The autopsy as a quality assurance tool: Last rites or resurrection?

The autopsy has been called the ultimate medical consultation. Particularly with the recent focus on

Uses of the autopsy

Most people are familiar with the forensic autopsy, authorized by law in situations where a death is sudden, violent, or suspicious, or when the cause of death cannot otherwise be established. This article will focus on nonforensic autopsies.

Although most people think the nonforensic autopsy is performed to determine the cause of death, the uses of the autopsy are in fact numerous. Some of these many uses are:

• To confirm and clarify clinical diagnoses.
• To evaluate the accuracy of diagnostic procedures, such as computerized tomography (CT), nuclear magnetic resonance, and positron emission tomography (PET) scans. 
• To evaluate the efficacy and potential adverse effects of new drugs, new surgical techniques, prosthetic devices, and genetic engineering.
• To aid in the discovery of new or previously unrecognized diseases (e.g., Legionnaires disease, toxic shock syndrome, and AIDS).
• To provide reassurance and information for surviving family members. 
• To provide information for medical and epidemiologic research.
• To facilitate investigation of environmental, occupational, and lifestyle-related diseases.
• To provide tissues for research and transplantation.
• To teach medical students and residents. 
• To influence expenditures on medical research, training, and disease control.¹

The rise and fall of the autopsy

The autopsy has a long history that has been influenced by the prevailing medical models of the time. In Classical Greek times, there was little medical interest in the autopsy since diseases were believed to result from the imbalance of theoretical humours, which were not considered to have a simple anatomic basis.² Since there was no particular need for looking into the body for the cause of death, autopsies were performed for simply academic reasons.

In Anglo-Saxon England, the prevailing belief was that disease was due to magical causes and that the power of this magic might still reside in the corpse. Further contact with the body might be dangerous, and ones energy would be much better directed toward identifying and neutralizing the evil person who had initiated the magic.²

Only in the early days of modern medicine was the basis of disease understood to be anatomical and physiological. The golden era of the autopsy occurred during the 19th century, when tens of thousands of autopsies on adults provided the database that underlies modern adult medicine.³ Somewhat later, the medical science of pediatrics developed in a similar way, with much of that knowledge gained from pediatric autopsies. More recently, the specialties of neonatology and perinatology have benefited from autopsies on infants, neonates, and fetuses.³ 

In the 1800s, before medicine became specialized, clinicians performed autopsies on their own patients and were motivated simply by their own interest in anatomic findings that would verify their diagnosis and evaluate treatment. In first half of the 20th century, the autopsy became the province of pathologists.

Autopsy rates have plummeted from approximately 50 percent in the 1940s to around 14 percent in the 1980s.^1,4 Recent estimates suggest that autopsy rates are currently less than 5 percent in many hospitals.⁵ Because law mandates forensic autopsies, the rates for forensic autopsies have not declined.

In the early 1970s, the Joint Commission on Accreditation of Healthcare Organizations (formerly the Joint Commission on Accreditation of Hospitals) eliminated its recommendation for a minimum autopsy rate of 20 percent. JCAHO recognized that no single autopsy percentage rate could accommodate the very different needs of community hospitals, teaching hospitals, and academic referral centers.^1,4,6 Interestingly, however, the decline in autopsy rates appears to have been underway before this requirement was dropped.⁶ Currently, there is no good definition of what constitutes an adequate autopsy rate.

Reasons for decline of the autopsy

While many people blame the demise of the autopsy on JCAHO, the real reasons are much more complex. A variety of societal, medical, and technological factors are responsible for the decline of the autopsy, and it is clear that there is no single cause, or even a principle cause, for this decline.^1,3,7 

Economic considerations.
The costs of autopsy services are supported through general operating expenses, and with some exceptions, autopsies are performed by pathologists as part of their professional service contract with the hospital. 

