Article Text

Download PDFPDF
Autopsy interrogation of emergency medicine dispute cases: how often are clinical diagnoses incorrect?
  1. Danyang Liu1,2,
  2. Rongchang Gan3,
  3. Weidi Zhang1,
  4. Wei Wang1,
  5. Hexige Saiyin4,
  6. Wenjiao Zeng1,
  7. Guoyuan Liu1
  1. 1 Department of Pathology, School of Basic Medical Sciences, Fudan University, Shanghai, China
  2. 2 Department of Pathology, Fudan University Shanghai Cancer Center, Shanghai, China
  3. 3 Judicial Authentication Institution of Shanghai Minhang District Central Hospital, Shanghai, China
  4. 4 The State Key Laboratory of Genetic Engineering, Fudan University, Shanghai, China
  1. Correspondence to Dr Guoyuan Liu, Department of Pathology, School of Basic Medical Sciences, Fudan University, 138 Yi Xue Yuan Road, Shanghai 200032, China; gyliu{at}fudan.edu.cn

Abstract

Aims Emergency medicine is a ‘high risk’ specialty. Some diseases develop suddenly and progress rapidly, and sudden unexpected deaths in the emergency department (ED) may cause medical disputes. We aimed to assess discrepancies between antemortem clinical diagnoses and postmortem autopsy findings concerning emergency medicine dispute cases and to figure out the most common major missed diagnoses.

Methods Clinical files and autopsy reports were retrospectively analysed and interpreted. Discrepancies between clinical diagnoses and autopsy diagnoses were evaluated using modified Goldman classification as major and minor discrepancy. The difference between diagnosis groups was compared with Pearson χ2 test.

Results Of the 117 cases included in this study, 71 of cases (58 class I and 13 class II diagnostic errors) were revealed as major discrepancies (60.7%). The most common major diagnoses were cardiovascular diseases (54 cases), followed by pulmonary diseases, infectious diseases and so on. The difference of major discrepancy between the diagnoses groups was significant (p<0.001). Aortic dissection and myocardial infarction were the most common cause of death (15 cases for each disease) and the most common missed class I diagnoses (80% and 66.7% for each), higher than the average 49.6% of all class I errors of the study patients.

Conclusions High major disparities between clinical diagnoses and postmortem examinations exist in emergency medical disputes cases; acute aortic dissection and myocardial infarction are the most frequently major missed diagnoses that ED clinicians should pay special attention to in practice. This study reaffirmed the necessity and usefulness of autopsy in auditing death in EDs.

  • cardiovascular diseases
  • autopsy pathology
  • diagnosis discrepancy
  • emergency medicine
  • medical dispute

Statistics from Altmetric.com

Introduction

Emergency medicine physicians are confronted with a special pool of patients on a daily basis whose diseases develop suddenly, progress rapidly and, most importantly, correct diagnosis, and thus corresponding treatment in a timely fashion has a significant impact on patient outcome. Thanks to great improvement in medical technology, new concepts and fresh new treatment modalities, diagnostic errors generally have been shown to decline steadily.1 However, no data have shown decreases in diagnostic errors in patients who died in emergency department (ED) especially those involved in medical disputes.

Postmortem examination has long been deemed as the ultimate and most important step for apprehension of the diseases and determining the cause of death in patients.2 3 However, physicians may be reluctant to request an autopsy for the following reasons: the fear of medicolegal litigation, or due to traditional belief that corpus disintegration is unacceptable, or overconfidence about the cause of death and misconception that further investigation is unnecessary and redundant with the advent of increasingly sophisticated premortem diagnostic methods.4 In China, autopsy rate was constantly staggeringly low.5 Clinicians seldom requested autopsy unless a medical dispute emerged and autopsy was mandated by law. To the best of our knowledge, limited studies that investigated emergency room diagnostic errors using autopsy have been conducted in China. It is still unclear to what extent does autopsy exert its influence in emergency medicine and how often clinical diagnoses are correct in emergency medicine dispute cases. Therefore, we analysed a consecutive autopsy series spanning 15-year period in our autopsy centre that involved in medical disputes from ED in order to: (1) investigate the concordance between clinical diagnosis and autopsy findings and (2) identify the most common major missed diagnoses of these patients.

