| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Practice of Radiology |
1 Department of Radiology, Rush North Shore Medical Center, 9600 Gross Point Rd., Skokie, IL 60076, and Rush Medical College, Chicago, IL 60612.
Case summaries are based on actual events and lawsuits, although certain
facts have been omitted or modified by the author, who has supplied and
obtained authorization for the reproduction of the radiologic image. All
opinions expressed herein are those of the author and do not necessarily
reflect those of the American Journal of Roentgenology or the
American Roentgen Ray Society.
The Case
A 66-year-old woman who had been complaining of lower back pain for several weeks was referred by her family physician to a community hospital for CT examination of the lower lumbar spine. The hospital-based radiologist who interpreted the study noted in his report that the examination was "limited due to the patient's weight and inability to cooperate" but nevertheless concluded that the diagnosis was "relatively severe facet joint arthropathy involving lower lumbar spine with the most prominent and severe involvement at the L4–L5 level. No evidence of disc herniation or spinal stenosis" (Fig. 1).
|
Little is known about the patient's followup care, but what is known is that 3 years after the CT examination, the patient experienced sudden severe back and abdominal pain and collapsed. She was taken by ambulance to a nearby hospital where physical examination and abdominal CT revealed a ruptured abdominal aortic aneurysm. The patient was taken immediately to the operating room but died during surgery.
Four months after the patient's death, the family of the deceased filed a medical malpractice lawsuit against the radiologist who had interpreted the initial CT study, alleging that the defendant radiologist had been negligent for failing to diagnose an abdominal aortic aneurysm.
Malpractice Issues
A radiology expert witness retained by the plaintiff's attorney testified in her discovery deposition that calcification within the wall of an enlarged abdominal aorta was "clearly visible" on multiple images obtained during the lumbar CT examination and that the defendant radiologist's failure to observe the calcifications and report the presence of an abdominal aortic aneurysm constituted negligence. A radiology expert witness retained by the defense attorney testified in his discovery deposition that the primary purpose of the CT examination was to assess the lumbar vertebrae and that the calcification anterior to the lumbar spine present on the scans was difficult to see. The defense expert stated that an ordinary radiologist could well have failed to note the calcification and that it assumed importance only because it was later learned that the patient had succumbed to a ruptured abdominal aortic aneurysm. The defense expert characterized this type of retro-spective analysis as outcome bias.
The defendant radiologist's attorney consulted two additional radiologists, but both thought that the aortic calcification should have been reported by the defendant radiologist. With the consent of the defendant radiologist, the claims manager of the professional liability insurance company began settlement negotiations and eventually reached agreement with the plaintiff to settle the lawsuit for $700,000.
Discussion
Outcome bias is the tendency for people to attribute blame more readily when the outcome of an event is serious than when the outcome is comparatively minor [2]. This reaction is somewhat related to but is nonetheless distinguishable from hindsight bias, which is the tendency for people with knowledge of the actual outcome of an event to believe falsely that they would have predicted the outcome [3]. To more fully appreciate the influence that outcome bias exerts in the context of malpractice litigation, let us first review the concept of hindsight bias.
In a previous article that centered on the failure to diagnose a mediastinal malignant thymoma on a chest radiograph [4], the phenomenon known as hindsight bias was introduced. During the closing argument at the conclusion of the trial of the missed-thymoma lawsuit, the attorney representing the defendant radiologist had attempted to convince the jury that the missing of the mediastinal thymoma did not constitute negligence because the tumor was barely evident and could well have been interpreted as a normal structure by any reasonable radiologist. The defense attorney had urged the jury to find in favor of the defendant radiologist because
...only by means of hindsight bias, in other words, only by having knowledge of the patient's subsequent clinical course and later radiographs, information not possessed by the defendant radiologist at the time of initial interpretation, could the tumor be diagnosed.
The jury, however, returned a verdict in favor of the plaintiff, awarding $872,000.
In the subsequent appeal to the Washington State Appellate Court, the defense attorney maintained that the trial judge had erred by refusing to give the jury an instruction on hindsight bias that read [5]:
A physician is not to be judged in light of any after-acquired knowledge in relation to the case, and the question of whether or not he exercised reasonable care and skill is to be determined by reference to what is known in relation to the case at the time of treatment or examination, and must be determined by reference to the pertinent facts then in existence of which he knew.
The appellate court rejected the defense's argument, ruling that the jury had been appropriately instructed to evaluate the defendant radiologist's actions according to the "standard of a reasonably prudent radiologist acting in the same or similar circumstances at the time of the defendant's conduct" [5]. To have added the additional wording as requested by the defense attorney would have created a "risk of jury confusion," the court concluded.
The mechanism that leads a radiologist to conclude that he or she would have spotted a radiographic abnormality that the radiologist who initially interpreted the study did not has been called the "wisdom of hindsight" [2]. How hindsight bias influences physicians in making medical decisions has been illustrated by an experiment in which separate groups of practicing physicians were given almost identical hypothetic case histories and were then asked to rank various diagnoses in order of probability. These case histories all described the same nonspecific findings indicative of arthritis, but each differed in that its opening line indicated that the case involved a patient with a specifically named type of arthritis. The physicians' ranking of the diagnostic probabilities adhered substantially to the disease indicated in each opening sentence. In concluding that the results reflected classic hindsight bias, the researchers who had undertaken the experiment explained that physicians, like all persons, try to make sense out of what they know has happened rather than analyze the data independently [6].
