Sage Journals: Discover world-class research

Abstract

The rate of requests for autopsies has dropped over the last four decades. In the present study we have investigated the clinical usefulness of autopsies in a university hospital for determining the cause of death and as a source of information for the relatives. The autopsy routines were sufficient to fulfil clinical needs and resulted in essential information. The next of kin were involved in the decision to perform an autopsy in half the cases. In 10% of cases the relatives requested the examination. We believe that a desired increase in autopsies performed could be reached if the relatives of the deceased were informed of their value. We further suggest that nurses be included in a team process informing the relatives of the value of autopsies and their right to share the information gained.

Keywords

autopsy cause of death information to kin Sweden

Introduction

The benefits of clinical autopsies in the healthcare system for determining the cause of death, for continuing to educate the medical profession and for informing the next of kin are well known and have been confirmed in a multitude of studies from all over the world. In spite of this, the number of clinical autopsies is rapidly decreasing, while the number of forensic autopsies is not. In the United Kingdom, almost all autopsies are performed after the coroner determines the cause of death, while clinical autopsies are as low as 0.69%.¹ In Sweden, as in other countries around the world, the number of autopsies performed has steadily decreased over the last four decades, from locally nearly 100% and nationally 50% to 10% or lower today in some areas.^2–7 This means that the cause of death is determined primarily based on clinical assessments and investigations performed prior to death, although we know from previous studies that autopsy provides additional information about more than 40% of deceased patients.^2,8,9 In addition, Pritt et al. identified multiple errors on 82% of death certificates, including major errors on more than 30%.¹⁰

Clinicians are requesting fewer autopsies for many reasons.^11,12 Studies have suggested that the low rate of requests for autopsy could be because the process required to obtain permission or consent from the next of kin is cumbersome; relatives are not asked for permission in half of cases; and autopsy reports provided by the pathology department are of poor quality or do not reach the responsible physician in time to be of clinical value.^1,5,11 In a high proportion of cases where an autopsy was performed, the relatives were never given the results of the post-mortem examination.^11,13 Taking into consideration that the relatives of the deceased are increasingly involved in patient care, and more frequently are asking for autopsies to be performed for the benefit of their own well-being and health,¹⁴ this retrospective study investigated the usefulness of routine autopsies, both for fulfilling the obligations of the medical staff and for informing relatives.

When a person dies at home, whether or not in the healthcare system, the cause of death must be determined according to Swedish law and reported to the authorities through completion of a death certificate. The cause of death is of interest not only to the authorities but also to the relatives and healthcare personnel. The cause of death is determined by the physician responsible for the care of the patient at the hospital, nursing home or outpatient setting. The physician can request a post-mortem examination to support the initial diagnosis after obtaining consent from the next of kin. When persons die from unnatural causes, such as accidents, intoxication or crime, an autopsy is requested by the police and performed in a forensic department. If an unnatural cause of death can be excluded, the post-mortem examination is conducted in the pathology department of a hospital by a clinical pathologist.

Until the end of the 1980s, autopsies were frequent in Sweden. Burton and Underwood reported that in 1984 the autopsy rate in the city of Malmö was 81%,⁴ while the national frequency was reported as 38%.⁵ Since then, the relative frequency of autopsies has steadily decreased. From a national frequency of 34% in Sweden, the number of autopsies performed is now below 10%: 7% for women and 15% for men, including forensic autopsies.¹⁵ The reason for this decrease has not been investigated, and many causes have been suggested, such as a lack of interest and consent from the relatives; a lack of need because physicians have enough information to determine the cause of death from the clinical investigation; cost; administrative burden; late and low-quality reports from autopsies performed; or insufficient value or quality of autopsies to contribute to the understanding of the cause of death. Another interesting statement is that the frequency of autopsies requested is connected to the earlier experiences of autopsies during education and clinical training.⁷

