Quality Management

The Autopsy in Quality Management of Clinical Care

Despite advances in imaging techniques and laboratory diagnosis, an autopsy is still the most accurate way to identify the cause(s) of death. In fact, even as medicine becomes more advanced technologically, the autopsy continues to disclose undiagnosed conditions contributing to the cause of death. A survey of discrepancies between clinical and postmortem findings reported since the early twentieth century is shown in Table 15-1 . Although the methods used by the investigators vary, the studies indicate that the autopsy continues to be a useful tool for quality assessment. The coding system of Goldman and coworkers as modified by Battle and associates ( Table 15-2 ) can be used effectively in autopsy-based quality assessment programs. Shojania and colleagues reported an excellent systematic comparison of clinical errors detected by autopsy from 53 published studies over four decades leading up to 2002. They concluded that (1) the rate of major errors (involving a primary cause of death) decreased approximately 20% each decade; (2) the underlying major error rate in 2003 was estimated to average approximately 8%; and (3) institutions with low autopsy rates tend to select for cases more likely to show major errors, but clinicians cannot reliably predict which autopsies will be of high diagnostic yield.

Postmortem examinations help identify misdiagnosis or unexpected findings in patients dying from trauma, cancer, or sudden death ; after surgery ; during the perinatal period or when elderly ; or while receiving care in the emergency department, intensive care unit, or psychiatric hospital. The most recent autopsy data show that the most common missed diagnoses are vascular events—such as pulmonary embolism, ruptured aneurysm, and myocardial infarction—and infectious diseases, most often pneumonia. The data from autopsy studies also may have direct and indirect effects on the living. Faye-Petersen and colleagues found that fetal and perinatal autopsies altered genetic counseling and recurrence risk estimates in 26% of cases in which an autopsy was performed. The autopsy remains the standard for assessing the diagnostic accuracy of imaging studies. In addition, in studies too numerous to cite, postmortem examination continues to play an important role in identifying or evaluating therapeutic complications.

In relation to its role in the medical audit, the autopsy complements clinical and laboratory medicine by providing definitive diagnoses. It follows, therefore, that the autopsy should play a significant role in hospital quality management programs. In fact, postmortem examination provides the framework for assessing not only the quality of clinical diagnoses but also the effects of a chosen therapy. Traditionally, autopsy findings have been associated mostly with quality assurance, that is, the identification of poor outcomes in order to correct individual performance, but increasingly there is a call for autopsy to be involved more with quality improvement programs that aim to improve outcomes for all patients within the institution by changes in policy, procedures, and guidelines. To date, use of autopsy in quality improvement has related mostly to evaluation of effectiveness of medical, surgical, and device-based therapies, but this will be expanded in the future as the College of American Pathologists (CAP) requires that “findings from autopsies are incorporated into the institutional quality management program. Some examples of this could include: 1) reporting newly diagnosed infectious diseases to the hospital infection prevention committee, 2) presentation and/or review by institutional quality assurance committees, 3) reporting issues related to quality of care to risk management or sentinel event review committees.”

Quality Assurance of the Autopsy

The Autopsy Committee of the CAP has published helpful practice guidelines for autopsy pathology. The committee promotes improvement in autopsy diagnosis through its performance improvement program for participating laboratories, and CAP has assembled a program for quality management in anatomic pathology that includes recommendations for autopsy performance improvement. Others have also outlined quality control and assurance programs for autopsies. However, the accuracy of autopsy diagnosis has rarely come under serious scrutiny. From a survey of hospital clinicians, Fowler and colleagues found that with respect to answering physicians' questions, autopsies were judged satisfactory in 83%, partially satisfactory in 12%, and unsatisfactory in 2% of examinations. A prospective audit of autopsy performance revealed that in 3 of 99 cases in which a clinician's questions were not answered definitively, the root cause was failure of the prosector to do a thorough postmortem dissection. In a study of 125 autopsies performed at a university teaching hospital, Bayer-Garner and colleagues found that the tasks most frequently asked of the autopsy pathologists by the clinical staff were to identify the pathologic conditions responsible for the clinical picture, to confirm a suspected diagnosis, and to determine the cause of death. Of 100 questions asked before these autopsies, 78.6% were answered in the final anatomic diagnosis, 9.7% were addressed in some part of the autopsy report, 9.7% were not addressed at all, and 1.9% could not be answered by the autopsy report. Veress and colleagues compared interobserver variability among four pathologists at the same institution with respect to macroscopic diagnoses and immediate cause of death. Although they found a high level of agreement on major diagnoses, there were moderate discrepancies in assigning the immediate cause of death and in identification of minor diseases.

