with evidence that only indirectly addressed the main issues.  No wonder they ended up unable to make a clear recommendation, and leaving the decisions to individual discussions, and individual discussions that would necessarily hinge on guesses about the unknown. 

One would think that a big point of discussion about breast cancer screening would be why after eight trials enrolling a total of about 350,000 patients reported over 20 years we still cannot answer the big clinical questions.  A related point for discussion in the US is why only one, and the earliest trial was conducted here.  If we here in the US think breast cancer screening is such a major concern (and we should think so), why have we been unable to mount a single important trial of it since the HIP trial conducted more than 30 years ago?

Instead, the rancorous debate in the US included...

Anecdotes, Some Irrelevant

The press found a number of women who said they would not be alive were it not for screening mammography before age 50.  With all due respect, one cannot tell whether an individual whose tumor was found on screening mammography would still have been diagnosed early enough for succesful treatment in the absence of screening mammography.  (And also with all due respect, we have no idea whether there also are cases of women who died as a result of treatments of tumors that never would have progressed, or cases of women who died of tumors caused by radiation from multiple mammograms.)  Reasoning from single cases when people, diseases, and treatment results vary so much is likely to mislead. 

It is somewhat ironic that some of the cases cited were of women who had breast cancer diagnosed before age 40, even though the debate was supposedly about screening from ages 40-49.  For example, in an inflammatory article that suggested that some "oncologists" might want the USPSTF sent to the prison at Guantanamo Bay, Washington Post editorialist Dana Milbank cited cancer activist Nancy Brinker, who mentioned her sister "whose breast cancer was found with a mammogram at age 37," (and apparently who tragically is no longer alive).(4) 

Going Well Beyond the Evidence

As noted above, no trial has shown that screening mammography for women under 50 increases overall longevity.  We all hope it does, but so far, there is no clear evidence that it does. 

Yet multiple media reports included assertions that screening mammography saves lives.   For example, the breast cancer activist mentioned above said, "mammography saves lives," apparently including mammography under age 50.(4)  An op-ed column by Dr Alan Kaye, chairman of radiology at Bridgeport Hospital, asserted "large, multinational research studies have shown that mammography saves lives in all age groups covered by the current guidelines."(5)  I would challenge him to show me a single such study that found a statistically significant increase in overall survival for patients under 50.  A Texas radiologist stated, "I diagnosed a 40-year old woman with breast cancer last week.  If she had waited 10 years, with pre-menopause breast cancer she would have been dead."(6)  Unfortunately, since she was just diagnosed, how can he be certain that she will survive any given amount of time?  How could he know that the cancer might not have become manifest, absent that single mammogram, later while still treatable? 

I do not want to be too hard on patients who do not appreciate that the outcomes of testing and treatment for breast cancer are not certain.  However, one would hope that physicians would be able to deal with this uncertainty.

Conflicted Opinions

Some of the more strident discourse came from those who may have had financial vested interests in promoting screening mammography.  Fugh-Berman and Bell pointed out numerous "fact-free emotionally charged statements" made by people who appeared to "reading from the same script-book."(7)  They identified that many of the loudest critics of the USPSTF guidelines were affiliated with not-for-profit organizations with impressive names, but also with substantial financial support from corporations that make products used in mammography.  Also, some had personal financial relationships with such corporations. 

An op-ed article by former US Food and Drug Agency (FDA) commissioner Dr Andrew von Eschenbach and Ms Nancy Desmond distorted the USPSTF guidelines to mean "most women should delay screening until they are 50," and claimed that was based on cost concerns, not clinical evidence.(8).  Desmond is the CEO of and von Eschenbach is now a senior advisor to the Center for Health Transformation.  The Center's members include numerous pharmaceutical and device manufacturing corporations, including several that make mammography equipment (e.g., GE Healthcare and Siemens).

Summary

Cancer, especially breast cancer, has major emotional connotations, and can be a difficult issue to deal with from many people.  The conflicting emotions cancer brings out in many patients may understandably affect their physicians, as well as friends and family.  Nonetheless, physicians, other health policy professionals, and health policy experts can serve patients better if they do not allow the patients' understandable affective responses cloud their understanding of the clinical and scientific issues. 

