Abstract

Context.—Fear that damaging information from autopsy may be introduced as evidence in lawsuits alleging medical malpractice is often cited as one factor contributing to the decline in autopsy rates.

Objective.—To determine how autopsy information influences the outcome of medical malpractice litigation.

Design.—We studied state court records in 99 cases of medical malpractice adjudicated from 1970 to the present to assess the role of information from autopsies in the outcomes.

Results.—The 3 largest groups defined by cause of death at autopsy were acute pulmonary embolism, acute cardiovascular disease, and drug overdose/interaction. Findings for defendant physicians outnumbered medical negligence in the original trial proceedings by a 3:1 margin. The appellate courts affirmed 51 acquittals and 19 findings of negligence, and reversed the original trial court decision in 29 cases for technical reasons. We found no significant relationship between accuracy of clinical diagnosis (using the autopsy standard) and outcome of a suit charging medical negligence. Even when a major discrepancy existed between the autopsy diagnosis and the clinical diagnosis, and the unrecognized condition was deemed treatable, defendant physicians were usually exonerated. Moreover, major diagnostic discrepancies were relatively uncommon in suits in which a physician was found to be negligent. Conversely, in about 20% of cases, autopsy findings were helpful to defendant physicians.

Conclusions.—Our study confirms that a finding of medical negligence is based on standard-of-care issues rather than accuracy of clinical diagnosis. Autopsy findings may appear to be neutral or favorable to either the plaintiff or the defendant, but are typically not the crux of a successful legal argument for either side in a malpractice action. We conclude that fear of autopsy findings has no rational basis and is an important obstacle to uninhibited outcomes analysis.

The autopsy is of critical importance in medical education, counseling, and as a source of tissue for studies of pathogenesis of disease and research.1–3 In different medical eras, the frequency of major discrepancies between clinical and anatomic diagnoses based on autopsy findings has not changed.4–9 Despite this fact, autopsy rates at both community and teaching hospitals in the United States have dropped substantially. Of the many contributors to this decline, one reason that is often cited is fear that damaging information from autopsies will be introduced as evidence in lawsuits in which malpractice is alleged to have occurred.7,10,11 Evidence in support of this assertion is extremely difficult to find. To investigate the matter, we undertook a study of readily available public court records of malpractice case proceedings to assess the role of information from autopsies in the adjudication of these cases.

METHODS

We studied state appellate court decisions to determine how autopsy results were used by the litigants. The written opinions of an appellate court typically contain a summary of the legal history, medical history, the basis for the claim, the legal issues on which the appeal is founded, a discussion of how legal precedents apply to the particular case, and conclude with a reasoned decision by the court. Westlaw is a public database that provides the entire text of court decisions at the appellate level throughout the United States. We limited the scope of our survey to the last few decades, 1970 to present. The Westlaw database, ALLSTATES, searches all cases in the United States, no matter the level, using the Natural Language approach. We used the following key words in our search: physician, negligence, (wrongful) death, malpractice, and autopsy. Variations of these key words provided additional cases. In a series of searches, 400 cases were retrieved. Each case was screened to determine its pertinence to the study. Approximately 300 cases that did not deal with medical negligence/malpractice claims were discarded. All product liability and criminal charges were excluded. All other cases were printed for detailed review. One hundred seventeen cases were reviewed at least twice, and in most cases 3 times, by members of the Autopsy Committee of the College of American Pathologists. The percentage of total cases reviewed from the 3 decades surveyed was 1970–1979, 15%; 1980–1989, 19%; and 1990–1999, 66%. Sixteen cases were identified during the review phase in which autopsy was mentioned in the transcript solely because an autopsy had not been performed. Two cases with autopsies were excluded because nonmedical issues were being adjudicated in a workman's compensation case and a medical examiner's case with a diagnosis of strangulation. Ninety-nine cases constituted the final study group. The senior author (K.B.) reviewed all cases. One of the authors (C.I.) performed the original search and screening, and then examined the transcripts to provide help in interpretation of complex legal issues.

