| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
1 From the Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Ambulatory Care Bldg, Suite 219, 15 Parkman St, Boston, MA 02114. Address correspondence to the author.
Index terms: Breast neoplasms, radiography, 08.111 • Breast radiography, quality assurance • Editorials • Medicolegal problems
There have been recent efforts to highlight the cost of medical malpractice and to try to find ways to reduce these costs (1). However, it is not just the cost of malpractice insurance that is a problem. What I believe is being overlooked is the unjustified psychologic damage that is being inflicted on some physicians, as well as the failure of the legal system to recognize science, with particular emphasis on the detection of breast cancers with mammography. Screening mammograms are saving lives, but the number of radiologists who are willing to read mammograms is decreasing because of their fear of being sued.
The only accepted way to demonstrate that a test for screening is efficacious is through the use of a randomized controlled trial (RCT) in which a large population is randomly divided into two groups. If the process of division is blinded (nothing is known about the participants) and truly random and the number of participants is sufficiently large, then every individual in one group will have a "twin" in the other group. In the absence of screening, women in both groups would be destined to die at the same time from the same type of cancer (2). If the first group is offered the screening test and the other group is not, and, as the groups are followed over time, there are fewer breast cancer deaths in the screened group than in the control group such that the differences in death rates are statistically significant, then the efficacy of the screening test is proved.
RCTs are the only way, at this time, to eliminate other biases that can produce the appearance of a benefit without providing a true decrease in deaths. Lead-time bias (detecting the cancer earlier but having no effect on the ultimate date of death), length bias sampling (slower growing, less lethal cancers are more likely to be detected with periodic screening), selection bias (those who participate in screening tend to be healthier than the general population), and overdiagnosis (pseudodisease) bias (screening appears to save lives by allowing detection of disease that would have never been lethal if left undetected) can make a test seem efficacious when it actually provides no benefit. Given the expense of screening and the fact that it involves healthy individuals, some of whom will have potentially harmful interventions that prove to be performed for benign reasons (they would not have had to undergo these tests had they not been screened), as well as the anxieties and inconveniences to which a large population is subjected, such rigorous trials are needed before the test can be put into routine clinical practice.
Results of RCTs have demonstrated that mammographic screening can decrease the death rate due to breast cancer (3–5). The ultimate demonstration of the effectiveness of a cancer screening test, however, is the application of the test in the general population with a concomitant decrease in deaths from the cancer. Such a demonstration has been made for breast cancer screening. In an important article published by Tabar and colleagues (6), the authors reviewed the death rate from breast cancer over a 29-year period in two counties in Sweden—Ostergotland and Kopparberg (now Dalarna). These were the two counties that participated in an RCT of mammographic screening for breast cancer between 1978 and 1987. The authors evaluated the breast cancer death rates during the 9 years preceding the trial (1968–1977), when screening was not offered to the women in the two counties. They did the same for the 9 years during which the two-county RCT was under way and part of the population was offered mammographic screening (the others were unscreened control subjects). They compared these rates with those from the most recent period (1988–1996), during which all women in the two counties aged 40–69 years have been offered service screening. They found that the breast cancer death rate during the period when part of the population was being screened was lower than the rate prior to screening and that the death rate was even lower after the entire population of the two counties was offered screening. For women who actually took part in screening, there was a greater than 60% decrease in breast cancer deaths. Of interest and importance is the fact that there was no change in the death rate among women who were not screened during the total period of time, suggesting that treatment, which was available to all women during the period of review, did not account for the observed changes.
The decrease in mortality rates in these two Swedish counties follows the well-known "dose response curve"—the more women screened, the greater the mortality reduction—which further supports the efficacy of mammographic screening. The Swedish review provided a clear demonstration that mammographic screening can dramatically reduce the death rate from breast cancers in generalized service screening and that the benefit is not just confined to RCTs. This study was followed by a more comprehensive study (7) of approximately 30% of the Swedish population, which provided even more evidence that the introduction of mammographic screening to the general population results in mortality reduction from breast cancer. This reduction is directly linked to the onset of mammographic screening.