Autopsies are expensive and include the costs of facility maintenance, such as space, overhead, personnel, supplies and equipment. An autopsy requires four to 10 hours of a pathologists time, as well as three to six hours of effort on the part of autopsy assistants, histologists, and secretaries.⁶ The autopsy also entails laboratory costs, including
histopathology, and if relevant, toxicology, chemistry, bacteriology, and serology. During the late 1970s and early 1980s, autopsy costs were estimated from $900 to $2,000, depending upon the procedures performed.¹ The costs were estimated to consist of approximately 47 percent direct personnel expenses, 40 percent for supplies and services, and 13 percent for space, equipment, depreciation, and indirect costs.¹ 

Third-party payers do not directly reimburse either hospitals or pathologists for autopsies. Medicare Part A (Hospital Insurance Program) makes no specific provisions for financing the autopsy.6 Autopsies are viewed as an overhead expense, and payment is bundled into hospital payments.⁸ Medicare Part B (Supplementary Medical Insurance for Physician Services) excludes reimbursement for pathologists because autopsies are not considered a service related to the diagnosis and treatment of a given patient.⁶

From an economic point of view, a critical factor in determining the appropriate rate of autopsies is marginal cost. Because the variable costs associated with autopsies are low, virtually all the costs of an autopsy service are fixed and unrelated to volume.¹ While the autopsy can be considered a public good, with benefits that are generally intangible and widely distributed, the actual monetary costs of performing autopsies fall directly upon hospitals and are difficult to recoup. This may contribute to the market failure of autopsies from an economic viewpoint.¹

Fear of litigation. Fear of litigation or embarrassment may discourage clinicians from ordering autopsies. Soaring litigation has created the misconception among both the public and some physicians that fallibility is unacceptable and that all errors are avoidable.⁵ This mindset discourages physicians from requesting autopsies for fear of disclosure of errors.⁵

Some research suggests that autopsies reduce the risk of financial loss from malpractice suits. Autopsies eliminate suspicion and provide reassurance to families. Autopsy findings replace conjecture with facts and allow defendants to construct a better defense. Autopsies have also been shown to reduce the number of capricious malpractice legal actions.¹ 

For medical performance to improve, the nature and causes of failures must be studied for clues that can help improve medical care in the future.⁹ Medicine is by its nature imperfect, and society must come to understand and accept a realistic degree of fallibility.⁹ 

Pathologists. We have met the enemy, and he is us. Many, if not most, pathologists are quite vocal about their own dislike of autopsies. Autopsies are often perceived as unscheduled nuisances that may extend into evenings or occur on weekends. Autopsies take the pathologist away from the directly reimbursable work of interpreting surgical and clinical laboratory specimens. Autopsies are lengthy, time consuming, dirty jobs, often unappreciated, and generally thankless.³ Finally, in todays litigious environment, when a negative unexpected finding is discovered at autopsy, the pathologist must possess the skills and tact of a diplomat.⁶ These unexpected findings may also result in further sacrifice of the pathologists time by having to spend time discussing the findings with the hospitals risk manager, giving depositions, or possibly testifying in court. 

Clinicians. Many clinicians are skeptical of the high levels of discrepancy between clinical and autopsy diagnoses and do not see a major role for autopsy in clinical audit.² Despite the opinions of most professional organizations and the wealth of evidence to support the medical and scientific value of the autopsy, many clinicians feel unable to justify the cost of autopsies within hospitals.²

Some factors considered most influential in a clinicians decision to request an autopsy include patient age, attitudes of relatives, and the clinicians confidence in the clinical diagnosis.² New and sophisticated diagnostic techniques have increased confidence in clinical diagnoses, with the result that autopsies are often considered to be outdated and unnecessary.²

Clinicians may become disenchanted by autopsies due to poor quality autopsy reports from unmotivated pathologists.¹⁰ Burning clinical questions may have lost their relevance due to the long turnaround time of final autopsy reports, which can take weeks to months to complete, depending upon the motivation of the pathologist. 

Finally, the attitudes of clinicians may be influenced by their personal experience as medical students or house staff, and later by the level of interest shown by senior colleagues and pathologists.² A self-perpetuating cycle of low interest in autopsies is created when these influences are consistently negative.

Nonmedical decision makers

Aside from medical personnel and the families of the deceased, other groups may influence autopsies, including hospital administrators, nursing home operators, and funeral directors.¹ 

Hospital administrators also dislike autopsies, since they are responsible for the financial support of the autopsy room, which uses money they could be using on clean linens, nurses, social workers, medicines, and even cafeteria food.³ Administrators generally prefer to spend their money on the living, not the dead.¹¹ Administrators may also dislike autopsies because they prefer not to dwell on the unfavorable outcomes that may occur in their hospitals.¹¹ 

Quality of care varies considerably among nursing homes, and it is possible that those homes with below average care might be less inclined to encourage autopsies lest the findings reflect poorly upon them.¹ 

Funeral directors and pathologists have traditionally had a subtly adversarial relationship. Funeral directors have strong disincentives to autopsy, including the inconvenience of transporting the body to the hospital when the death has occurred elsewhere and delay in preparation of the body for the funeral service.¹ In addition, the embalming procedure is a little more difficult and time consuming in an autopsied body than in an intact body. 