The following factual information was extracted from each appellant court transcript: clinical setting, clinical diagnosis, autopsy cause of death, existence of major discrepancy between clinical and autopsy diagnosis, and original outcome (negligence or acquittal, affirmation or reversal on appeal, and reasons for reversal). The autopsy findings in each case were classified as follows: clinical diagnosis confirmed; cause of death identified (or not); major discrepancy between clinical and autopsy diagnosis, in which case it was presumed that autopsy findings may have contributed to the decision to sue; and autopsy findings were important to successful defense. The case reviewers, all experienced autopsy pathologists, were asked to make a judgment on whether they thought the autopsy findings favored the plaintiff (P), the defendant (D), or were neutral (N). Autopsy findings were classified as favoring the plaintiff if there was a major discrepancy between the clinical and autopsy diagnosis and if the unrecognized condition was treatable with a reasonable chance of success. Autopsy findings were viewed as neutral if the clinical diagnosis was confirmed or the findings were construed as being of equal value to both sides of a case. If the clinical diagnosis was confirmed and the death was an expected outcome for that condition, the autopsy findings were considered favorable for the defense. The final rating for each case was based on 2 independent reviews, as follows: P-P, autopsy strongly favored the plaintiff; P-N, autopsy slightly favored the plaintiff; N-N or P-D, autopsy findings were neutral; D-N, autopsy slightly favored the defense; and D-D, autopsy strongly favored the defense.

RESULTS

The classification of autopsy results in the 99 study cases appears in Table 1. Autopsy results unequivocally confirmed the clinical diagnosis in 27 of the 99 cases. There were 54 cases with major discrepancies between the clinical diagnosis and autopsy findings, of which 40 were diseases that if known were deemed potentially treatable by the case reviewers. No cause of death was identified at autopsy in 8 cases. The autopsy conclusions were unclear in 10 cases because of differing interpretation by expert witnesses or autopsy reporting irregularities, including 3 private autopsies that were excluded by the court for violation of discovery procedure. A total of 82 cases could be evaluated from the dual standpoints of nature of autopsy findings versus outcome of the litigation and whether the autopsy findings with respect to the cause of death favored the plaintiff, favored the defense, or were neutral.

Table 1. 

The Relationship of Autopsy Findings to Outcome in 99 Cases of Alleged Medical Malpractice

The Relationship of Autopsy Findings to Outcome in 99 Cases of Alleged Medical Malpractice
The Relationship of Autopsy Findings to Outcome in 99 Cases of Alleged Medical Malpractice

The appellate court affirmed 51 acquittals and 19 findings of medical negligence by defendant physicians (Table 1). The original trial court decision was reversed in 29 cases and remanded for retrial. In decisions that were reversed, physician defendants had been acquitted by summary judgment in the first trial in 14 cases, acquitted by jury trial in 10 cases, and found to be medically negligent by jury trial in 4 cases. One reversal decision dealt only with a technical legal issue and contained little medical information. The reasons for reversal of trial court judgments varied, but tended to fall into the following categories. Summary judgments for defendants were reversed for a variety of procedural errors in 8 cases. Most were cases that had been dismissed at the pretrial phase; the appeals court disagreed, holding that fact issues remained to be determined by trial before a fair judgment could be made. The technical errors that resulted in reversal of trial judgments included multiple examples in which expert testimony had been disallowed, discovery procedure rules had been violated, or instruction to the jury was deemed faulty on complex issues such as proximate cause or multiple causation. Two convictions were reversed and remanded because the statute of limitations law had been improperly applied. Eighteen cases reviewed by the appeals courts, mostly reversals, involved questions about autopsy performance or reporting, or admissibility of autopsy information or death certification, all of which will be discussed in a separate article.

The relationship of autopsy findings to the legal outcome of the malpractice suits is presented in Table 2 in a manner that permits comparisons between alleged medical malpractice in several different clinical practice settings. The table also demonstrates the outcomes in the 3 largest groups defined by autopsy diagnosis, namely, fatal acute pulmonary embolism, fatal acute cardiovascular disease, and fatal drug overdose/interaction. Fifty percent (49/99) of the total number of study cases and 69% (34/49) of the cases with major diagnostic discrepancies are represented in 4 of the 5 clinical settings tabulated: postoperative deaths, deaths shortly after an emergency room visit, postpartum maternal deaths, and perinatal deaths. As expected, surgery, emergency medicine, and obstetrics were the medical specialties of the physicians named in these suits. The resumes of the original trial proceedings typically lacked information about credentials, unless that had been an issue raised by the attorneys.

Table 2. 

