Snowball in a Blizzard

by Steven Hatch

  “Data and evidence have not, do not, and never will be the sole determinants of health coverage”—surely this is an enlightened view of medicine. But its truth is limited here, and it is a rhetorical trick, a play to the crowd, rather than a deep consideration of the real-world consequences of scientific policy. Should we not heed that very data if it means that we don’t have to put one hundred women through Monica Long’s experience? Or should we be comforted that most of them will remain blissfully ignorant that they have been carved up unnecessarily?

  O Canada

  Nearly seven years after the USPSTF guidelines were issued, it is unclear whether they have had any effect. In 2013, a group of researchers looked at changes in mammography rates among nearly 30,000 women who participate in a study known as the National Health Interview Survey. While it is an annual survey, questions about mammography screening are given only every three years, the most recent years being 2005, 2008, and 2011. The final two time points allowed for a natural experiment to see whether the new guidelines had any impact on screening.

  They didn’t. In fact, the trend over these periods showed a slight increase in mammograms, although it was not a statistically significant difference and thus could still be within the natural range of variation. Nevertheless, the study highlighted the fact that the rates didn’t go down in the immediate period after the guidelines were issued. Compare this to the dramatic decline in the number of women who took hormone replacement therapy following the early termination of the Women’s Health Initiative: one study estimated that in 2000 approximately 38 percent of women used some form of hormone replacement; by 2010 it had dropped to just under 6 percent. (I’ll discuss what happened with hormone replacement therapy in a later chapter.) Both events were hugely publicized, so an inability to get the message out can’t account for the difference. The authors of the study noted that the ongoing recommendations of several professional and advocacy organizations probably account for at least some of the effect. Why the new screening mammography guidelines, years after their publication, have not yet caused a rethink among those organizations about the wisdom of screening mammograms in women under fifty is left unaddressed in their analysis.

  In the meantime, further evidence has accumulated supporting the USPSTF’s conservative approach to mammography. In February 2014, the news media reported on the completion of a massive study on the effectiveness of mammograms in Canadian women that was published in the British Medical Journal. The study followed women aged forty to fifty-nine over a twenty-five-year period; women in the control group received annual physical exams designed to palpate lumps, while the other group underwent these breast exams in addition to mammography. The bottom line was stark: the researchers found no benefits at all to women in this age range. Each group had roughly 45,000 women. Roughly speaking, about 3,200 women in each group developed breast cancer, of which 500 in each group died. Compared side by side, one doesn’t even have to resort to statistics to see that there was no benefit to mammography.* Moreover, that held true for women in the fifty to fifty-nine age range as it did for those under fifty. No matter how one looked at the data, one couldn’t find any evidence of benefits, which made the harms of false-positive diagnoses even more dramatic.

  The screened group enrolled 44,925 women, and 3,250 were diagnosed with breast cancer, of whom 500 died. The control group included 44,910 women and had 3,133 cases of breast cancer, of whom 505 died.

  As I completed the first draft of this book in the summer of 2014, the Canadian mammography study—which by all accounts should have caused a massive public reconsideration of the value of mammograms for all women under sixty but especially for women under fifty—had not led to any announcements about a change in policy at Komen or the American Cancer Society. Considering that the Canadian study wasn’t the first of its kind and had come nearly five years after the USPSTF’s systematic analysis, it becomes increasingly difficult to understand their rationale.

  It is possible that by the time this book is published, one or both of these groups will have decided to change their recommendations. Nevertheless, the fact that those recommendations have not yet changed is a story unto itself, and worthy of consideration in its own right. Why was there such resistance to the 2009 guidelines, whose wisdom has now been borne out by further research? I think one does not have to stray far beyond that NBC News online essay I’ve quoted above to arrive at an answer. When health technologies are discussed in terms of righteousness, and aren’t simply thought of as tools by which we might, or might not, extend our lives, it becomes difficult to evaluate the technology for what it is. For screening mammograms are not, and never were, a moral good: they are X-rays whose interpretations incorporate an element of uncertainty. The rhetoric surrounding the value of screening mammograms has been so overdetermined, so loaded with metaphor, that we are blind to what the data tells us, which is that the old policies made matters worse for women’s bodies rather than better. And that is quite astonishing.

  The Medical-Industrial Complex, Up Close and Personal

  Just after the publication of the Canadian mammography trial, I came to the office of Dr. Judy Ockene at the University of Massachusetts with these thoughts on my mind. Judy was one of the dozen and a half experts who formed the committee that issued the 2009 USPSTF report. Her work on that committee is only the latest accomplishment in a career that has spanned five decades, and she considers it one of her proudest. I met her through some seminars that I had attended about the structure of modern scientific collaboration—a subject on which she had considerable expertise after working with colleagues from across the country and for several years on those guidelines.

  As one of the senior faculty at UMass who has been centrally involved in resident and faculty development, she’s a woman who always has a kind word and an encouraging smile. She’s one of the types that large academic institutions are exceptionally good at producing: a scholar of the highest order who also fosters the careers of the next generation of researchers and beyond, a pillar of the academic medical community. Even though I tower over her physically, to me and many other junior faculty members who have encountered her, she’s a titan. Yet you would brush past her in a supermarket without any awareness of her importance in this critical chapter in the history of American medicine.