The present study aims to investigate the clinical usefulness of autopsies in the healthcare process for determining the cause of death and providing information to relatives and next of kin. In this work, we have addressed the question of the value of autopsy in a university clinical setting. The study is a follow-up to two earlier studies performed in the 1970s by Mona Briton^16,17 and in the 1990s by Bela Veress and Irina Alafuzoff,⁹ when autopsy rates were significantly higher in Stockholm and at the Huddinge Hospital (now part of Karolinska University Hospital). These studies showed that autopsies contributed significantly to understanding the cause of death and the diseases that led to death. In 25% of cases, the cause of death suggested before autopsy by the physician was incomplete or misleading, and in 15% of cases, autopsy results revealed that the disease at the focus of the treatment effort was incorrectly diagnosed. Fifteen per cent of all cancer cases, half of them fatal, were discovered by a pathologist during post-mortem examination.⁹

We performed this study on the value of autopsies in a setting where the autopsy rate is low, and post-mortem examination is an exception rather than a rule, to compare our data with the two earlier studies performed when the autopsy rates were significantly higher in the same university hospital context. This study investigated, for the first time in Sweden, the post-mortem examination procedure from start to finish, that is, from obtaining consent to reporting the cause of death to the authorities and the family of the deceased. This question is of particular interest today, when relatives are increasingly knowledgeable about disease and often participate in patient care.¹⁸

Materials and methods

Design

In this retrospective study, we investigated 54 autopsies performed at a university hospital in Stockholm. Of these, 53 autopsies were complete and included dissection and evaluation of all visceral organs including the brain. One case that was only partially examined was excluded from the study. The patients were cared for at different clinical departments where personnel were accustomed to taking care of patients with terminal diseases and their relatives. We asked the following questions:

When were the autopsies performed?

Was the report released prior to issuing the death certificate?

Did the autopsies add clinical value for the healthcare personnel and the relatives?

Who requested the autopsy?

Who was informed of the results of the autopsies?

We also compared death certificates with the cause of death determined after autopsy. The clinical usefulness of the autopsies and the autopsy reports were defined based on whether the above questions were answered or not. We did not include data on the use of the autopsies for educational purposes or research due to the fact that data supporting these useful activities were not reported in the medical records. Thus, in this article, we address clinical autopsies performed during 2008 and 2009 from the end-user perspective, with a special focus on how the relatives participated in the decision to perform the autopsy and how they were informed of its results.

The clinical diagnoses registered, reported or suggested in the medical records were compared to the diagnoses registered in the pathology reports. In this study of the value of the routine clinical autopsy we did not evaluate the quality of the diagnoses described in the autopsy reports.

Materials

Patients registered at the Karolinska University Hospital between 2008 and 2009 who died at the hospital were identified in this study. During these two years, 1588 adult patients died at the hospital, and of these patients 174 (11%) were autopsied. In order to evaluate the routines in the wards where personnel were accustomed to caring for deceased patients and more likely to consider the possibility of an autopsy, we selected three departments – gastroenterological surgery, haematology and geriatrics – from which we studied all of the autopsies requested. These clinics were the same as those selected for a previous questionnaire study on nurses’ and physicians’ attitudes about autopsies.¹⁴ In addition to these cases, we randomly selected 20 cases from other departments that were experienced in caring for dying patients. In sum, we studied 54 autopsies. In each of the 54 cases, the medical record, including the nurse’s record;¹⁹ the request form for autopsy; the autopsy report; and the death certificate were examined and compared. One case was excluded because the autopsy requested was partial and incomplete. In this limited study we have not considered ethnic, cultural or religious differences as possible explanations for the reduced frequency of requested autopsies. These factors as well as socio-economic and geographic differences are included in an ongoing national registry study by the same authors.

The deceased patients (Table 1) consisted of 32 men and 21 women with a mean age of 69.6 years (range 45 to 88 years, median 66.5 years). Twenty-five (47%) of the deceased, 10 women and 15 men, came from gastroenterological surgery; 6 patients, 3 women and 3 men, came from haematology and only 2 were women from geriatrics. The rest of the cases, 6 women and 14 men, were in other departments where the units primarily providing care were focussed on lung, kidney or neurological diseases, or were patients in intensive care units. In this group, the autopsy frequency was 15%, whereas the mean frequency for the entire hospital during these years was 11%.

Table 1.

Patient age and sex from different clinics and documented times for autopsy request, performance and report.