Influence of Autopsy on Outcome of Medical Malpractice Suits in Major Clinical Subsets and by Cause of Death

Influence of Autopsy on Outcome of Medical Malpractice Suits in Major Clinical Subsets and by Cause of Death
Influence of Autopsy on Outcome of Medical Malpractice Suits in Major Clinical Subsets and by Cause of Death

Findings for the defendant outnumbered findings of medical negligence in the original trial proceedings by a large margin, 75 versus 23 (approximately 3:1). Physicians accused of malpractice were acquitted in 8 of 13 cases in which the autopsy confirmed the clinical diagnosis, compared to acquittals in 32 of 36 cases in which a major condition, deemed by the original prosector, the plaintiff, and case reviewers to be the cause of death, had not been recognized or suspected before death. Thus, there seemed to be no relationship between accuracy of clinical diagnosis (using autopsy as the ultimate standard) and the outcome of a lawsuit alleging medical malpractice. The significance of these statements is tempered by the fact that 14 of the acquittals and 1 conviction were reversed and remanded for retrial, the results of which are unknown to us.

Characteristics of Cases in Which Defendants Were Found to Be Negligent, Affirmed on Appeal

Defendants were found to be negligent and decisions were confirmed on appeal in 19 cases (see Table 3). Seventeen cases were distributed among 6 medical specialties as follows: surgery, 4; emergency medicine, 4; obstetrics, 3; general medicine, 3; anesthesia, 2; and psychiatry, 1. Two defendants were chiropractors. In 15 of 19 cases, autopsy findings may have played a role in the decision to sue, in conjunction with other factors that are unknown to us. In 10 of 19 cases, a major discrepancy existed between the clinical diagnosis and the cause of death, as determined at autopsy. An anatomic cause of death was found in 17 of 19 cases.

Table 3. 

Critical Information in 22 Medical Malpractice Cases With Jury Verdict of Medical Negligence*

Critical Information in 22 Medical Malpractice Cases With Jury Verdict of Medical Negligence*
Critical Information in 22 Medical Malpractice Cases With Jury Verdict of Medical Negligence*

Clarifying data for 19 cases in which the original trial finding of medical negligence was affirmed by the appeals court are listed in Table 3, along with 3 other cases in which a conviction of medical negligence was reversed and remanded for retrial. In 4 cases, an emergency room physician was the last doctor to see the patient. There were 4 postoperative deaths, 3 postpartum deaths, 2 deaths related to anesthesia, 3 drug-related deaths, 1 spinal cord injury caused by neck manipulation, and 1 case in which clinical findings were erroneously interpreted as untreatable fatal disease. Two deaths, including 1 postoperative death, were caused by unrecognized pulmonary emboli. A finding of negligence in each of these cases was based on standard-of-care issues rather than existence of diagnostic discrepancy.

Acute Pulmonary Embolism

Eight trial records concerned patients in whom the cause of death based on autopsy was massive pulmonary embolism, multiple pulmonary emboli, or multiple pulmonary emboli of varying age. All of these patients were in basic good health and had at least 1 well known risk factor for pulmonary embolism. Four were hospitalized for serious trauma, 1 had recent cesarean section and another a recent hysterectomy, 1 was on bed rest for chronic low back pain, and 1 had spontaneous onset of bilateral calf tenderness and inflammation. Two patients had multiple risk factors for deep vein thrombosis. The duration of restricted mobility or absolute bed rest in the 8 patients prior to death ranged from 5 days to 7 weeks, with an average of 22 days. Leg deep vein thrombosis had been diagnosed clinically in 2 patients, and in retrospect, probably was present in 3 others based on signs and symptoms mentioned in the court record. Acute respiratory symptoms, chest pain, or unexplained syncope were noted in 6 of the 8 patients 24 to 48 hours before death. In 3 of these 8 patients, pulmonary embolism had been diagnosed antemortem, but in 2 of these patients, diagnosis was made 2 hours or less before death and no treatment had been started before death supervened. In one patient diagnosed correctly, aggressive treatment was underway. In another case, failure of a patient in a leg cast to visit his physician as scheduled over a 5-week period may have contributed to delayed diagnosis of deep vein thrombosis. Communication difficulties with the physician's office were alleged by the plaintiff to have contributed to the delay.