Miscellaneous factors. Miscellaneous factors contributing to the decline of autopsy rates include concerns about potentially transmissible diseases such as AIDS, decreased emphasis on the autopsy in formal medical education, and the increasing number of people who die at home.^1,3,4,5

Quality assurance and the autopsy

Multiple studies have demonstrated the enduring ability of the autopsy to identify clinically important missed diagnoses.^5,9,12,13 While the types of diagnoses missed have changed over the decades, the percentage of missed diagnoses has remained surprisingly constant over the last 90 years, ranging from 15 percent to 80 percent of autopsied cases .^5,12,13,14 Autopsies identify an unexpected cause of death in approximately 20 percent of cases, and in half these cases, correct premortem diagnosis could have prolonged the patients survival.⁴ 

While autopsies have the power to audit some aspects of medical care, no one has yet shown that simply conducting autopsies and publicizing the findings directly benefit anyone.¹⁰ No data have substantiated the widespread belief that autopsies lead to improvements in patient care, which means that the beneficial effect of autopsies on the quality of care remains a matter of faith.^4,10 Most commonly, autopsy reports reside in the obscurity of the pathology and medical records departments, becoming orphan data.¹⁰ 

Accuracy of diagnosis is not the only important factor in the quality of medical care. Therapy, prognosis, patient satisfaction, and outcome are also important. Effectiveness of therapy may be reflected in autopsy findings, but the accuracy of prognosis, patient satisfaction and outcome, are not easily tested at autopsy.¹⁰ 

Autopsies provide a measure of the accuracy of death certificates. In most hospitals, the death certificate is completed before the autopsy, if there is one; consequently, discrepancy rates between death certificate and autopsy diagnoses are in the range of 30 percent to 50 percent.^3,15 Mortality statistics are generally compiled from death certificates and are used to estimate the incidence and prevalence of diseases.^6,15 Inaccuracies in these statistics have the potential to cause misplaced emphasis on research of specific diseases and the creation of irrelevant health policies.⁶

The autopsy itself is not error-free, since pathologists, like any physicians, can and do render erroneous diagnoses. No studies have been performed to determine error rates in autopsy diagnosis.^9,16 Nonetheless, the autopsy is still essential to quality assurance because it establishes a standard for diagnostic accuracy and therapeutic effectiveness.⁴

With the aging of the baby boomers, demographics are shifting toward an older population, and gerontology (geriatric medicine) is still a relatively young medical field. Important and unique attributes of geriatric medicine, completely distinct from adult medicine, could be discovered through autopsy studies, yet multiple studies consistently show an extremely low autopsy rate in the elderly.^1,3,4 The aged are far more likely to suffer from more than one disease at a time, and puzzling interactions occur among coexisting diseases, producing new clinical manifestations and new complications.³ Elderly patients frequently have suffered with chronic diseases for long periods, and the cumulative effects of these long-standing illnesses have not been cataloged.³ Not surprisingly, the rate of error in clinical diagnoses is highest among elderly patients.³ 

The future of the autopsy 

The cost effectiveness of autopsies should be examined carefully. Autopsies are expensive, and establishment of new funding streams that would reimburse costs will be difficult, if not impossible.⁷ Fair reimbursement for both the professional and institutional costs of performing autopsies is needed.⁴ Third-party payers need to recognize the cost of autopsies as a necessary investment for measuring and maintaining quality, not simply as an overhead expense for research and education.⁴ Some have suggested that autopsies be assigned a CPT code with
DRG-approved payment, or that hospitals be allowed to charge a surtax on bed occupancy to help support autopsy costs.⁷

Managed care organizations and other payers will be more inclined to reimburse for the autopsy if and when they clearly perceive both its role and value.⁸ Clinical quality is measured much more systematically in managed care than in more traditional settings, and the emphasis is increasingly on outcomes. The final element in a comprehensive study of prevention, screening, primary and secondary intervention, ambulatory disease management, and hospital management should be an autopsy study of deaths.⁸