			Care	Time to (days):
Code	Sex	Age	time (days)	Request	Autopsy	Autopsy report
Gastroenterological surgical clinic
18	F	58	28	1	12	13
50	F	64	11	1	14	17
30	F	65	13	3	8	9
42	F	66	17	3	8	11
38	F	72	162	1	8	8
8	F	73	4	1	4	21
36	F	75	21	1	8	43
10	F	77	10	4	13	14
2	F	79	5	3	9	12
37	F	85	10	1	8	28
32	M	50	5	5	9	20
29	M	54	54	1	7	7
43	M	59	16	1	16	17
6	M	61	13	1	9	10
19	M	62	5	1	15	16
35	M	66	24	1	9	21
17	M	66	11	1	4	7
33	M	66	2	1	7	20
3	M	67	22	1	8	11
1	M	68	5	1	4	11
34	M	70	10	1	16	31
31	M	74	5	5	15	16
9	M	74	14	1	6	11
7	M	75	14	2	4	8
28	M	76	7	1	7	8
Haematology clinic
22	F	45	56	2	5	6
21	F	69	4	1	7	7
20	F	71	40	1	7	9
15	M	82	7	2	3	3
4	M	72	12	1	6	10
14	M	74	47	1	8	10
Geriatric clinic
16	F	85	24	1	10	16
5	F	88	10	1	7	12
Other patients
Lung clinic
24	F	57	5	1	14	19
11	F	78	6	1	6	8
23	M	59	12	1	5	5
13	M	74	6	1	5	17
12	M	74	11	1	3	3
Urology clinic
25	M	57	27	1	5	171
26	M	87	19	2	5	16
44	M	92	3	1	8	9
Neurology clinic
46	F	71	1	5	9	10
27	F	81	1	1	7	8
48	M	55	22	3	12	13
45	M	66	8	2	10	36
47	M	70	10	1	5	9
Orthopaedic clinic
40	F	76	2	14	20	147
39	M	85	1	1	5	37
Intensive care clinic
41	M	60	29	2	3	16
49	F	82	1	1	7	19
Emergency clinic
51	M	58	1	1	9	11
52	M	60	1	2	7	10
53	M	60	1	1	3	262

The diagnoses reported in the records and in the autopsy protocol were compared and registered according to a Goldman protocol modified by us (Table 2). The Goldman protocol, published in 1983 by Lee Goldman et al.,²⁰ has been used in several reports^21–23 and a modified Goldman protocol was used by others.^1,24 The Goldman protocol classifies diagnoses that have been determined after autopsy but have not been described in the medical record. Goldman uses four classes of missed diagnoses. Goldman class I concerns major diagnoses that were missed, with a potential impact on survival, that would have changed management; Goldman class II concerns major diagnoses that were missed, with no impact on survival, which would not have changed therapy; Goldman class III concerns minor diagnoses that were missed which were related to the terminal disease but not to the cause of death and Goldman class IV concerns minor diagnoses that were missed which were not related to the terminal disease or the cause of death. Two modifications were introduced in the Goldman protocol to adapt to the goal of this study. The first modification was increasing the number of categories to six: class V concerned known diagnoses of earlier diseases that were confirmed by autopsy and class VI concerned other findings of significance to the medical history. The second modification was an extension of the phrases used to describe each category. In addition to the new classes of missed diagnoses, we introduced a phraseology using the term ‘diagnosis confirmed’ to accommodate suggested diagnoses that were not confirmed in the medical records. The reason for the addition of classes V and VI was to evaluate the autopsy report as an information source for the relatives, who have detailed knowledge of the medical history of the deceased.

Table 2.

Modified Goldman criteria for autopsy discrepancies.

Major discrepancies	Class I	Missed major diagnosis with potential adverse impact on survival and that would have changed management. In Class I we have also included major diagnosis with impact on survival, that was confirmed by the autopsy.
	Class II	Missed major diagnosis with no potential impact on survival and that would have not changed therapy. In Class II we have also included major diagnosis with no impact on survival, that was confirmed by the autopsy.
Minor discrepancies	Class III	Missed minor diagnosis related to terminal disease but not related to the cause of death. In Class III we have also included minor diagnosis related to the terminal disease but not related to cause of death, that was confirmed by the autopsy.
	Class IV	Other missed minor diagnosis not related to terminal disease or cause of death. In Class IV we have also included minor diagnosis not related to terminal disease or cause of death, that was confirmed by the autopsy.
Earlier known diseases	Class V	Earlier known diseases confirmed by autopsy.
Other findings	Class VI	Other findings of significance to the medical history

Each case was classified and categorised by the first author (AR) together with an experienced pathologist (LE). All of the authors discussed the classification and categorisation, and participated in the preparation of the manuscript.