Acute Cardiovascular Disease

The study group included 10 apparently previously well adult patients with acute illness followed by death within 24 to 48 hours. Most of these patients were last seen by a physician in the emergency department of a hospital; a few had been seen in private physician offices. In only 1 case was suspicion of cardiac disease raised by the physician. Findings at the last examination of each of these patients before death could reasonably be ascribed to the disorder found at autopsy to be the cause of death. The clinical diagnoses were viral syndrome (n = 2), pneumonia (n = 2), chest myalgia (n = 2), pleuritis (n = 1), indigestion (n = 1), brief seizure with tachypnea (n = 1), and pain due to old pulmonary fibrosis (n = 1). The autopsy causes of death in these patients were acute or subacute myocardial infarction due to coronary atherosclerosis (n = 7), myocarditis (n = 2), and aortic dissection due to Marfan syndrome (n = 1). The outcomes of the original malpractice trials in these 10 cases were 7 acquittals and 3 findings of medical negligence. The appellate courts reversed the original decisions because of technical errors in 4 acquittals and in 1 finding of negligence, mandating new trials.

Death Due to Drug Reaction, Overdose, or Abuse

In 9 study patients, death occurred as a result of administration of 1 or more drugs, usually prescription drugs, alone or in combination with over-the-counter medications, ethanol, illicit drugs, other forms of substance abuse, and in 1 case, intramuscular administration of a non–FDA-approved alternate therapy for cancer. In this group, 8 physicians and 1 chiropractor were sued for unexpected death of patients under their care, in whom both the mechanism and manner of death were initially unknown or poorly understood. Many of the autopsies in this group were performed under the authority of the local medical examiner. Postmortem drug levels and toxicology screenings were critical in classifying and explaining most of these deaths. In 1 fatality following electroconvulsive therapy, aminophylline was suspected to be involved. In 3 cases, suicide was strongly suspected.

Voluntary supplementation, without medical advice, of prescription drugs by other medication available in the home was implicated in several deaths. In 3 cases involving death after ethanol abuse (n = 2) and paint sniffing (n = 1), the plaintiff argued unsuccessfully that the physician who had prescribed medication shortly before death was responsible for the outcome. Medical negligence was the finding in 3 of the 9 cases in this group. A psychiatrist was found negligent in his care of a patient with a diagnosis of depression who died after an overdose of salicylate. A generalist was held to be negligent in the death of a patient who was discovered at autopsy to have massive liver necrosis after taking cyclobenzaprine hydrochloride and propoxyphene napsylate/acetaminophen with suicidal intent. The physician had failed to forward laboratory results indicating liver injury to the psychiatric hospital; the medications were prescribed originally for traumatic neck strain. A chiropractor was negligent in the death of a patient with cancer who developed acute liver failure after self-injection of an unapproved alternative therapy, consisting of animal protein (neytumorin and neythymin), supervised by the chiropractor. Estimated chance of survival with conventional medical therapy was 50%.

In 5 cases, it was alleged that an anesthetic agent or the process of its administration was the cause of death. In a death due to acute liver failure following halothane anesthesia, the proximate cause of liver failure could not be established and the physician was acquitted. One sudden death occurred after local administration of lidocaine hydrochloride prior to drainage of an abscess; the autopsy reportedly confirmed the clinical diagnosis of “anaphylaxis” and the physician was acquitted. In 3 cases, multiple defendants, including 2 nurse anesthetists, were found to be negligent in deaths that occurred during induction of general anesthesia. No cause of death was found at autopsy in 1 of these cases. Another was due to cardiac arrest during epidural anesthesia; resuscitation was delayed because no intravenous access line was in place. In the third case, an emergency cesarean section was performed using open-drop ether anesthesia without a secure airway, a fact that was misunderstood by the surgeon; the patient aspirated and died immediately.

Role of the Autopsy

Physician defendants were acquitted in 30 of 49 cases in which the reviewers decided that the findings at autopsy slightly or strongly favored the plaintiff (Table 4). In 14 cases in which the autopsy findings did not clearly favor either the plaintiff or the defense, there was a finding of negligence in 3 cases. In 19 cases in which the autopsy findings slightly or strongly favored the defense, all 19 defendants were acquitted. In these 3 groups, the number of judgments that were reversed on appeal were 18, 1, and 2, respectively. In 16 of the 19 convictions affirmed by the appeals court, the reviewers thought that the findings at autopsy favored the plaintiff.

Table 4. 