Yet the late 2009 debate in the US about screening mammography included many responses in which emotion seemed to overwhelm reason.  It may also be that some such responses came from people who had vested interests, or whose employers had vested interests that supported the emotional, rather than the reasoned approach.  Meanwhile, no one seemed to acknowledge that a big reason we are still debating this topic is that we have not made the effort or expended the resources to do good trials of sufficient size to answer the questions that need answering.  Of course, such trials might provide answers that would upset some people, or threaten others' incomes.  (As one news article pointed out, mammography is now a $5 billion a year industry in the US.)(6)

So my end of annus horribilis 2009 message on Health Care Renewal is to better serve our patients, from 2010 onward we health care professionals need  to try harder to put evidence and logic ahead of our own emotions, and certainly ahead of our financial self-interest.  

Note that numerous bloggers have taken on this topic, so see posts on Respectful Insolence, GoozNews,  Health Care Organizational Ethics, and the Evidence in Medicine blog.
References

1. US Preventive Services Task Force. Screening for breast cancer: U.S. Preventive Services Task Force recommendation statement.  Ann Intern Med 2009; 151: 716-726. [link here]
2. Nelson HD, Tyne K, Naik A et al. Screening for breast cancer: an update for the U.S. Preventive Services Task Force.   Ann Intern Med 2009; 151: 727-737.  [link here]
3. Partridge AH, Winer EP. On mammography - more agreement than disagreement.  N Engl J Med 2009; 361: 2499-2501. [link here]
4. Milbank D. Feeling farther from the finish.  Washington Post, Nov 24, 2009.  [link here] 
5. Kaye A. An alarming retreat on early detection.  Hartford Courant, Nov 25, 2009 [link here]
6.  Jacobson SJ. Dallas-area clinics ignore proposed rules, still push for mammograms. Dallas News, Nov 27, 2009.  [link here]
7. Fugh-Berman A, Bell A. Mammography and the corporate breast.  Bioethics Forum, Nov 24, 2009.  [link here]
8. von Eschenbach A, Desmond N. Government panels can't put price on human life. Associated Press, Nov 24, 2009.  [link here]

On Automobile, and Health Care Companies Run by Finance People

The New Republic published "Upper Mismanagement" about what happens when businesses are run by people who do not understand their companies' businesses.  Although the article was focused on the decline of manufacturing in the US, its applicability to health care is obvious:

Harvard business professor Rakesh Khurana, with whom I discussed these questions at length, observes that most of GM�s top executives in recent decades hailed from a finance rather than an operations background. (Outgoing GM CEO Fritz Henderson and his failed predecessor, Rick Wagoner, both worked their way up from the company�s vaunted Treasurer�s office.) But these executives were frequently numb to the sorts of innovations that enable high-quality production at low cost. As Khurana quips, �That�s how you end up with GM rather than Toyota.�

How did we get to this point? In some sense, it�s the result of broad historical and economic forces. Up until World War I, the archetypal manufacturing CEO was production oriented�usually an engineer or inventor of some kind. Even as late as the 1930s, business school curriculums focused mostly on production. Khurana notes that many schools during this era had mini-factories on campus to train future managers.

After World War II, large corporations went on acquisition binges and turned themselves into massive conglomerates. In their landmark Harvard Business Review article from 1980, 'Managing Our Way to Economic Decline,' Robert Hayes and William Abernathy pointed out that the conglomerate structure forced managers to think of their firms as a collection of financial assets, where the goal was to allocate capital efficiently, rather than as makers of specific products, where the goal was to maximize quality and long-term* market share.

By the 1980s, the conglomerate boom was reversing itself. Investors began seizing control of overgrown public companies and breaking them up. But this task was, if anything, even more dependent on fluency in financial abstractions. The leveraged-buyout boom produced a whole generation of finance tycoons�the Michael Milkens of the world�whose ability to value corporate assets was far more important than their ability to run them.