Ethics

This project was approved by the Regional Ethical Review Board in Stockholm (2011/208–31). The patient records were coded to assure the anonymity of the deceased and their next of kin.

Findings

The time of care for the patients in different units varied from 27 days for haematology cases to 8 days for the 20 cases from mixed wards (Table 1). The mean time at the hospital for all cases was 16 days. In the majority of cases from all of the units, the autopsy request was written on the day of, or the day after, the death of the patient, but in five cases the request was written five or more days after the time of death. The request form arrived at the morgue two days after the time of death, and the post-mortem examination was carried out within seven days. However, in five of the 53 cases, the examination was performed two weeks after the time of death for reasons that were not explained in the records. The written autopsy record was registered in the medical record 12 days after the death of the patient. However, in two cases, the registration of the written report in the medical record was more than 100 days after death, with no explanation or obvious reasons. Telephone reports or direct demonstrations to the responsible physician were not registered, but we would expect that they took place. In all cases, tissue samples were taken, fixed, embedded, cut, stained and stored. In 23 of the 53 cases (43%), a microscopic examination was performed and reported.

From the data shown in Table 1, it is obvious that requests for autopsy were submitted within the first day or first few days after the time of death. On 23 occasions, representing 43% of the cases, the physician suggested or required the post-mortem examination. In these cases, dialogue or consent was not reported in the medical record, although an oral consent for autopsy is mandatory. In 25 cases (45%), the request was the result of a discussion between the physician and the deceased person’s relatives (Table 3). In six cases (11%), autopsies were requested by the relatives. According to the medical records, in two-thirds of the cases, the relatives were informed of the autopsy result, while in one-third, the relatives’ request to be informed was not registered. In 40% of the cases, this information was supplied by the physician over the phone, while in 8% it was given in a personal meeting. In two cases the relatives received the information in a letter written by the physician. In six cases (11%), the relatives asked for and received the protocol from the autopsy. Interestingly, only three of the six families that requested an autopsy were given the entire protocol. In 43% of the medical records, we have no information regarding if and how the relatives were informed. We could not find information showing that the nurses documented discussions with the relatives, or were involved in the decision to request an autopsy, in any of the nurses’ records that we investigated, despite their interest in being more involved in this process.¹⁴ It should be added that the decision to request an autopsy is not the nurses’ responsibility.

Table 3.

Party requesting the autopsy and how kin were informed of the autopsy results.

	Number of patients	Autopsy request			Kin informed		How kin were informed
	Number of patients	Doctor	Kin	Both	Yes	n.d.	Phone	Letter	Report	Contact	n.d.
Surgical clinic	25	11	3	11	13	12	8	0	3	2	13
Haematology clinic	6	3	0	3	3	3	3	0	0	0	3
Geriatric clinic	2	0	0	2	2	0	0	0	1	0	1
Other clinics	20	9	3	8	17	3	10	2	2	2	6

Note. n.d. = not documented

In Table 4 the Goldman classification of clinical diagnosis is presented. Thirty-three patients (62%) fulfilled the criteria for Goldman class I–II diagnoses (number of diagnoses, 42) and 40 (75%) had Goldman class I–III diagnoses (number of diagnoses, 59). Goldman class I–III diagnoses were found in all of the participating clinics and the distribution among the different clinics appeared to be even. Fourteen of the patients with Goldman class I–III diagnoses fit into more than one class (Table 4). The Goldman class IV diagnoses were all found in patients that also had class I–III diagnoses. Only one-quarter of the cases had no findings that qualified for Goldman class I–IV. The most dominant diseases in the Goldman classes were cardiovascular diseases, but lung diseases, cancers and neurological diseases were also common. In Table 4, the number of autopsy reports with diagnoses in Goldman classes V and VI, representing earlier known diagnosis confirmed by autopsy (V) and other relevant findings (VI), are also provided.