Additional support for population screening is evident in breast cancer mortality rates in the United States. The death rate from breast cancer remained essentially unchanged for over 50 years, until it suddenly began to decrease in 1989–1990. Peto et al (8) attributed the decline to better therapies, but they did not take mammographic screening into account in their analysis. A review of the annual incidence rates of breast cancer in the United States provides more direct evidence that much, if not most, of the decrease is a result of screening. Data from the Connecticut tumor registry and the Surveillance Epidemiology and End Results, or SEER, program (9) show that the annual incidence (new cancers diagnosed each year) of invasive breast cancer in the United States has increased steadily since 1940. This increase has never been satisfactorily explained but has been attributed to changes in lifestyle and undefined environmental factors. The incidence increased annually from approximately 60 new cases of invasive breast cancer per 100,000 women in 1940 to more than 90 new cases per 100,000 women by 1982.
In 1983–1984, there was a sudden and dramatic increase in the annual incidence of breast cancer over and above that which had been experienced in the preceding years. This more rapid increase continued, reaching as high as 115 new cases per 100,000 women per year. Initially, there was great concern that something had stimulated a true increase in the number of breast cancers in the United States. It became apparent, however, that most of this increase was due to the earlier detection of cancers as a result of the onset of a much higher rate of mammographic screening (10). Physicians and women had recognized the mortality reductions that had occurred in the RCTs from screening, and women began to participate in screening in increasing numbers. The sudden increase in breast cancer screening resulted in breast cancers being detected earlier, adding to the baseline incidence that had been present without screening. Cancers that would not have been detected for 1 or more years in the absence of screening were being detected earlier and were adding to the baseline incidence. The rate of screening was sufficiently high to be reflected in national statistics.
The abrupt increase in cancers being detected, especially ductal carcinoma in situ (a lesion that can be virtually only detected with mammography), provides an objective marker for the onset of widespread mammographic screening that began in 1983–1984 in the United States. Since periodic screening is more likely to depict cancers with moderate growth rates, RCT findings have shown that mortality reduction is likely to appear 5–7 years after the onset of screening. This is precisely the time that mortality rates began to decrease suddenly after screening began in the United States. For the first time in over 50 years, the death rate from breast cancer began a rapid decrease that started around 1989–1990. This downward trend has continued for the past decade, mirroring the increased rate of detection in the mid-1980s and early 1990s. Although I believe that some of the reductions in death are a result of better treatment, the abrupt change in incidence that clearly marked the start of screening in the United States coupled with the equally abrupt decrease in deaths at the appropriate time thereafter would have to be a major coincidence. I believe that the evidence strongly suggests that mammographic screening is the major reason why the rate of death from breast cancer is decreasing (11).
Good News/Bad News
Although mammographic screening is far from perfect, the benefits that were shown objectively in RCTs have now been demonstrated for populations outside of trials. Unfortunately, just as the benefits of mammographic screening have produced a major decrease in deaths from these cancers, I believe we are headed for a screening crisis. The number of radiologists who are willing to interpret mammographic images appears to be decreasing. There has been a marked decrease in the number of radiologists who are applying for breast imaging fellowships. In a recent survey (12), it was found that the major reason why radiologists were not going into the field was that they were not interested. However, the second most important reason was the stress involved in interpreting mammograms and the fear of being sued. There has been a dramatic increase in the number of lawsuits in the United States for "failure to diagnose" breast cancer (13). Radiologists are well aware that this has become one of the leading causes of medical malpractice claims. Many radiologists are asking themselves why they should expose themselves to the increased risk of a lawsuit.
One of the reasons for these suits is that the expectations of women have been raised beyond the reality of the situation. In the attempt to convince women to participate in mammography screening, the potential of the technique was "oversold." Although I do not believe it was ever stated, I believe that many of the advertisements and articles led women to believe that screening was a guaranteed lifesaver. Despite repeated warnings by radiologists over the years that mammography does not depict all cancers and does not allow detection of all cancers early enough to save all lives, the damage was done and perfection was expected. Consequently, radiologists are being held to a standard that is impossible to achieve. In my experience, in reviewing numerous malpractice cases, the most common scenario is a woman with a diagnosis of breast cancer who wants to sue or is suing by claiming that the cancer was missed on an earlier mammogram. I have been asked by the plaintiff to review prior mammograms to determine whether or not the cancer was visible previously, and I have been asked by the defense to do the same, since a lawsuit has already been initiated because the plaintiff’s expert testifies that the cancer was visible on a prior mammogram. The failure to have detected it is termed a breach of the "standard of care," and the radiologist is accused of negligence.