  She is, in some sense, the face of the medical-industrial complex—that nebulous entity by which certain patients and advocates, sometimes with justice and sometimes without, conceive of the modern medical system. Or at least she is one face that comprises it in part, and mine is another. In its featureless form, the medical-industrial complex is vaguely nefarious; looking at Judy, who stands not much more than five feet, if that, you see a real person who has devoted a lifetime to research on a subject that leaves her eminently qualified to be listened to with care. Talk to her about mammography, and you’ll have a much harder time envisioning the USPSTF as a disembodied aggregation of cold and uncaring scientists, and, in doing so, it becomes a lot harder to ignore the substance that led to the guidelines.

  “I think it was a failure of communication, a big failure of communication,” Judy said as she reflected back on the aftermath of the guidelines in 2009. She noted that the actual report was submitted well in advance of the actual publication date, which could have provided an opportunity to carefully consider the publicity that would inevitably ensue, and perhaps to control the message to a much greater extent. “The failure was that we didn’t ask the questions, ‘How do we frame it in a way that doesn’t get people irritated? How do we help them understand what the recommendations mean?’ There’s a lack of understanding by the public—and even by clinicians and researchers—what these guidelines mean, and that’s understandable,” she said.

  “The people on the committee are scientists, and they’re very busy scientists, but there are communications experts with important public relations skills and knowledge at most of the institutes involved in this work who could have bee
n thinking about those questions. Unfortunately, communication errors were a large preventable problem.”

  I see her point, although I am less sure that even the best PR firm could have added enough spoonfuls of verbal sugar to help these medicine recommendations go down. The concept equating mammography and health had become so entrenched in the public mind that any departure from that, no matter how mild and carefully packaged such statements may have been, would have been a very difficult sell. Coupled with the story line that the medical establishment is capable of bureaucratic indifference—or even hostility—to women through any number of policies, the 2009 guidelines faced a steep uphill battle to be accepted by the public as being in the best interests of women.

  Having a feel for uncertainty in medicine allows us to think about risks and benefits. We compare the size of a benefit with the level of risk, and then consider how sure we can be about those risks and benefits to locate a spot on the spectrum of certainty. The reason the debate about the 2009 USPSTF screening mammogram guidelines is so instructive is that the public, and a good many physicians, was saying, in effect, Screening mammograms are way over here on the spectrum! The Task Force’s reply was, in effect, No, actually it’s over here, much closer to the middle. The red-shift in perception was caused, to a very large extent, by a lack of appreciation for the uncertainty that comes with the technology.

  The experts, however, who were fully aware of that uncertainty, were in consensus. In the following chapter, we’ll see what happens when the uncertainty is so significant that even experts can’t agree what to do. It is perhaps the most difficult and technical to read in this book, thick with studies and numbers, so it is worth considering getting a cup of coffee before proceeding (and I’ll discuss coffee in a later chapter, too).

  4

  THE PRESSURES OF MANAGING PRESSURE

  Dum in dubio est animus, paulo momento huc illuc impellitur. (When a mind is in a state of uncertainty, the smallest impulse directs it to either side.)

  —ROMAN PLAYWRIGHT TERENCE, ANDRIA, ACT 1, SCENE 5

  We’ve just seen how uncertainty can have a major impact on public health guidelines and how the consequent public discussion is framed. In approaching guidelines, doctors and patients alike might regard them as simple yes-or-no, up-or-down recommendations. However guidelines could likewise be understood as policy statements of relative benefits versus relative risks, documents that eschew the hard-and-fast language of “do this, don’t do that” and opt instead for a more nuanced view of what the scientific research can actually tell us about some topic. One of the reasons the US Preventive Services Task Force’s 2009 mammogram recommendations induced so much cognitive dissonance in the public when the report was issued was that the vast majority of the public (to say nothing of the many doctors who focus on women’s health) assumed the recommendations were going to look like the first kind of “marching orders” guideline, although the task force itself consciously intended to provide the second kind of recommendation, which put the uncertainty surrounding screening mammography’s benefits front and center.

  Yet it is important to note that although many physicians strongly disagreed with those guidelines (and, indeed, nearly seven years after the report’s publication, many still do), the committee itself was unanimous. The experts assembled, sixteen in all, reviewed a mountain of data and presented a picture of screening mammography as a technology that occupied more than one spot on the spectrum of certainty, depending on a woman’s demographics. The value of screening mammograms for healthy women above the age of fifty but under the age of seventy-four was regarded as being to the left of center. They believed there was either a big suggestion of a small benefit or possibly a small hint of a big benefit. In other words, there was moderate uncertainty of its benefits, but the benefits seemed persuasive (although these recommendations came out before the results of the large and well-designed Canadian mammography trial that showed no benefits at all, which may have influenced the 2009 report). But the evidence in the benefits of screening mammograms in women under fifty was regarded as far less persuasive, and so screening mammograms in that group was placed either in the absolute center, or possibly a little to the right of center on the spectrum of certainty, where it suggested net harms.