Table 4.

Number of diagnoses in each Goldman class and number of patients with diagnoses in Goldman classes I to III.

	Number of diagnoses in Goldman classes						Total number of patients	Number of patients Goldman diagnoses Class I–II	Percentage of patients with Goldman diagnoses I–II	Number of patients Goldman diagnoses Class I–III	Percentage of patients with Goldman diagnoses I–III
	I	II	III	IV	V	VI	Total number of patients	Number of patients Goldman diagnoses Class I–II	Percentage of patients with Goldman diagnoses I–II	Number of patients Goldman diagnoses Class I–III	Percentage of patients with Goldman diagnoses I–III
Surgical clinic	13	5	7	12	10	7	25	14	56%	17	68%
Haematology clinic	3	2	3	2	3	2	6	4	66%	5	83%
Geriatric clinic	1	0	1	1	2	1	2	1	50%	1	50%
Other clinics	11	7	7	5	12	15	20	14	70%	17	85%

In Sweden, the law dictates that everyone who dies shall have a certificate written by a physician specifying the cause of death and the major disease that ultimately led to death, in addition to a certificate stating that the patient is dead. In this study, we found the death certificates in 52 of the 53 cases (Table 5). One certificate was missing, which is surprising but consistent with the national figures indicating a loss of 1% of death certificates. Thirty-three (62%) of the death certificates were based on a report from the post-mortem examination, and nine (17%) were based on a clinical examination before death in combination with information from the autopsy report. In nine cases (17%), the certificate was based only on clinical examination. In one case (2%) the information was missing and in one case (2%) the certificate itself had not reached the authorities or was found in the medical record. The information on the death certificate was consistent with the information in the autopsy report in 45 cases (85%), but in seven cases (13%) the information was different from what was written in the autopsy report or missing. These patients had a mean care time of 19 days and received the autopsy report within two weeks from the time of death. In five of these seven cases we found Goldman class I diagnoses that were clinically unexpected (pulmonary embolism, myocardial infarction, unknown malignancy and vascular central nervous system insults).

Table 5.

Consistency of the death certificate with the autopsy report.

	Death certificate		Reported cause of death based on
	Consistent	Inconsistent	Autopsy	Clinical data	Both	No report
Surgical clinic	22	3	14	7	4	0
Haematology clinic	5	1	3	2	1	0
Geriatric clinic	1	1	2	0	0	0
Other clinic	18	2	14	0	4	2
Number	46	7	33	9	9	2

Discussion

The goal of this study was to determine the clinical value of routine autopsies and to evaluate the value of the information gained at autopsy for the clinical process, for diagnosis before autopsy, for informing relatives and healthcare personnel and for the death certificate. The diagnoses per se were not re-evaluated, confirmed or critiqued. In a clinical situation, where the frequency of autopsies is decreasing rapidly, we addressed whether autopsy today lacks clinical value.^25,26 The frequency of forensic autopsy, performed on patients who had died from unnatural causes, including accidents and crime, has not decreased.^15,27 Several reports have shown that in a significant number of cases, the next of kin was never contacted or involved in any discussion concerning autopsy, and was not informed of the value or possibility of a post-mortem examination. It is also reported that in many situations where an autopsy was performed, the protocols were delayed, incomplete or unclear, and that the relatives never received the results of the investigation.^5,6,11,13,26

We selected departments at the Karolinska University Hospital in Stockholm that were experienced with patients dying on their wards and also which had also been included in an earlier study using questionnaires to evaluate how healthcare personnel used and perceived post-mortem examinations. The departments selected represented different active research departments in the hospital: gastroenterological surgery, haematology and geriatrics. The selected departments had very different rates of post-mortem examinations, ranging from 23% for surgery to 4% for geriatrics. Haematology had a 10% autopsy frequency, which was close to the mean value for the hospital.

By examining data from 53 consecutive autopsies, we concluded that the request for autopsy was generated by the responsible physician within one or two days of death, and that the autopsy was performed and reported within seven to ten days of death, with few exceptions. These routines satisfied the physician’s need for information to write the death certificate and the next of kin’s need to inform relatives in time for the funeral.