The fact is, it is not uncommon for a cancer to be visible in retrospect. A few lesions are even obvious. What is being missed in this area of litigation is the fact that there are ample data to show that all radiologists, no matter how skilled, will periodically fail to see a clinically important abnormality that is visible in retrospect. This is true among radiologists looking for lung nodules (14), those looking for polyps on barium enema images (15), and those looking for breast cancer on mammograms (16). Human perception is complex. We are often unable to "see" something until we are told what we are looking at. This is graphically demonstrated in an illustration provided by Kundel and Nodine (17) in their 1983 article on perception. When the observer’s eye movements are tracked while looking at an unfamiliar picture, the movements are fairly unstructured and the observer cannot deduce what the picture is showing. Once the observer is told what is in the picture and it is pointed out, then subsequent eye movements follow the expected pattern, and the observer is able to "see" what is in the picture. Once a cancer has been pointed out to an observer, it becomes much easier to "see" it. Perceiving it in the first place is a challenge.
Given the fact that all radiologists, no matter how skilled or dedicated, cannot avoid periodically missing a clinically important lesion, it is unclear how this could be a breach of the standard of care. Since this is apparently an immutable fact of the human condition, how can it be considered negligence? Although I am told that many of the lawsuits are being won by radiologists, some are lost, and others are settled. A loss or settlement places the radiologist on a public list, and he or she must explain the suit and its result with every change in job and for licensure. The legal costs are substantial, but what is not quantifiable is the enormous psychologic trauma that many who are sued go through. In my experience, it is not so much the concern over the possible monetary loss that might be incurred, but it is the self-doubt, embarrassment, and loss of self-esteem that accompanies the accusation that one was "negligent" that can plague radiologists who are accused of malpractice. One radiologist whom I know, who was exonerated at trial, lost a substantial amount of weight just waiting for the trial and required psychotherapy. For a lawyer, lawsuits are a way of life. For a physician, it strikes at the core of our being.
In my experience, after discussing this with physicians who have been sued, the suggestion that one was negligent in the care of a patient is devastating. Care of the patient is the meaning of being a physician. Not only is the accusation very traumatic, but the legal process may go on for years before the physician is exonerated. The individual who is sued must live with the uncertainty and with little support. We are warned to not discuss the case with anyone except our lawyers. If we discuss the case with someone else, to try to reassure ourselves that we were not negligent, that person can be subpoenaed. Consultation between physicians is a way of life and the way we seek "the truth" in caring for our patients. Being denied this basic interaction with colleagues at a time when we are under substantial stress only adds to that stress.
Anyone Can Be an Expert
I believe that the epidemic of lawsuits is sustained by the fact that virtually anyone can be offered as an "expert witness." Physicians are willing to testify that another physician should not have overlooked a finding and was guilty of negligence. Since an expert witness can charge virtually anything he or she wishes, there is great financial incentive to be an expert witness. Experts can be found on the Internet. Do all experts have true expertise? I am aware of cases in which the expert witness was primarily involved not in the practice of breast imaging but in another imaging subspecialty, or the expert was practicing in a nonimaging medical specialty. Moreover, since the expert witness is merely providing an opinion, there is additional "flexibility" for the expert in that he or she may not be held strictly to the rules of perjury. It is up to the defendant’s lawyer to try to discredit the expert in the eyes of the jury.
We are all aware that mammograms from the past are kept on file so that new images can be compared with the prior images to see if there has been a change that might indicate cancer development. These prior images, which are retained for the benefit of the patient, are available for retrospective review. Such prior images can be used to sue radiologists whether or not an abnormality is truly visible and suggestive of malignancy on the prior images.
The Standard of Care in Our Legal System Is a Myth
The law requires that physicians must function at or above the standard of care, which is defined as what an average physician would have done under the same circumstances at the time that the negligence was alleged to have occurred. The problem is that there is no way of ever knowing what the standard of care actually is at the time the care is being provided. The standard of care is decided in retrospect by a jury that makes its decision on the basis of the testimony of experts for the plaintiff and those for the defense. Compounding the problem is that the standard for a certain situation, which is decided by one jury for the circumstances defined in one case, is not binding in any other case, even if the circumstances are identical. Another jury could hear the same testimony and arrive at the opposite conclusion. One jury might decide that, in 1995, the standard of care was to image all palpable masses with ultrasonography (US), while another could decide that that was not the standard. Whichever jury you have determines whether or not you are guilty of negligence.
