On December 17, 2015, JAMA Surgery published an opinion piece written jointly by a surgical oncologist (Dr. Ismail Jotoi) and the chief architect of the Canadian Screening Trials (Dr. Anthony B. Miller). To add gravity and authority to their cause, they invoke the timeworn admonition in their title (in Latin, of course) – Primum Non Nocere – “Above all, do no harm.” The irony is too much – if we follow their advice to quit screening at 70, then approximately 1,000 additional breast cancer deaths will occur each year, give or take a few. This is a conservative estimate, by the way, given that mammographic compliance isn’t that great to begin with. Primum Non Nocere?
This overworked “do no harm” is mostly used today when someone wants to preface their opinion that everyone should stop doing something the author believes is harmful. It’s a sure bet to strike a chord because there’s potential harm in everything we do in medicine. The phrase is attributed to the Hippocratic Oath, which actually stated, “to abstain from doing harm.” To follow this admonition literally, of course, one would have to cease practicing medicine.
So, what was Primum Non Nocere intended to mean? Originally, it stated that if you weren’t sure whether or not a therapeutic intervention would help your patient, then it’s better to do nothing than risk harm. Of course, when the Hippocratic Oath was widely adopted, doing harm was the norm. Very few medical practices of the day actually benefitted the patient, so it’s best we forget about the original meaning. Today, the phrase is meant to imply a weighing of risks vs. benefits. If the risks (harms) outweigh the benefits, “don’t do it.”
And this is where the controversy in mammography screening lies – harms vs. benefits – with anti-screening forces exaggerating harms to the point of publishing studies of permanent psychological damage after a benign biopsy. At the same time, benefits are minimized, such as the insistence on quoting mortality reductions based on “invitation to screen” rather than the benefit as measured in those who actually screen.
The authors in this opinion piece seem to have no restrictions on what they’re willing to call a “harm.” For instance, “lead time” is listed as a harm, complete with a reference. Lead time is NOT a harm, but a built-in bias. For the sake of clarity, “lead time” is one of the primary epidemiologic biases that can be controlled through prospective, randomized trials with mortality as the endpoint. Unchecked, lead time may exaggerate the benefit of screening, but that is not a direct harm to the patient. To list it as such is a remarkable display of anti-screening bias, the irony being that it is misusing a term involving bias.
Only one of the epidemiologic biases can actually result in harm, and this is “overdiagnosis bias,” which can result in overtreatment. As for the other biases – lead time bias, length bias, selection bias, etc. – they don’t harm anyone. Furthermore, these biases have been held in check with meta-analyses of all prospective, randomized trials of mammographic screening where there is a proven mortality reduction. It is bizarre that “lead time” would be newly tagged as a harm. What’s next? Mammographic screening contributes to global warming?
Dr. Anthony Miller can be forgiven for his opinions. After all, he crossed over to the Dark Side of the Force decades ago, and can only interpret the world of screening in one way – a good clinical exam is good enough. He has spent his entire career promoting clinical exam over mammography, in spite of the fact that there’s not a shred of evidence that lives are saved through physical exam alone. His own CNBSS studies are used to denigrate mammography, even though quality in the Canadian trial was so outrageously bad that only 32% of the breast cancers diagnosed in the mammography limb were actually discovered by mammography. No wonder clinical exam looks good to Dr. Miller – his mammograms missed most of the cancers! A typical screening trial will flip-flop the numbers seen in Canada, that is, 68% of the cancers will be found on mammography while 32% will be felt on exam. One can trace Dr. Miller’s writings on the superiority of clinical exam over mammography back over 4 decades, even before he knew the results from his own trial. Absolutely nothing is going to alter his opinion.