We concluded from the documentation in the medical records that two-thirds of the relatives were informed of the results of the autopsy. We also determined that the autopsy results were used to correctly complete the death certificate in just over three-quarters of the cases. However, in more than 20% of the cases, the certificates were completed based only on clinical information. The correlation between the cause of death reported on the death certificate and in the autopsy report was adequate in 85% of the cases, but in 13% of the cases the information was not in agreement with the autopsy report or was missing. It is difficult to find a consistent reason in this limited material for the discrepancies, but in five of seven cases unexpected Goldman class I diagnoses were discovered. It is obvious from our investigation that post-mortem examination supports the clinical process and that it is very important for determining the cause of death. The importance of post-mortem examination is illustrated by studying the diagnoses that were added, confirmed or changed by the autopsy. We classified these diagnoses according to the Goldman protocol to indicate diagnoses that were important factors leading to death or the major disease that led to death (Table 4). We also added findings that were less significant for the cause of death but could have caused symptoms that were obvious to the relatives, or were of epidemiological interest, but that we did not expect to find in the medical record. Diagnoses classified in Goldman class I–III were either missing in the medical record or were suggested but not confirmed while the patient was alive. It is notable that 75% of the patients fulfilled the Goldman I–IIII criteria, and 14% had diagnoses that fit more than one class. Only one-quarter of the patients did not have a diagnosis in Goldman class I–III. The relative number of discrepancies found by our investigation was of the same magnitude as had been previously reported from the same area,^9,16,17 and in other countries at hospitals in studies performed 20–40 years ago, when the relative number of autopsies performed was significantly higher.^9,11,20,21 It is clear that post-mortem examinations in a clinical setting are of as much value today as they were in the 1970s and 1990s in Stockholm. In a recent report on patients from intensive care units, surgical wards and nursing homes it was stated that the discrepancy in diagnosis before and after autopsy was not dependent on ante-mortem diagnostic evaluation.²⁸ In recent reports from Fröhlich et al. and Wong et al. the number of missed diagnoses are lower than reported in previous studies and in this work.^1,24 These studies of patients in intensive care units are, however, based on Goldman modifications, and included the possibility of successfully treating a patient for the diagnosis reported in the autopsy report before death, or the possibility of actually confirming the diagnosis ante mortem.

We also concluded that the process and flow of information were acceptable for the results to be of value in the clinical routine, although there were exceptions due to delays in reports, microscopic examination and supplying relatives with information. In 40% of cases, a microscopic examination had not been performed, which was not surprising in cases where the macroscopic examination was conclusive. We also observed a significant number of situations in which the relatives asked for the autopsy to be performed. This appears to be an increasing trend, which we view as an indication of relatives’ engagement, and which may indicate that relatives want a confirmation of the diagnosis and of how the patient was treated by the healthcare system.^14,26,27,29

One could argue that the findings presented in this work were expected because we examined situations only in which an autopsy was requested by the responsible physician or the relatives, and that these results may not be relevant for deceased patients who are not examined after death. Although we cannot provide data for the deceased who were not autopsied, we believe it reasonable to assume that there are misdiagnoses in these situations as well, likely on the same order of magnitude as seen among the deceased that were examined after death. We base this assumption on the fact that the number of missed or confirmed diagnoses has not changed over the past 50 years and is independent of the autopsy rate. This statement confirms the observations made by Goldman et al. stating that the rate of diagnostic errors has been stable during the last three decades in spite of increased ante-mortem investigations.²⁰ Scordi-Bello et al. also stated that ante-mortem diagnostic evaluation did not affect the levels diagnostic discrepancies in a study comparing data from patients who had died in intensive care units, surgical wards and nursing homes, respectively.²⁸ We are concerned about the value of the death certificate register, which among those that had been autopsied contained missed causes of death in 14% of the cases. We suspect that this number is much higher (75% from the Goldman class I–III cases) for the total register of all people who have died in Sweden. Error rates in death certificates were estimated to be around 30% in the United Kingdom.¹