Relationship of Autopsy Findings to Outcome in 82 Medical Malpractice Suits Based on 2 Reviews

Relationship of Autopsy Findings to Outcome in 82 Medical Malpractice Suits Based on 2 Reviews
Relationship of Autopsy Findings to Outcome in 82 Medical Malpractice Suits Based on 2 Reviews

In 17 cases, the findings at autopsy were deemed to have been important, and in some cases critical, to the acquittal of physician defendants. In many of these cases, the autopsy confirmed the antemortem clinical diagnosis, but in several an unexpected major finding at autopsy was helpful to the defense if there had not been sufficient time to conduct a reasonable medical investigation. Included in this total were several instances in which experts disagreed on the significance of minor autopsy findings, thereby neutralizing the influence of the autopsy. One case of a young child who collapsed mysteriously at home and was declared dead on arrival, should have been a medical examiner case in the opinion of the reviewers, but inexplicably was not. The suit involved a dispute over the significance of peritoneal blood found at autopsy after unauthorized “dry” postmortem peritoneal taps in the emergency room. The defendant emergency room physicians were acquitted. In several other cases, the interpretation of the autopsy findings by the original prosector seemed implausible to 1 or more of the reviewers, but was accepted by the court at face value and seemed, in both examples, to have influenced outcome in favor of the defendants.

No unequivocal cause of death was identified at autopsy in 8 cases. This group included 3 perinatal deaths and 5 unexpected deaths in adults; these deaths occurred in association with a hospital procedure, including 2 during anesthesia, 1 instance of bleeding after spinal surgery in which no source was identified, and 2 sudden deaths, one occurring 4 days after a hysterectomy and the other 3 weeks after a motor vehicle accident. In this group, there were 7 acquittals (2 reversed) and 2 findings of negligence (anesthesia-related). In one of the perinatal deaths, placental abnormalities seemed to have influenced the court to think that the baby was already compromised.

Outcome of Medical Malpractice Actions When No Autopsy Has Been Performed

Our search yielded 16 case records in which the word autopsy appeared in the appeals court record, but no autopsy had been performed. All records were reviewed in detail. There were 6 findings of negligence and 9 acquittals, of which 4 were reversed and remanded for new trials because of error. The ratio of outcomes is similar to that in cases with autopsy. One decision involved a technical issue only.

Three of these cases are selected to illustrate autopsy-related issues that may emerge during trial in deaths in which no autopsy was performed.

Example 1

A patient who had a history of previous heart attack and of gastroesophageal reflux developed chest pain for 2 days and visited a cardiologist who scheduled a catheterization for the next morning. The patient had a fatal cardiac arrest on the way home from the visit. Speculation on the cause of death by experts at the trial included myocardial infarction, electromechanical dissociation, ruptured aneurysm, and peptic ulcer eroding the aorta from the lower esophagus. In the absence of an autopsy, the jury was expected to evaluate the experts' testimony with respect to the most probable cause of death, a decision that could conceivably influence the decision about negligence. In this case, acquittal by jury was reversed because the defendant-physician's trial testimony favoring an ulcer penetrating the aorta had never been mentioned during pretrial discovery. The outcome is unknown.

Example 2

A middle-aged man developed chest pain and underwent cardiac catheterization, during which a catheter wire broke and remained within the heart to be discovered in a routine postoperative chest film a few hours later. An anonymous informant told the family about the wire after the patient died suddenly in the hospital 17 days later. The family alleged that multiple physicians had engaged in a conspiracy to conceal the accident by not telling them about it, by not mentioning it in the chart or during discussions of the need for a second operation, and by not requesting permission to perform an autopsy. The physicians told the family that the patient had died of another heart attack. In conceding the conspiracy theory, the defense argued that the wire could not be established as the proximate cause of death. Plaintiff's expert argued that thrombosis on the wire and thromboembolism were the most probable cause of death. A summary judgment by the court for the defense was overturned by the appeals court, which found that major fact issues regarding the proximate cause of death required a jury trial to resolve. The outcome is unknown.

Example 3

The defendant physician in a case of alleged malpractice claimed spoilation of evidence by the survivors of a patient when they refused his request for an autopsy and had the body cremated. The circumstances of the death had not required reporting to the local coroner. The appeals court noted that the responsible survivor has the right to determine disposal of the body of a relative and held that an autopsy would never be ordered by a court for the purpose of providing evidence in a civil case. On this basis, the physician-defendant's claim of spoilation by cremation was denied. The outcome of the suit is unknown.