Methods

Selection of cases

We analysed retrospectively the medical and necropsy records of 117 selected patients (from 1 January 2001 to 31 July 2015) that involved in medical disputes who died in the EDs of several hospitals in Shanghai, as well as those who were brought to the EDs with loss of consciousness and declared death in the EDs, while the family requested autopsy. The patients underwent a postmortem examination at Department of Pathology, School of Basic Medical Sciences of Fudan University, Shanghai, China. The study was approved by the Medicine Ethics Committee of School of Basic Medical Sciences, Fudan University.

Analysis of reports

We recorded the age, sex, time (from admission to EDs to death), premortem clinical diagnosis or symptoms and necropsy diagnosis of each case. Full copy of medical records and related medical information of the deaths were routinely requested from the hospitals or the relatives. A complete necropsy including histological assessment of each organ was carried out by three pathologists for included patients. The pathologists reviewed macroscopic and microscopic findings and made necropsy diagnosis for each case. Then clinico-pathological conferences (CPC) attended by clinicians and consultant pathologists were held to discuss the cause of death and the appropriateness of clinical management for each case.

Necropsy diagnoses were grouped according to International Classification of Disease, 10th edition (ICD-10): infectious diseases, neoplastic diseases, cardiovascular diseases, pulmonary diseases, gastrointestinal diseases and miscellaneous (remaining diagnoses). After a thorough review of patient’s clinical information, necropsy diagnosis and the cause of death in autopsy reports and CPC discussion notes, the discrepancy occurrence and its degree for each case was evaluated according to the method (table 1) of Goldman and colleagues,6 the modification of Battle and colleagues,7 and as non-classifiable cases8 by three independent investigators. When disagreement occurred, it was resolved by debate until a consensus was reached.

Table 1

Description of diagnosis discrepancy classes

Data analysis

In this study, we only compared the main diagnosis that directly involved in patient death and did not take minor discrepancies into consideration for the reason that in the practice of emergency medicine, it is the ED physician’s top priority to pin down problems that threaten patient’s life rather than focusing on problems that can be handled non-urgently. We compared the difference of major discrepancy and class I diagnostic errors between diagnosis groups with Pearson χ2 test the SPSS V.15.0 statistical software package. All reported p values were two sided and p≤0.05 was taken to be significant.

Results

Characteristics of study cases

A total of 117 medical disputes cases (male/female: 83/34) with median age 46 years old (range: 32 days–88 years old) that happened in the ED were included in this study. Time from admission to EDs to death ranged from 0 hour to 13 days and 12 hours for all the cases.

Discrepancies between clinical diagnoses and autopsy diagnoses

Table 2 presents the composition of ICD-10-based autopsy diagnoses of the study patients. Fifty-eight cases (49.6%) were classified as class I errors, and 13 cases (11.1%) as class II errors. Thus, 71 cases (60.7%) were revealed as major discrepancy when class I and class II missed major diagnosis were pooled (table 2). The detailed data of each case with major diagnosis errors are listed in the online supplementary table S1

Supplementary Material

Supplementary material 1
Table 2

Distribution of the primary causes of death, major missed diagnosis and class I errors of the study patients

Complete agreements between clinical diagnosis and postmortem examination (class V discrepancy) were found in 32 cases (27.4%), and eight cases (6.8%) were classified as class VI discrepancy who were brought to the ED with loss of consciousness due to cardiac arrest, brain haemorrhage, ectopic pregnancy rupture and so on, and subsequently declared death in the ED with no diagnostic procedure (figure 1). Six cases were classified as minor discrepancy with the sum of class III and class IV missed diagnoses.