We conclude that the decrease in the autopsy rate in Stockholm cannot be explained by the hypothesis that autopsies are of less value today compared to when the rate was much higher. We also conclude that the validity of the cause of death on the death certificate is dependent on whether an autopsy was performed. Hospital routines have been sufficient to allow the nursing personnel to perform their duties to the government and to relatives. If the autopsy routines from the pathology department have been sufficient, the diagnostic support has been substantial, and the interest shown by the relatives to the deceased has been expressed, why has the autopsy rate been so low and continued to decrease? We¹⁴ and others^30,31 believe that the autopsy rate would increase if both healthcare personnel and relatives of the deceased had more information regarding the autopsy process and its value. Hooper and Geller⁷ and Liao and Singh³² also confirm that the propensity of the responsible physician to request autopsy is dependent on the earlier experiences of autopsies during education and clinical training and also on training in the process to collect family consent and to file the request. Liao and Singh³² and Hooper and Geller⁷ also propose automatic routines in the electronic health records to remind physicians to consider autopsy and suggests that nurses and other non-physicians could also have that role.

Autopsy experience in the medical education and training of young physicians is significantly less than 20 years ago. About half of the medical doctors employed in Sweden today were educated in other countries, and many of these doctors have had no or very limited experience in post-mortem examinations. In routine healthcare, where autopsies are rare, the experience of the pathologists has also decreased, threatening the value of the autopsy. We and other authors have noticed that there is a high and growing interest among nurses to be included in the information and consent process with next of kin and other relatives.³⁰ Nurses often have closer and more frequent contact with family members and can present the possibility of a post-mortem examination, preferably before the death of the patient but when it is clear that the patient is going to die.¹⁴ We suggest that educational programmes for nurses include training in handling cases of terminally ill patients, with an emphasis on understanding the value of autopsies, especially for the next of kin. In the national guidelines for treating strategies in intensive care, where the patients as well as their closest relatives have to consider whether to interrupt treatment or decide not to initiate treatment to prolong life, or to prepare for organ and tissue donation, there are recommendations on how to use information and decide what to do. In this context, the relatives and the next of kin are involved early in the process, and decisions are made by the doctor together with the kin and nursing personnel. We suggest that these recommendations also become guidelines for how to decide on the benefits of post-mortem examination. Hull et al. have suggested that this information should be included in an instructional brochure designed for families to help the next of kin make an informed decision as to whether the deceased should undergo an autopsy.³³

Clinical usefulness

In this work, which is a follow-up study on a recent survey study from the same hospital on the attitudes of nurses and physicians towards autopsy and the decreasing tendency to use post-mortem examination, we confirmed the value of autopsies to gain information about the cause of death for the relevance of the death certificate as well as to inform the next of kin. We believe that the use of autopsies should increase and we discuss the role of information to relatives and education of physicians and nurses to reach this goal.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

The authors declare that there is no conflict of interest.

References

Wong

Osborn

Waldmann

. Autopsy and critical care. J Intens Care Soc 2015; 16(4): 278–281.

Winters

Custer

Galvagno

Jr et al.

Diagnostic errors in the intensive care unit: a systematic review of autopsy studies. BMJ Qual Saf 2012; 21(11): 894–902.

Shojania

Burton

McDonald

et al.

The autopsy as an outcome and performance measure. Evidence Report/Technology Assessment 2002; 58: 1–5.

Lindstrom

Janzon

Sternby

. Declining autopsy rate in Sweden: a study of causes and consequences in Malmo, Sweden. J Intern Med 1997; 242(2): 157–165.

Burton

Underwood

. Clinical, educational, and epidemiological value of autopsy. Lancet 2007; 369(9571): 1471–1480.

Burton

Phillips

Covinsky

et al.

The relation of autopsy rate to physicians’ beliefs and recommendations regarding autopsy. Am J Med 2004; 117(4): 255–261.

Hooper

Geller

. Relevance of the autopsy as a medical tool: a large database of physician attitudes. Arch Pathol Lab Med 2007; 131(2): 268–274.

Sinard

. Factors affecting autopsy rates, autopsy request rates, and autopsy findings at a large academic medical center. Exp Mol Pathol 2001; 70(3): 333–343.

Veress

Alafuzoff

. A retrospective analysis of clinical diagnoses and autopsy findings in 3,042 cases during two different time periods. Human Pathol 1994; 25(2): 140–145.

10.

Pritt

Hardin

Richmond

et al.

Death certification errors at an academic institution. Arch Pathol Lab Med 2005; 129(11): 1476–1479.