COMMENT

In law, as in medicine, error and negligence are not synonymous.12 To establish liability in a medical malpractice action, a plaintiff must prove 3 elements. The duty inherent in the doctor-patient relationship must have been established. A breach of duty defined by failure to exercise the required standard of care must have occurred. The breach of duty must be shown to be the proximate (underlying) cause of injury or death to a reasonable degree of medical certainty, that is, that the injury or death could have been avoided except for the breach of duty.

Our study sample is a highly focused subset of cases that were tried in various states and appealed in the state courts. This sample may not be representative of medical negligence lawsuits in general, because most are settled out of court, many allege injury without lethal outcome, and the decision to file suit and to appeal the outcome of a trial may be influenced to an unknown extent by the availability and nature of autopsy findings. Additionally, although we believe it to be true, we are unaware of published data to support the common assertion that autopsy findings, more often than not, are useful in averting suits alleging medical negligence.13 

We found that information from the autopsy can be helpful to either the plaintiff or defendant, or can be neutral in a given case. Especially noteworthy is our finding that in 30 (61%) of 49 of cases in which the reviewers concluded that information provided by the autopsy favored the plaintiff, the defendant physicians were acquitted of medical negligence. Conversely, in 19 (100%) of 19 cases in which reviewers thought that the autopsy findings favored the defense, the defendants were acquitted. We found little evidence that a finding by a court for the plaintiff in a medical malpractice case is influenced by the existence of an autopsy report in which a major discrepancy exists with the working clinical diagnosis. On reflection, this should not be a surprise. Standard-of-care issues were at the heart of every case, and medical perfection, which is unattainable in any event, is not the standard of care. On the other hand, a finding by the court of medical negligence was significantly more common in cases in which we thought the autopsy findings favored the plaintiff (40%) than in cases in which the autopsy findings were neutral or favored the defense (10%).

The medical facts in the 99 cases examined in detail clearly indicate that most of these deaths were unexpected outcomes for the patients and their physicians, as well as the institutions or private offices where the alleged breech of standard of care occurred. When death rather than injury is the endpoint, an essential component of the definition of medical malpractice, proximate cause, cannot in most cases be thoroughly examined without information derived from autopsy. Exceptions occur in cases in which the clinical information gathered before death is sufficient to clarify diagnosis and the cause of death with reasonable medical certainty. Thus, autopsy information may help to focus the claims and counterclaims made in cases of alleged medical malpractice specifically on the questions of whether the alleged breach of duty pertains to the proximate cause of death and whether the death was avoidable, or whether chance of survival would have been greater had that breach not occurred.

Medical malpractice suits are usually filed because absence of honest disclosure causes the family to feel ignored and suspicious of deception or concealment.14 Unexpected autopsy results may influence the decision to sue by focusing attention on whether the patient was treated properly. When the basis for an emergent and incompletely defined lethal illness is determined at autopsy, this discovery inevitably will result in a more detailed review of the actions or inactions of physicians relative to a patient's complaints than would be possible in the absence of an autopsy. Depending on the facts in a case, such scrutiny can be beneficial to either the plaintiff or the defendant. The integrity of the process of adjudicating a dispute alleging medical negligence clearly benefits from availability of autopsy information in one very important way: speculation about the proximate cause of death is limited to examination of physician behavior in the context of clinical facts and an autopsy-proven cause of death. Without an autopsy, the attorneys of both defendants and plaintiffs may be tempted to introduce a broad range of probable, possible, and perhaps even implausible sequences in support of their claims.

To estimate how many patients might have had a greater chance of survival had the clinical diagnosis not been mistaken is a daunting task, even with autopsy information on the cause of death. This judgment was a component inherent in each reviewer's opinion on whether the autopsy favored the plaintiff or defendant. The class 1 error rate was 46% in cases in which the final outcome was known, but class 1 errors (death due to undiagnosed, potentially treatable disease) had no impact on judgments regarding negligence. That loss of chance of survival is itself an evolving concept was suggested by court positions on this point, which differed during the 30 years covered in our review.

In our series, a substantial number of cases dealt with the question of physician responsibility for suspected or unsuspected fatal adverse drug events, in which the results of postmortem analyses were decisive for clarifying the role of drugs, both licit and illegal. Recent studies have emphasized the frequency of adverse drug events in hospitalized patients, especially in elderly patients with multiple diseases.15,16 Autopsy has an indispensable role in the identification of the role of drugs in deaths, whether intentional or accidental. Juvin et al17 made the point that the autopsy may not be helpful if specimens appropriate for drug testing or determination of anaphylaxis mediators are not saved. Another situation in which autopsy information was particularly helpful was demise after a medical procedure, an event in which both autopsy rates and suspicion of egregious error are likely to be high.