How hindsight bias influences radiologists who interpret mammograms was made evident in a study by Elmore et al. [7]. Radiologists were asked to reinterpret the same mammograms they had interpreted months earlier, but on the second occasion, they were given different information regarding symptoms and family history of breast cancer. The differing histories caused 40% of the interpreting radiologists to change their original diagnostic interpretation.
The effect of hindsight bias in a medical legal context was studied in still another experiment [8]. Participants were given case histories in which a hypothetic psychiatric patient told his therapist that he was contemplating injuring a third person. In one scenario, the patient did not injure the person; in a second scenario, the patient caused minor injury to the person; and in the third scenario, the third person was seriously injured. Respondents were then asked to imagine that they were jurors in a malpractice lawsuit and to determine whether the failure of the hypothetic therapist to inform the third party of impending injury constituted negligence. Participants were told that they should base their decision only on whether the actions of the therapist had breached the standard of care and that they should not be influenced by the end result of the case history (i.e., whether the hypothetic patient had actually caused injury to the third person). Twenty-four percent of respondents found the therapist negligent when major injury had been inflicted, whereas only 6% found the therapist negligent when no injury had occurred. The researchers concluded that hindsight bias may contribute to an unfair and inaccurate assessment of negligence.
In summary then, hindsight bias is the tendency to predict the correct result once the final result becomes known. Closely related is the tendency to attribute blame once we know that the outcome has been bad, which brings us to outcome bias.
Outcome Bias
As stated earlier, outcome bias is the relationship between the severity of an outcome and the degree of harshness with which we judge the perpetrators of that outcome. In the context of medical malpractice, the ramification of outcome bias is that if a patient sustains a serious debilitating or fatal outcome as a result of medical care, then the desire to find someone on whom to cast blame may cause us to find negligence in circumstances under which, were we to examine the situation dispassionately, there is none [2]. With this in mind, then, we can consider the question of whether the missing of the aortic aneurysm by the defendant radiologist in the case presented in this article would have been considered negligence if the patient had remained asymptomatic, rather than having died from rupture of the aneurysm. As we shall see, the answer to the question could well be, No.
Brennan et al. [9] reviewed the medical records of 30,000 patients hospitalized in acute care hospitals in New York to determine the overall incidence of adverse events and the percentage of those events that resulted from negligence. Brennan and his research colleagues found that the likelihood that an adverse event was judged to be due to negligence varied directly with the severity of the adverse event. Of those adverse events that led to temporary disability of patients, 22% were judged to be caused by negligence. Of those events that led to permanent total disability, 34% were judged to be caused by negligence, and of those events that led to death, 51% were judged to be caused by negligence. In a later study, Brennan et al. [10] found that not only did the determination of negligence increase with the severity of injury but also the amount of compensation paid to the injured claimant. In fact, the researchers found that it was neither the presence of negligence (as determined by independent objective analysts) nor even the presence of an adverse event that was predictive of payment. The sole factor that was predictive of compensation was the degree and permanence of the disability.
In another study of outcome bias [11], 112 practicing anesthesiologists were asked to judge the appropriateness of care in 21 cases involving adverse anesthetic outcomes. The anesthesiologists were asked to review hypothetic patient case histories that were identical in every respect with the exception that a different but plausible outcome of severity was included in each. The reviewing anesthesiologists were asked to rate independently the care in each case as either appropriate or less than appropriate, on the basis of their personal judgment of reasonable and prudent practice. Given the same set of facts, 67% of reviewing anesthesiologists felt the care was appropriate when the outcome was not severe and temporary, but the figure dropped to 36% when the outcome was more debilitating. Not surprisingly, the researchers concluded that knowledge of outcome exerted not only a major effect on the opinion relative to appropriateness on the part of the reviewing anesthesiologist but on the harshness of the reviewers' judgments as well. These researchers pointed out that these kinds of judgments are widely used in settings such as hospital quality assurance committees, local disciplinary review boards, and malpractice proceedings.
More recently, surgeons at Vanderbilt University [12] reviewed 130 patients with surgical adverse events to determine whether knowledge of the patient's outcome influenced reviewers' judgment as to whether the adverse event resulted from physician negligence. "Ideally, settlement discussions and jury awards should be a function of the magnitude of deviation from the standard of care, not lost productivity, unmet expectations, or courtroom theater," observed these researchers. However, as in previous similar studies, the researchers found that financial liability was not related to the degree or presence of negligence but instead varied directly with the severity of the patient's disability.
One final study [13] showing the influence that outcome bias exerts on both the public's and physicians' attitudes about medical errors is quite noteworthy. Researchers gave 831 physicians and 1,207 members of the public from around the nation the following vignette: A 67-year-old man goes to the hospital for surgery. He has an allergy to antibiotic drugs, which is noted on his medical records. The surgeon does not notice the information about the allergy and orders an antibiotic to be given at the end of the surgery. A hospital nurse gives the patient the antibiotic.