11.

McPhee

Bottles

et al.

To redeem them from death: reactions of family members to autopsy. Am J Med 1986; 80(4): 665–671.

12.

Gaensbacher

Waldhoer

Berzlanovich

. The slow death of autopsies: a retrospective analysis of the autopsy prevalence rate in Austria from 1990 to 2009. Eur J Epidemiol 2012; 27(7): 577–580.

13.

McManus

Suvalsky

Wilson

. A decade of acceptable autopsy rates: does concordance of clinician and pathologist views explain relative success? Arch Pathol Lab Med 1992; 116(11): 1128–1136.

14.

Mjörnheim

Rosendahl

Eriksson

et al.

The attitudes of nurses and physicians about clinical autopsy in neonatal and adult hospital care: a survey in Sweden. Nurs Res 2015; 64(4): 264–271.

15.

The National Board of Health and Welfare. Causes of death 2012, http://www.socialstyrelsen.se/statistics/statisticaldatabase/help/causeofdeath (2013, accessed 4 October 2016).

16.

Britton

. Diagnostic errors discovered at autopsy. Acta Med Scand 1974; 196(3): 203–210.

17.

Britton

. Clinical diagnostics: experience from 383 autopsied cases. Acta Med Scand 1974; 196(3): 211–219.

18.

Tobiano

Marshall

Bucknall

et al.

Patient participation in nursing care on medical wards: an integrative review. Int J Nurs Stud 2015; 52(6): 1107–1120.

19.

Björvell C. Sjuksköterskans journalföring: en praktisk handbok. Lund: Studentlitteratur, 2000.

20.

Goldman

Sayson

Robbins

et al.

The value of the autopsy in three medical eras. N Engl J Med 1983; 308(17): 1000–1005.

21.

Battle

Pathak

Humble

et al.

Factors influencing discrepancies between premortem and postmortem diagnoses. JAMA 1987; 258(3): 339–344.

22.

Kuijpers

Fronczek

van de Goot

et al.

The value of autopsies in the era of high-tech medicine: discrepant findings persist. J Clin Pathol 2014; 67(6): 512–519.

23.

Schwanda-Burger

Moch

Muntwyler

et al.

Diagnostic errors in the new millennium: a follow-up autopsy study. Mod Pathol 2012; 25(6): 777–783.

24.

Frohlich

Ryan

Murphy

et al.

Are autopsy findings still relevant to the management of critically ill patients in the modern era?

Critic Care Med 2014; 42(2): 336–343.

25.

Oluwasola

Fawole

Otegbayo

et al.

The autopsy: knowledge, attitude, and perceptions of doctors and relatives of the deceased. Arch Pathol Lab Med 2009; 133(1): 78–82.

26.

Oppewal

Meyboom-de Jong

. Family members’ experiences of autopsy. Fam Pract 2001; 18(3): 304–308.

27.

Tsitsikas

Brothwell

Aleong

JAC

et al.

The attitudes of relatives to autopsy: a misconception. J Clin Pathol 2011; 64(5): 412–414.

28.

Scordi-Bello

Kalb

Lento

. Clinical setting and extent of premortem evaluation do not predict autopsy discrepancy rates. Mod Pathol 2010; 23(9): 1225–1230.

29.

Bassat

Castillo

Alonso

et al.

Resuscitating the dying autopsy. PLoS Med 2016; 13(1): e1001927.

30.

Dziobon

Roberts

Benbow

. Attitudes of nursing staff to the autopsy. J Adv Nurs 2000; 32(4): 969–974.

31.

Souder

Terry

Mrak

. Autopsy 101 (CE). Geriatr Nurs 2003; 24(6): 330–337.

32.

Liao JM and Singh H. Reviving the autopsy as a diagnostic error-reduction tool. Lab Med 2013; 44(2): 186–190.

33.

Hull

Nazarian

Wheeler

et al.

Resident physician opinions on autopsy importance and procurement. Human Pathol 2007; 38(2): 342–350.

The clinical value of autopsies at a university hospital in Sweden

Abstract

Keywords

Introduction

Materials and methods

Design

Materials

Ethics

Findings

Discussion

Clinical usefulness

Footnotes

Funding

Conflict of interest

References