The only precedents that are set in the legal system have to do with matters of law, not medicine. There will be sanctimonious arguments to the contrary, but I strongly believe that physicians are being held to a standard of care that does not exist. How can it exist if two juries can arrive at opposite definitions on the basis of the same data? Physicians are being held to a standard of care that is determined, in retrospect, at a single instant, by the opinion of nonmedical jury members who, in part, base their decision on how they feel toward the plaintiff and the defendant and how they feel about the experts who, themselves, are financially invested in the proceeding. Lawyers will argue that the circumstances in each case are unique, and therefore, each case must be decided individually. This, of course, is often nontenable. Either US was the standard of care for a palpable abnormality in 1995 or it was not, yet in the system as it exists today, a jury in one case might decide it was the standard of care, while a jury in another case could decide it was not. In my opinion, in the courtroom the standard of care is, thus, a myth.
The Eyes Have It
Radiologists are being sued for not having seen a cancer that may or may not be visible in retrospect on previous images. How do you determine a standard of care for failure to perceive something that is visible in retrospect? Only the images from the examination in question are presented to the jury, with the area of concern highlighted to emphasize its appearance. The images are not provided to the jury along with those of 40 other patients to see if they would find the lesion the way it was originally presented to the radiologist. As noted earlier, results in study after study have shown that even experts in a field fail to perceive abnormalities that are visible in retrospect. Investigators have shown that in as many as 70% of cases, a finding that could potentially represent breast cancer is visible, in retrospect, on a preceding image (18).
In an informal survey that I have undertaken around the country, I have been unable to find a single experienced radiologist who has not, at one time or another, overlooked a cancer that was visible in retrospect. Given this reality, it would seem that either all radiologists are negligent or that the failure to periodically perceive an abnormality is expected of the "average" radiologist and should not be considered negligence. I remain baffled by the rationalization of radiologists who freely admit that they have overlooked cancers that are visible in retrospect but believe that it is ethical to turn around and testify that another radiologist was negligent for doing the same thing.
The Vagaries of Human Perception
Attorneys who make a living from lawsuits should also be concerned about breast cancer and the looming mammography crisis. At a recent conference attended primarily by attorneys, I pointed out my concerns and the need for reform. I reviewed my opinion that our "adversarial" legal system was not designed to get at the truth but "to see who wins." I complained that doctors are being held to a standard that is unpredictable, and, in some cases, impossible to meet. When I suggested that radiologists should not be considered negligent for not seeing a cancer, an indignant reply was that failing to see a stop sign is not an excuse for driving through it. Another attorney at another meeting made the same comment to me, making me wonder if this analogy is being taught in law schools. It is clearly nonsense. There are laws that pertain to stop signs. There are no laws that pertain to perceiving a breast cancer. Stop signs all look the same, and there is no dispute as to their appearance. The same is not true for breast cancers. Stop signs are always found at intersections and are always on the right side of the road. Breast cancers occur in various locations and with different appearances. Stop signs are also very common. They do not occur randomly, once in every 200–1,000 intersections. I suspect that if stop signs occurred randomly, we would fail to see them from time to time. Certainly, if the stop sign was green and hidden by trees, failing to stop would be defensible.
The Legal System Is Learning About Science
The legal system is beginning to recognize science, but not willingly, I believe. I was once asked at a deposition whether a lesion that was seen on only the mediolateral oblique view in a patient with two lesions (a cancer and a fibroadenoma) was the cancer that was later found laterally in the left breast. My reply, based on the scientific facts, was that "no one in the world could tell you that it was the lateral lesion." This precipitated a storm of objections, and I was told that I could only say "in my opinion," not "no one in the world." The court would not recognize that it is impossible to derive three-dimensional information from this single two-dimensional projection.
Is There a Solution?
The legal system needs to take into account universal problems with the human visual system. When you are searching for your keys and you do not see them lying "right in front of your nose" in plain view on the table, this is not negligence. It has nothing to do with inattentiveness, a lack of expertise, or a lapse in concentration. It is a fact of human perception. We all fail to see things periodically, even when we are trained to recognize them and we are trying very hard to see them. I was recently sent a graphic example of the vagaries of human perception. Unfortunately, I do not know the source of the following example, but it is very striking.