  The bottom line was that these sixteen experts, who had laboriously reviewed the evidence for months, and who must have had marathon discussion sessions about how to convey the subtleties of these findings, were completely unified in standing behind these guidelines. While there was a “controversy” in the public debate, there was nowhere near the same level of contentiousness within the task force itself.

  The following story, however, is about a guideline task force in which the experts did split, and as such this chapter represents the relatively rare moment when true biomedical controversy exists on a large scale. While there are many disputes in medicine about the optimal treatment for a condition, or the best strategy to manage a given disease, or the most accurate method to arrive at a diagnosis, most of these arguments deal with fairly fine points, and there is much consensus on the general approaches to various maladies. But here we’ll see an example of a situation where even highly trained doctors looked to the experts and saw nothing but discord.

  It may seem ever so slightly sadistic to include a chapter dealing with this degree of uncertainty in a book for a general audience. If the experts can’t even agree, then how on earth should a layperson be able to sort it out? I include this story not because I wish to induce a sense of throw-your-hands-up hopelessness, but rather because understanding the role that uncertainty plays in genuinely controversial policies allows all of us to establish what we do know and do not know about a medical topic. And as I hope to underline further at the conclusion of the book, having the ability to sort out what is known from what is not can be genuinely liberating for patients. Moreover, I include it to underline that treatments always carry the risk of harms, and any time we pop a pill, we should regard that action as the culmination of a risk-versus-benefit equation that should land squarely on the side of benefit. As patients, we should always be cognizant of that equation.

  So there is good news in this chapter, although at first glance it may seem daunting and frustrating. But bear with me. Go slow as you read this, for the data can seem confusing—not surprising, because it caused the expert consensus to crack! But even for nonmedical laypeople it’s not impossible to understand the parameters of the debate. If you start to get bogged down in the studies, just remember that the question that divided the specialists was this: For people with high blood pressure, what is the optimal level to which we should attempt to lower it?

  Time to dive in.

  JNC8, or Breaking Up Is Hard to Do, Expert Edition

  In February 2014, the members of a group called the Joint National Committee (a panel of more than a dozen experts assembled by the National Heart, Lung, and Blood Institute, which is a key division of the National Institutes of Health) issued its long-awaited recommendations for the diagnosis and management of high blood pressure. Because high blood pressure has been one of the central and most lethal problems facing Americans for several generations, the JNC has been periodically updating its guidelines since its first report in 1977. This most recent report was the seventh revision of the original, and thus has been shorthanded in the medical community as the JNC8.

  It is safe to say that there will not be a JNC9 anytime in the near future, for the process of producing JNC8 was so fraught with contentiousness that the entire method by which national guidelines are generated is unlikely to continue. Seven months before the publication of JNC8, the leadership of the National Heart, Lung, and Blood Institute issued a long statement discussing the process of the guideline formation, and although the text of the announcement is somewhat cryptic—they state, for instance, that they would publish “five integrated cardiovascular guideline products . . . as evidentiary reviews,” whatever that means—it seemed clear that they
were getting out of the guideline business.

  The source of the dispute could be found in matters both practical and theoretical to the function of the JNC8 committee. The practical issues centered on blood pressure targets for one particular group of patients, while the theoretical issues related to how one included or excluded evidence in forming the guidelines. By the time the JNC8 guidelines were issued, a minority group had published its own separate report in a different journal.* Not only would patients find such conflicting recommendations difficult to absorb, workaday docs would be equally hard put to the task of knowing which set of recommendations to follow unless they immersed themselves in the data to arrive at their own conclusions. In the case of the JNC8 guidelines, immersion can quickly lead to drowning.

  Five of the members of the JNC8 committee peeled away from the remaining twelve to issue their report in 2014. Again, contrast this with the 2009 USPSTF recommendations on screening mammograms, in which all sixteen authors supported the report. Controversy has its gradations, and, as we’ll see in the next chapter, on Lyme, sometimes patients think there is a raging debate with much uncertainty, when in fact there’s virtually no debate at all and a pretty high degree of certainty about the major points of diagnosis and treatment.

  JNC8 represented a departure from JNC7 in a manner similar to the revised mammography recommendations by the USPSTF in 2009 in that they proposed scaling back treatment goals. The table on the next page summarizes the differences between the two sets of guidelines:

  In patients over the age of eighty, there were no changes: the goal blood pressure was 150/90 or lower. There were new, higher goal blood pressures for patients with either concomitant diabetes or chronic kidney disease, with the previous target being 130/80 and the new target 140/90. This may seem counterintuitive for people who have long heard that a “normal” and healthy blood pressure is 120/80, but when the committee reviewed the data, they believed that there was no evidence to suggest that the lower blood pressure prevented the complications of these diseases. Although these were higher goals, the full JNC8 concurred on the new targets for these patients.