The authors have loaded this opinion piece with their own personal bias. For instance, the claim is made that the UK Age Trial “failed to show a significant benefit from screening.” The reference for this statement is an article Dr. Jatoi wrote in 2011. Yet, in the latest 2015 update from that trial, widely accessible through Lancet Oncology (2015; 16:1123-1132), the data shows something else entirely. For background, the UK Age Trial was designed to study screening women in their 40s. (No other age groups, just 40-49, so I’m not sure why it was even included in this article about women 70 and older.) With far better design than the Canadian trial and vastly superior mammography to the Canadian trial, when Dr. Jatoi wrote his 2011 article noting “no benefit,” there was a clear trend toward benefit even though it had not yet reached statistical significance. Then, in September 2015, the final results were announced –
For the first time ever for women in their 40s, a statistically significant 25% reduction in breast cancer mortality was demonstrated for the 1st 10 years after diagnosis for those in the mammography limb. Beyond 10 years, this difference gradually declined to 12% (and barely lost its statistical significance). It was totally predictable that the anti-screeners would call the trial a failure. But what really happened? Did mammography save lives?
In fact, women in the control group (no mammography), after a decade of no screening in the Age Trial, began their routine mammographic screening as done in the UK, starting at 50. Thus, the “no mammography” became a “mammography group.” In fact, there was zero difference between the two groups from age 50 on, diluting the earlier benefit. No surprise…the screening phase of the study is over by then, and both groups are getting screened. This is a problem in long-term follow-up of all screening studies, in that controls screen anyway, and women in the screening group may stop screening. Oh, yes, another interesting feature of the AGE Trial – there was no measurable overdiagnosis, in spite of the never-ending emphasis and exaggeration of this bias.
In another controversial statement, the authors make the claim that, with improved systemic therapies, we can back off on screening – “with modern adjuvant systemic therapy, a less sensitive screening modality such as CBE (clinical breast exam) may now prove as effective as mammography.” It is true that when that day comes in the future when systemic therapy cures all breast cancers, we will no longer need screening. But it is disastrously premature to state that this day has come already, or even that we are easing into an era where we can back off on screening. The 40,000 women who died of breast cancer this year would not be nearly as enthusiastic about those systemic therapies as these authors seem to be.
I don’t have time or space here to go over the Harvard study that showed, overwhelmingly, that most breast cancer deaths occur in unscreened women. Systemic therapies and screening are complementary, and until that 40,000 number is dramatically reduced, we cannot back off on either screening or systemic therapies. In fact, we need to be doing more in both categories. One does not preclude the other, nor do they need to operate in an inverse proportion. I could easily claim that if we screened all women with breast MRI, we could back away from systemic therapies. And though I could defend that position to a degree, it would be every bit as misleading as the inverse, claiming that systemic therapies allow us to back off on screening.
Since there is very little data from the prospective, randomized trials on the mortality reduction for screening women in their 70s, what would rational thought tell us? If we divide screening into the decades, the benefit of mammography gains momentum over time. It takes fewer and fewer screens to save a life as women age. Screening in the 50s is more effective than the 40s, while screening in the 60s is better than the 50s (by quite a bit). Would any rational person on the planet truly believe that this accelerating benefit would come to an abrupt stop at age 70? In fact, barring co-morbidities, it is likely that the effectiveness is every bit as good, if not better, than for women in their 60s. Yet, we’re told Primum Non Nocere, and to drop mammography at its peak performance level; instead, adopt clinical exam for which there is no evidence anywhere that lives can be saved with exam alone.
A full 25% of eventual breast cancer victims are over the age 70. We’re already being told by some to stop screening under age 50, which is another 25%. So, if the anti-screeners get their way, screening only between 50 and 70, we will be excluding 50% of the eventual breast cancer victims from the potential of early detection. Compare this to the current 5% of eventual victims excluded by virtue of being under age 40 at the time of diagnosis. I submit that the leap from 5% to 50% will be deadly for many women, several thousand a year, whereupon the term Primum Non Nocere takes on a new twist.
It only takes a few pages of the written word to sling mud. In contrast, there are so many misunderstandings, false claims, erroneous misrepresentations of available data, and flat-out errors, that it takes a book to wipe the mud away. That’s why I wrote one – the working title is
Breast Cancer Early Detection: Past, Present, Prologue (McFarland & Company, Inc., Publishers) with release estimated late 2016.