The applicable standard of care in the emergency room setting differs from that applied to a hospitalized patient. A major mission of a hospital emergency department is to screen patients for life-threatening disorders that require immediate appropriate inpatient investigation and care. The question is whether to do more per patient to avoid error or to screen efficiently so as to maximize net benefit to those who present with a particular complaint.18 Court records suggest awareness that no screening paradigm is perfect. It was uncommon for a court to conclude that emergency room physicians contributed to loss of chance of survival, despite the fact that autopsy findings of major unrecognized and potentially remediable disease seemed to provide solid support for the position of plaintiffs. Successful lawsuits alleging medical malpractice in this setting appear to be based on a low index of suspicion by an emergency room physician who fails to investigate in sufficient depth to make a reasonable judgment about the basis for the visit. In any event, performing an autopsy to determine the cause of unexpected death occurring within a few hours or days of a visit to a physician is vital to physicians, the family of the deceased, and the community and may be required by the local coroner or medical examiner.

Among patients who died of pulmonary embolism, 3 convictions and 8 acquittals were recorded. The low frequency of either recognition or anticipatory treatment for pulmonary embolism may have contributed to the decision to sue the treating physician(s). One defense expert opined that acute pulmonary embolism is an autopsy diagnosis, not a clinical diagnosis. That is a debatable position. A more reasonable proposition is that failure to diagnose pulmonary embolism is not necessarily evidence of medical error or of malpractice. Whether recognition of massive or multiple episodes of pulmonary embolization conflicts with the standard of care is dependent on multiple factors, of which the difficulty of recognition is but one. Trial courts seem to recognize these distinctions. Cases in which physicians were convicted of malpractice in management of patients who died of pulmonary embolism displayed certain features, such as lack of aggressive preventative measures and delayed recognition of both deep leg vein thrombosis and penultimate episodes of pulmonary embolization in high-risk settings. Inattention may result in failure to detect early clinical signs and delayed requests for tests, such as lung scans, that may be helpful. As previously stated, in malpractice actions, the court must determine whether the physician's actions or inactions or the disease itself is the proximate (underlying) cause of death. Multiple causes may be operative, of which physician behavior is only one.

Much useful information was derived from this study. Our findings support the essential role of the autopsy in elucidating the nature of lethal disease processes, especially among patients for whom death was an unexpected outcome. However, autopsies are not infallible. About 8% of autopsies were not contributory, for one reason or another, to understanding the cause and mechanism of death. In a few cases, the most appropriate postmortem studies, given the clinical problem posed, may not have been performed. In about 5% of cases, the interpretation of the autopsy findings either seemed implausible to the reviewers or elicited differing opinions among experts for the plaintiff and defense, which we cannot comment on, never having reviewed the original materials. In most cases, the autopsy findings were accepted at face value in the appeals court summary and presumably had not been successfully challenged in the original trial. In 19 study cases in which the autopsy findings favored the defense, the court ruled no negligence in every case. Of more importance, in 39 cases in which autopsy findings were thought to favor the plaintiff, the courts held in the majority that no medical negligence had occurred, because the standard of care had not been breached. This pattern underscores an important general concept. Perfection in the practice of medicine is not possible and is neither demanded nor expected of physicians. The autopsy has much to offer in the elucidation of causes of unexpected, as well as expected deaths. Improvement of the quality of care depends on many things, including autopsy information. While autopsy findings can be used to justify a physician's actions in a medical malpractice lawsuit, it remains a fact that information derived from autopsies is also used to initiate lawsuits.19 We submit that our findings support the proposition that autopsy information is generally not harmful to defendants in lawsuits alleging medical malpractice. We conclude that fear of autopsy findings is unbecoming to the medical profession and an obstacle to the pursuit of excellence through uninhibited outcomes analysis.

Table 3. 

Continued

Continued
Continued

Acknowledgments

Members of the Autopsy Committee include Johannes Bjornsson, MD; Micheal Bell, MD; Kevin E. Bove, MD; John J. Buchino, MD; Kim A. Collins, MD; Gregory J. Davis, MD; Marsella Fierro, MD; Stephen A. Geller, MD; Randy Hanzlick, MD; Dean Havlik, MD; Grover M. Hutchins, MD; Eun Young Lee, MD; Lawrence C. Nichols, MD; Joseph Parisi, MD; John Sinard, MD; J. Thomas Stocker, MD; and Keith Volmar, MD.