Count the number of F’s in the following text (this is not a trick question): Finished files are the result of years of scientific study combined with the experience of years. Most people (who have not seen this before) will count three F’s, when, in fact, there are six. You are an expert in F’s. You were told that there are F’s in the text. You even suspected a trick, so you were extra diligent at looking, yet most everyone will only see three F’s. You have been provided with far more information than a radiologist has when looking at a mammogram. In a screening population of 1,000 women, there are likely only four to 10 who have a breast cancer. Each woman has a minimum of four images (two of each breast) for a total of 4,000 images for the radiologist to review, and breast cancers have a wide range of appearances. You had a far simpler task, and many of you still missed three of the six F’s that could not be more "obvious." I would submit that the failure to see a cancer that is visible in retrospect is certainly unfortunate but not negligent, since there is no way that we know of to prevent this from happening periodically.
This is not to say that there are not cases of true negligence. I once reviewed a case where an irregularly shaped spiculated mass at mammography was examined by using an abdominal US probe. The images that were recorded showed an irregularly shaped lesion with no enhanced through-transmission of sound and with internal echoes. Despite all of these highly suspicious findings, the lesion was dismissed as a benign simple cyst. Nevertheless, given the fact that every radiologist who has any long-term experience has failed to perceive something of importance that is visible in retrospect, the failure to perceive an abnormality that is visible in retrospect should not be a cause for litigation. In fact, since it happens to every radiologist periodically, it is well within the standard of care.
Another problem that needs some attention is the definition of an "expert." The fact that anyone can qualify needs to be changed. Experts should have to be truly expert. If a physician is to testify in a trial where mammography figures heavily in the case, then that physician must be an expert in mammographic interpretation. How is expertise defined? To borrow from Justice Potter Stewart—"You know it when you see it." Judges should be able to determine when someone is really expert or not and should not rely on the attorneys to have to discredit the expert in the eyes of the jury. I believe that definitions of expertise could be written that would be satisfactory to the medical as well as legal professions.
Something needs to be done, and quickly. Mammographic screening can clearly save lives. Consequently, lives are, quite literally, at stake. Berlin (19) has suggested that a way to reduce the number of lawsuits is to educate the public as to the fallibility of mammography. I agree completely, but I suspect that this will be insufficient. We have been educating women and their physicians for years, and the number of lawsuits continues to increase. The legal system needs to acknowledge reality. There is a precedent. When confronted with the huge amount of litigation involved with silicone gel implants, Sam C. Pointer, Jr, coordinating judge for the Federal Breast Implant Multidistrict Litigation, convened a review by a panel of experts to determine if there was any credible scientific evidence that silicone gel implants were causing systemic disease. Their objective review showed that implants were not the cause of the systemic illnesses, as was being claimed (20). It is my understanding that this review is admissible in all implant cases and has resulted in a major decrease in litigation. I would urge that a similar review be undertaken for the detection of breast cancers on mammograms. Radiologists could be monitored for the number of cancers missed. The goal is not to absolve radiologists of any responsibility but to acknowledge the facts.
Some protection based on facts is needed. There has been no enthusiasm for this in the past. Hopefully, the situation will change. Reform is needed before women are informed that they are on the verge of losing a life-saving test. Once this becomes public, the situation will change. If reform does not occur, then access to breast cancer screening will diminish and may even become unavailable. Regardless, radiologists are greatly concerned, and if they refuse to interpret mammograms, the benefit from mammographic screening will be lost.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  C. D. Furtado, D. A. Aguirre, C. B. Sirlin, D. Dang, S. K. Stamato, P. Lee, F. Sani, M. A. Brown, D. L. Levin, and G. Casola Whole-Body CT Screening: Spectrum of Findings and Recommendations in 1192 Patients Radiology, November 1, 2005; 237(2): 385 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Elmore, S. H. Taplin, W. E. Barlow, G. R. Cutter, C. J. D'Orsi, R. E. Hendrick, L. A. Abraham, J. S. Fosse, and P. A. Carney Does Litigation Influence Medical Practice? The Influence of Community Radiologists' Medical Malpractice Perceptions and Experience on Screening Mammography Radiology, July 1, 2005; 236(1): 37 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Mulshine Clinical Issues in the Management of Early Lung Cancer Clin. Cancer Res., July 1, 2005; 11(13): 4993s - 4998s. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Berlin Mammography Screening Can Survive Malpractice ... If Radiologists Take Center Stage and Assume the Role of Educator Radiology, December 1, 2004; 233(3): 641 - 644. [Full Text] [PDF]
|
|
|
|
|
|
D. B. Kopans Be Careful to Not Willingly Suspend Disbelief Radiology, December 1, 2004; 233(3): 645 - 647. [Full Text] [PDF]
|
|
Read all eLetters
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