Figure 1

Distribution of discrepancy classes I–VI of the study patients. Note: seven patients had diagnosis in more than one discrepancy category, totally.

The probability of six different discrepancy classes did not depend on gender or age (in online supplementary file), but depend on time from admission to death (p=0.018). Moreover, the differences of class I and class II discrepancy between different time groups was significant with p=0.04, and the cases with deaths within 48 hours were more prone to have class I diagnosis errors (p=0.02).

Primary causes of death of the studied patients

Of the six necropsy diagnosis groups (table 2), the most common major diagnoses (principal underlying disease and primary causes of death) for study patients as determined by postmortem examination were cardiovascular diseases (54 cases), which were followed by pulmonary diseases (26 cases), infectious diseases and gastrointestinal disease (12 cases for each group). The difference of major errors (class I and class II) between the six groups was significant at p<0.001 by Pearson χ2 (table 2) as well as the difference of class I errors. However, the probability of missed major diagnoses did not depend on gender, age or time from admission to death in(online supplementary file).

Most common diseases involved in medical disputes cases in ED

Aortic dissection and myocardial infarction (15 cases for each disease) and its related complications were the most common cause of death of the included patients (table 3), followed by interstitial pneumonia (10 cases), enteral infections with toxic shock (7 cases), airway obstruction and brain haemorrhage (6 cases for each). Among them, aortic dissection was the most common class I missed diagnosis (80%), while myocardial infarction was the second most common missed diagnosis (66.7%), higher than the average 49.6% of all class I errors of the study patients (table 2).

Table 3

The rate of class I error of the top six frequently diagnoses (case number ≥5)

Noteworthily, of the 117 dispute cases, six cases had enlarged thymus glands detected by postmortem findings and termed as status thymicolymphaticus, which may contribute sudden unexplained death.9Five patients were diagnosed as schizophrenia clinically, with one case ruled out eventually, which was diagnosed as encephalitis after autopsy (see online supplementary file).

Discussion

Even though abundance of investigations have been carried out to compare premortem clinical diagnosis with postmortem autopsy conclusions in medical patients8 or surgical patients10 or intensive care unit patients,11 12 there were only limited research focusing in ED.13 14 Our findings, that high disparities between clinical diagnoses and postmortem examinations exist in emergency medical disputes cases, suggest that autopsy plays an essential role in uncovering the true cause of death of patients who have died in ED and that death certificate issued by clinicians may be misleading and potentially skew the demographic characteristics of epidemiological data.

In this study, the major discrepancy between clinical and postmortem examinations is in the higher range (60.7%) compared with median rate of 23.5% (range, 4.1%–49.8%) for major errors in the study that reported by Shojania et al,1 which analysed published 53 autopsy series of different case mix, while the present study was focusing in EDs. It is virtually impracticable for ED physicians to detect the underlying cause of symptoms and accordingly formulate correct diagnosis in a limited period of time when a patient presents with rather common and non-specific symptoms that entailed meticulous differential diagnosis. However, certain diseases progressed rapidly that may completely go out of control, as we found that the cases with deaths within 48 hours were more prone to have class I diagnosis errors. All these factors would contribute to the seemingly high rate of discrepancy.

However, the discordance rate of this study on ED was much higher than other reports. O’Connor et al reported the incidence of major discrepancy was 40.6%14 for patients who die in ED. In a study carried out specifically in a paediatric emergency room setting, Goldman class I errors and class II error rate was a total of 15%.13 The major reasons for the wide disparity between our study and others maybe stem from the low autopsy rate in China and the fact that all the cases included in this study were medical dispute ones, which may have an increased possibility of clinical diagnostic errors and result in the selection bias.

As reported by Jayawardena et al, cardiac system involvement was the leading cause of death in both male and female patients that died within 48 hours of admission to the ED3; the most common causes of death in this study were cardiovascular disorders (46% of cases) including coronary heart disease and rupture of vascular. The sudden unexpected deaths in the EDs may provoke dissatisfaction of the relatives, and prompt them to ask the physicians to make up for a perceived treatment error, whether the given case involved actual medical malpractice. Similarly, these conditions have constituted the majority of the medical disputes of this study.