Efforts to promote quality assurance in hospitals have generally neglected the autopsy.² The medical community needs to determine whether autopsies truly can be used systematically to measure and improve the quality of medical care, rather than being used simply as a spot-check.⁴ Healthcare providers and organizations need to develop systems which provide feedback and act on autopsy information, and these systems need to be evaluated with measurable benchmarks as in any other quality assurance monitor.⁴ Logistics for implementation of such a system would include how data are collected and reported, and how the information is shared and utilized beyond the practice of pathology (e.g., epidemiology, technology transfer, and imaging).⁷ An independent panel should review autopsy data, with recommendations as appropriate.¹⁰ 

Patients coming to autopsy do not represent a random group from the hospital population, since clinicians are more likely to request an autopsy when the patient has created diagnostic dilemmas during life.¹⁵ For data to be meaningful, a mechanism must exist for obtaining a sufficiently representative sample of hospital deaths to permit statistical reliability.¹⁰

Other proposed steps for enhancing the contributions of autopsy to clinical medicine include: improving the process of obtaining autopsy consent, standardizing the method for classifying unexpected autopsy findings, integration of autopsy findings into educational feedback systems, and research to determine the yield and cost effectiveness of the autopsy in specific clinical situations.⁴

The prognosis for the autopsy is grave, and multiple political and financial forces imperil its future. These difficulties are in large measure the unintended consequences of policies aimed at cost containment. While it is not likely that the autopsy will ever disappear altogether, in the absence of its integration into effective and realistic quality assurance programs and improved reimbursement, it may remain little more than a quaint relic in nonforensic deaths.

Dr. Steigman is a member of the Department of Pathology at Garden City Hospital in Garden City, MI.

References:

1. Nemetz PN, Ludwig J, Kurland LT. Assessing the Autopsy. Amer J Pathol 1987; 128:362-378.

2. Start RD, Cotton
DWK. The meta-autopsy: Changing techniques and attitudes towards the autopsy. Qual Assur Health Care, 1993; 5(4):325-332.

3. Hill RB. The current status of autopsies in medical care in the USA. Qual Assur Health Care 1993; 5(4):309-313.

4. Landefeld CS, Goldman L. The autopsy in clinical medicine (editorial). Mayo Clin Proc 1989; 64:1185-1189.

5. Hasson J. Medical fallibility and the autopsy in the USA. J Eval Clin Pract 1997; 3:229-234.

6. Reichert CM, Kelly
VL. Prognosis for the autopsy. Health Aff Summer 1985; 4:82-92.

7. Setlow VP. The need for a national autopsy policy. Arch Pathol Lab Med 1996; 120:773-777.

8. Chernof D. The role of managed care organizations in autopsy reimbursement. Arch Pathol Lab Med 1996; 120:771-772

9. Anderson RE, Hill
RB, Key CR. The sensitivity and specificity of clinical diagnostics during five decades. Toward and understanding of necessary fallibility. JAMA 1989; 261:1610-1617.

10. Hill RB, Anderson RE. The autopsy crisis reexamined: The case for a national autopsy policy. The Milbank Quarterly 1991; 69:51-78.

11. Lundberg GD. College of American Pathologists Conference XXIX on restructuring autopsy practice for health care reform. Arch Pathol Lab Med 1996; 120:736-738.

12. Goldman L, Sayson R, Robbins S, et al. The value of the autopsy in three medical eras. NEJM 1983; 308:1000-1005.

13. Kirch W, Schafii C. Misdiagnosis at a university hospital in 4 medical eras. Report on 400 cases. Medicine 1996; 75:29-40.

14. Saracci R. Autopsy as the yardstick for diagnosis: An epidemiologists remarks, in ed. Ribolt E & Delendi M. Autopsy in Epidemiology and Medical Research, Lyon, France: International Agency for Research on Cancer; 1991; pp185-196.

15. Peacock SJ, Machin D, Duboulay
CEH, et al. The autopsy: A useful tool or old relic? J Pathol 1988; 156:9-14.

16. Saracci R. Problems with the use of autopsy results as a yardstick in medical audit and
epidemiology. Qual Assur Health Care 1993; 5(4):339-344.

© 2002 Nelson Publishing, Inc. All rights reserved.