The story has two endings: One subgroup of participants heard a version that involved no harm coming to the patient; these subjects were simply told, "The patient wakes up with a rash, the mistake is noticed, the antibiotic is stopped, and the patient fully recovers." The second subgroup of participants, however, was told, "The patient wakes up with a rash, is gasping for air, the mistake is noticed, and the antibiotic is stopped, but the patient goes into respiratory arrest and dies."
When asked whether the offending surgeon should be sued for malpractice, 4% of physicians and 30% of the public in the first subgroup, (i.e., those who were told that the patient sustained no harm) answered in the affirmative. In the subgroup that was told that the patient went into respiratory arrest and died, the percentage of physicians who felt that the offending surgeon should be sued for malpractice rose 14 times to 55%, whereas the percentage of the public who felt similarly rose 2.3 times to 69%. The percentage of physicians who believed that the offending surgeon's medical license should be suspended climbed from 0% in the subgroup who heard the version in which the patient sustains no serious injury to 87% in the subgroup who heard the version in which the patient dies, whereas the percentage of the public who were in favor of suspension of the surgeon's license increased from 23% to 50% between the two scenarios.
It is probably not surprising that more than two thirds of the public believe that a physician should be sued for malpractice if he or she commits a medical error that leads to a fatality, but it perhaps is surprising that, notwithstanding the medical community's push to reduce malpractice litigation through tort reform [14], more than half of physicians polled who do not believe that a physician should get sued for committing an error that does not cause permanent injury to a patient think that the physician should be sued if the error leads to the death of the patient.
Summary and Risk Management
The terms hindsight bias and outcome bias are similar in that both types of bias lead people, once they know the actual outcome of an event, to believe that they would have accurately predicted the eventual outcome without that knowledge. The two biases differ in that hindsight bias refers only to the tendency for people with knowledge of the actual outcome of an event to believe falsely that they would have predicted the outcome. Outcome bias, on the other hand, is the tendency for people to attribute blame and to do so more readily when the nature of the outcome is serious than they would do if the outcome were comparatively minor. The attribution of blame perhaps satisfies a psychologic need to find an object to punish, for by punishing another we may feel that we annul the wrong and lessen the hurt [2].
Neither hindsight bias nor outcome bias is supposed to exert influence on the determination of medical negligence, whether that determination is being made by a judge, jury, expert witness, attorney, or claims manager of a malpractice insurance company. Indeed, at the conclusion of a medical malpractice trial, jurors are instructed by the judge to consider, first, only whether the conduct of the defendant physician adhered to the standard of care, independent of any harm the patient may have suffered, and then, only if and after they have made that determination, should they consider the nature and extent of patient injury and the amount of compensation to be awarded to the patient. The paradigm that has evolved in American civil law that bifurcates the determination of whether a physician's actions breached the standard of care from the determination of the extent of compensable patient injury generally fails to be implemented in American courtrooms. Hindsight bias, in other words, knowing what additional radiologic and other laboratory findings disclosed and what ultimately happened to the patient, substantially influences jurors' determination of whether the defendant radiologist conformed to the standard of care. Similarly, outcome bias— knowing the extent of injury sustained by the patient—substantially influences jurors' determination of not only whether the defendant radiologist breached the standard of radiologic care but also, if they believe such a breach occurred, the harshness with which the defendant should be dealt.
Even if a patient is seriously injured or dies as a result of medical care, a defendant radiologist should be found to be free of negligence as long as the defendant adhered to the standard of care [15]. It is clear from the studies reviewed in this article, however, that the determination of negligence is indeed influenced by the nature and extent of patient injury. Both hindsight bias and outcome bias tend to assume greater influence in cases in which there has been an adverse event such as a missed radiologic diagnosis of an aortic aneurysm or perhaps a missed breast cancer on a mammogram, especially when expectations are high regarding professional performance [8]. Radiologists are particularly vulnerable in this regard because it is difficult to convince a jury that a radiologist who is well trained (and well compensated) to discover every abnormality on a radiologic study should be excused for failing to do so [15]. These biases are exacerbated by testimony of the expert radiology witnesses retained by the plaintiff and the defendant respectively, who themselves are influenced by these biases.
Outcome bias, like hindsight bias, affects all human judgment, and it probably can never be eliminated from litigation and judicial proceedings. Acknowledging its existence, and recognizing its influence on expert witnesses and jurors alike, might mitigate its pejorative effect.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Berlin Radiologic Errors and Malpractice: A Blurry Distinction Am. J. Roentgenol., September 1, 2007; 189(3): 517 - 522. [Full Text] [PDF]
|
|
|
|
 |
|
 E. Crosby Medical malpractice and anesthesiology: literature review and role of the expert witness: [Fautes medicales et anesthesiologie : revue de la litterature et role du temoin expert] Can J Anesth, March 1, 2007; 54(3): 227 - 241. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