These cardiovascular diseases are notoriously difficult to diagnose, as we found in this study that acute aortic dissection and myocardial infarction accounted for a great majority of class I missed diagnoses. Inexperienced physicians may be baffled by the situation especially when patients were relatively young or without corresponding risk factors for cardiovascular diseases. In addition, these cardiovascular disorders are catastrophic unless prompt diagnosis and relevant interventions are implemented. For example, given the high mortality dissection of the ascending aorta which most presenting chest or back pain,15 rapid diagnosis of acute aortic dissection is imperative, and an emergency surgical procedure such as thoracic endovascular aortic repair is needed, which has dramatically changed the treatment paradigm for the disease.16 17 Thus, in practice, ED clinicians should pay special attention to acute aortic dissection and myocardial infarction, which are easily involved in emergency medical disputes, while correct diagnosis and thus corresponding treatment may have a different impact on patient outcome.

Noticeably, four of cases (3.41%) were diagnosed as schizophrenia clinically that has a much higher rate than the average incidence of this disease (0.3%–0.7%). The major causes of the death of these patients include suicide, cardiovascular diseases and acute alcoholic hepatitis, as reported by Ifteni et al.18 Schizophrenia is associated with premature mortality and a high rate of sudden, unexpected deaths.18 The relatives always request autopsy for the cause of the death when this happens, though there were no major missed diagnosis errors of these patients as we discovered in this study.

Some limitations of this necropsy study should be addressed. First, this study was carried out in a retrospective manner that needed interpretation of medical files, thereby underinterpretation or overinterpretation may potentially happen. This bias was overcome to some extent by having three independent assessors for each case and taking CPC notes for reference. Second, though deemed as the gold standard, autopsy can only reveal the nature of underlying disease only when it inflicts morphological alterations that are visible to the naked eye or under the microscope. Thus. metabolic or endocrine or cardiac electrical dysfunction may be missed, leading to the autopsy failing to reveal the cause of patient death. This problem, however, is inherent in all autopsies. Third, the current study represented a single-centre retrospective review of emergency medical dispute cases about a 15-year period; there was an inherent selection bias as we mentioned above. Fourth, all the cases in this study were medical disputes autopsies rather than clinical autopsies, which lead to seemingly high diagnostic errors. Patients who died in the EDs were seldom subjected to autopsy in China unless medical disputes demand autopsy to be conducted. If autopsy of death in the EDs not mandated by law are included, the incidence of major diagnostic errors were likely to be much lower.

Thus, in contrast with reported findings, our results confirmed highly frequent occurrence of disparity among clinical diagnosis and autopsy findings. Acute aortic dissection and myocardial infarction are the most frequently major missed diagnoses for the emergency medical disputes that ED clinicians should pay special attention to in practice. Though autopsy rate is continuing decline, it has an irreplaceable role in clinical and pathology practice, and our study reaffirmed the necessity and usefulness of autopsy in auditing emergency room deaths.

Take home messages

  • High major disparities between clinical diagnoses and postmortem examinations exist in emergency medical disputes cases.

  • Acute aortic dissection and myocardial infarction are the most frequently major missed diagnoses that ED clinicians should pay special attention to in practice.

  • Autopsy has an irreplaceable role in clinical and pathology practice.

Supplementary Material

Supplementary data

References

Footnotes

  • Contributors GL and DL conceived the study and participated in data collection, data interpretation and analysis, and manuscript writing. RG, WZ and WenZ participated in the data collection and interpretation. HS participated in statistical analysis. WW participated in critical analysis.

  • Funding This work was supported by the Natural Science Foundation of China (NSFC 31671461, NSFC 31301162).

  • Competing interests None declared.

  • Ethics approval The Medicine Ethics Committee of School of Basic Medical Sciences, Fudan University.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.