References

References
Autopsy.
A comprehensive review of current issues: Council on Scientific Affairs.
JAMA
1987
.
258
:
364
369
.
McPhee
,
S. J.
Maximizing the benefits of autopsy for clinicians and families: what needs to be done?
Arch Pathol Lab Med
1996
.
120
:
743
748
.
Lundberg
,
G. D.
Low-tech autopsies in the era of high-tech medicine: continued value for quality assurance and patient safety [editorial; comment].
JAMA
1998
.
280
:
1273
1274
.
Goldman
,
L.
,
R.
Sayson
,
S.
Robbins
,
L. H.
Cohn
,
M.
Bettmann
, and
M.
Weisberg
.
The value of the autopsy in three medical eras.
N Engl J Med
1983
.
308
:
1000
1005
.
Landefeld
,
C. S.
,
M. M.
Chren
,
A.
Myers
,
R.
Geller
,
S.
Robbins
, and
L.
Goldman
.
Diagnostic yield of the autopsy in a university hospital and a community hospital.
N Engl J Med
1988
.
318
:
1249
1254
.
Papadakis
,
M. A.
,
C. M.
Mangione
,
K. K.
Lee
, and
M.
Kristof
.
Treatable abdominal pathologic conditions and unsuspected malignant neoplasms at autopsy in veterans who received mechanical ventilation.
JAMA
1991
.
265
:
885
887
.
Nichols
,
L.
,
P.
Aronica
, and
C.
Babe
.
Are autopsies obsolete? [see comments].
Am J Clin Pathol
1998
.
110
:
210
218
.
Burton
,
E. C.
,
D. A.
Troxclair
, and
W. P. I. I. I.
Newman
.
Autopsy diagnoses of malignant neoplasms: how often are clinical diagnoses incorrect?
JAMA
1998
.
280
:
1245
1248
.
Roosen
,
J.
,
E.
Frans
,
A.
Wilmer
,
D.
Knockaert
, and
H.
Bobbaers
.
Comparison of premortem clinical diagnoses in critically ill patients and subsequent autopsy findings [see comments].
Mayo Clin Proc
2000
.
75
:
562
567
.
Hastings
,
M.
and
S.
Andes
.
Autopsy—Past and Present: Summary of the Institute's Autopsy Research Program and of the 1995 Annual Autopsy Survey.
Brief report series 97-A. Chicago, Ill: The Institute of Medicine of Chicago; 1995
.
Friederici
,
H. H. R.
Turning autopsy liabilities into assets.
JAMA
1983
.
250
:
1165
.
Kapp
,
M. B.
Medical error versus malpractice.
DePaul J of Health Care Law. Summer 1997
.
Zarbo
,
R. J.
,
P. H.
Baker
, and
P. J.
Howanitz
.
The autopsy as a performance measurement tool: diagnostic discrepancies and unresolved clinical questions.
Arch Pathol Lab Med
1999
.
123
:
191
198
.
Fielding
,
S. L.
When patients feel ignored: study findings about medical liability.
Acad Med
1997
.
72
:
6
7
.
Lazarou
,
J.
,
B. H.
Pomeranz
, and
P. N.
Corey
.
Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies.
JAMA
1998
.
279
:
1200
1205
.
Ebbesen
,
J.
,
I.
Baujardet
, and
J.
Erikssen
.
et al
.
Drug-related deaths in a department of internal medicine.
Arch Intern Med
2001
.
161
:
2317
2323
.
Juvin
,
P.
,
F.
Teissiere
,
F.
Brion
,
J. M.
Desmonts
, and
M.
Durigon
.
Postoperative death and malpractice suits: is autopsy useful?
Anesth Analg
2000
.
91
:
344
346
.
Davis
,
M. A,
,
A.
Keerbs
,
J. R.
Hoffman
, and
L. J.
Baraff
.
Admission decisions in emergency department chest pain patients at low risk for myocardial infarction: patient versus physician preferences.
Ann Emerg Med
1996
.
28
:
606
611
.
Wecht
,
C. R.
Medical malpractice suits and autopsies.
JAMA
1991
.
360
:
361
362
.

Author notes

Reprints: Kevin E. Bove, MD, Department of Pathology, Children's Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229 (kevin.bove@uc.edu)