The new managerial class tended to neglect process innovation because it was hard to justify in a quarterly earnings report, where metrics like �return on investment� reigned supreme. 'In an era of management by the numbers, many American managers � are reluctant to invest heavily in the development of new manufacturing processes,' Hayes and Abernathy wrote. 'Many of them have effectively forsworn long-term technological superiority as a competitive weapon.' By contrast, European and Japanese manufacturers, who lived and died on the strength of their exports, innovated relentlessly

Furthermore,

The business schools had their own incentives to channel students into high-paying fields like finance, thanks to the rising importance of school rankings, which heavily weighted starting salaries. The career offices at places like Harvard, Stanford, and Chicago institutionalized the process�for example, by making it easier for Wall Street outfits and consulting firms to recruit on campus. A recent Harvard Business School case study about General Electric shows that the company had so much trouble competing for MBAs that it decided to woo top graduates from non-elite schools rather than settle for elite-school graduates in the bottom half or bottom quarter of their classes.

No surprise then that, over time, the faculty and curriculum at the Harvards and Stanfords of the world began to evolve. 'If you look at the distribution of faculty at leading business schools,' says Khurana, '�they�re mostly in finance. � Business schools are responsive to changes in the external environment.' Which meant that, even if a student aspired to become a top operations man (or woman) at a big industrial company, the infrastructure to teach him didn�t really exist.

If business schools did little to teach about manufacturing, they did almost nothing to teach about health care. But at the same time the finance people were taking over manufacturing, health care organizations were pushed to turn over their leadership to business people to improve efficiency and break the physician's "guild." Would there be any reason to expect that a finance background would be better preparation to run a health care corporation than to run an automobile company?

For the latest thought- and wince-provoking example of how leaders of health care corporations seem to know almost nothing about the actual health care their companies provide, see a DailyFinance interview with Mr Kent Thiry, CEO of DaVita, a for-profit corporate provider of dialysis services. According to the company web-site, "prior to working for Vivra, Mr. Thiry was a partner at Bain & Company, an International management consulting company. He earned his BA degree, with distinction and Phi Beta Kappa, in Political Science from Stanford University in 1978, and his MBA, with honors from Harvard Business School in 1983"  So he got his MBA from an elite US business school at the time in which finance was becoming dominant as described above. 

Asked to explain his business model, Mr Thiry responded:

Most of us have a couple of kidneys. These kidneys are amazing organs -- some of the most complex, sophisticated organs in the human body, which is why they've been so difficult to replicate compared to other organs like the heart and lung and others. And when the kidney fails, you need to go on dialysis, unless you're one of the fortunate few to get a transplant. And we operate the centers that people come to if their kidney fails and they can't get a transplant.

And what we do in our centers is take care of these people typically three times a week -- four hours each time -- where we take their blood out of their body, clean out all the toxins that they would normally clean out themselves through the act of urinating. But you don't do that anymore once you've lost your kidney function. And we take that part out, take the toxins out and then put the blood back in with some other nutrients.

To be charitable, I do not think that would merit a "C" on a high school biology test. [Medical science cannot "replicate" hearts or lungs. Kidney function is not the same thing as "urinating." The functions of the kidney are far more complex than "cleaning out toxins."]

Does it make any sense to put someone who obviously understands so little about kidneys in charge of a kidney dialysis company?  (On the other hand, see this post on accusations that DaVita's ruthless business practices treat patients like "dialysis dollars.")

So, if putting finance people in charge of automobile companies turned out to be a recipe for bankruptcy, why should we expect from putting finance people in charge of dialysis companies, or hospitals, or drug, biotechnology or medical device companies, or health care insurance companies, or health care information technology companies?

The CEOs of big health care organizations, most of whom have business, not health care backgrounds, have mainly been good at paying themselves and their cronies well. (For example, according to the 2009 DaVita proxy statement, in 2008, Mr Thiry owned over 2 million shares of stock, 1.9% of all shares outstanding, and received more than $11 million in total compensation.  Clearly, he was not paid according to his knowledge of kidney biology.)  Meanwhile, health care costs rise, access falls, and quality degrades.

If we really want to reform health care, maybe we should take a lesson from Toyota. Put the car guys and gals in charge of car companies. And put the health care guys and gals in charge of health care.