Zarbo and coworkers reviewed completeness of information on autopsy consent forms, time intervals between patients' deaths and postmortem examination, and turnaround time of reports listing preliminary autopsy diagnoses. Although their study identified fairly consistent documentation on autopsy permit forms and greater than 80% compliance with the CAP's laboratory standards for completion of preliminary autopsy reports, a wide range of time intervals was found between patients' deaths and start of the postmortem examination. A similar study of final autopsy report turnaround time revealed that 47.6% were completed in 30 days or less, 28.8% in 31 to 60 days, 12.2% in 61 to 90 days, and 11.5% in more than 90 days. Thus, even if all of these cases were to be considered complicated, an astonishingly high number of cases exceeded the 60-day (Joint Commission) and even the 90-day (CAP) guidelines of the two major laboratory-certifying organizations of the United States.

It is unlikely that autopsy practice will receive any additional financial support or direct reimbursement by third-party payers. Furthermore, reinstatement of minimal autopsy requirements for hospital, residency, and managed care plan accreditation is not on the horizon. Thus the impetus for improving autopsy practice will remain exclusively the responsibility of pathology departments and pathologists who practice in the community, at a medical college, or at the local coroner's or medical examiner's office. Fortunately, autopsy pathologists and their institutions can accomplish significant quality assurance by following straightforward policies and ensuring adequate communication with clinicians, next of kin, and hospital quality-of-care committees ( Boxes 15-1 and 15-2 ). Additionally, many of the requirements for an autopsy service are apparent from the CAP lab accreditation checklist ( Box 15-3 ). The CAP requirements for retention of materials are shown in Table 15-3 , but many autopsy services will choose to retain dry materials and records for much longer, and will retain gross photographs for the duration of the report retention. We retain autopsy consents indefinitely so that they may provide provenance for materials that are used for education or research purposes.

The first step in performing an autopsy is the review of the medical record. The pathologist must obtain a clear understanding of the patient's present and past medical history, therapeutic interventions, and the terminal events leading to the patient's death. Review of the medical record should include a review of relevant laboratory data, reports of imaging studies, and surgical pathology or cytopathology reports or material. Before beginning the prosection, the pathologist must attempt to contact the patient's primary physician or physicians. The clinician usually has extensive knowledge of the patient developed over days, weeks, or even years of contact. It cannot be overemphasized how useful a brief discussion of the case with the treating clinicians can be in clarifying the clinical thought processes during the management of the patient's care. Discussion of the case with the clinician is part of the role of the pathologist as consultant. This provides the pathologist with an opportunity to ask the clinician about unresolved issues and concerns that arose during the patient's illness.

A thorough understanding of the clinical issues and questions allows the pathologist to anticipate the need for special studies that might elucidate anatomic diagnoses or best demonstrate pathologic findings. Is there a need for electron microscopy or immunofluorescence? Should any tissues or fluids be submitted for microbiologic cultures or collected for biochemical analysis? Should the pathologist alter his or her usual prosection technique or dissect an anatomic site not normally examined? Would radiography or angiography enhance the examination? The pathologist and his or her staff can prepare needed equipment and supplies before beginning the case.

After completing the prosection, the pathologist should communicate his or her findings to the patient's clinicians. The clinical team should be invited to visit the autopsy suite during or after the autopsy. Regrettably, attendance of clinicians at autopsies is often prohibited by other demands on their time. Recognizing this, the pathologist should be flexible in meeting with the clinical staff. Active dialog with the physicians who cared for the patient during life allows the autopsy pathologist to make a greater contribution. Questions remaining after the gross autopsy can be addressed by engaging the clinicians again during a review of the microscopic findings and reassessment of the case as a whole. The clinician's knowledge of the pathophysiology of a specific case can help the pathologist interpret the morphologic findings more accurately.

The autopsy pathologist is a consultant specialist. As such, the pathologist must report the anatomic findings to the responsible clinician in a timely manner. The preliminary autopsy report should be available within 48 hours; however, sooner—for example, the same day—is better. We feel that the final autopsy report, including neuropathologic findings, should be completed in 30 calendar days or less. This falls under both CAP and Joint Commission guidelines and regulations for routine cases. For a few rare cases requiring special studies that delay the autopsy report, we suggest that an interim report (including a statement specifying pending studies and/or consultations) be completed by 30 days and that the subsequent findings and additional interpretations be communicated as soon as possible.

The clinician is in the best position to interpret the autopsy for the family, taking the anatomic data and placing them in a clinical context. Many clinicians meet with the family to explain the autopsy report. Nevertheless, some families may contact the pathology department directly to discuss autopsy results. If so, most institutions would direct the family to the medical records department to receive a copy of the report and would encourage the family to reach the clinician to discuss the contents. At the same time, the attending clinician would be notified and supplied with an additional copy of the autopsy report and the address and telephone numbers of the next of kin so that contact can be initiated. Occasionally, pathologists are asked to discuss the autopsy findings with family members, and in those cases it is best either to record the conversation for the medical record or make a note of the conversation in the pathology records.

Peer review can serve as an easy check on autopsy quality. At academic medical centers this is part of the teaching program; autopsy faculty review findings at a daily or weekly gross conference, or “organ recital.” Periodic conferences held at a multiheaded teaching microscope corroborate, elucidate, or correct gross pathologic specimen interpretations and provide a forum for discussion of the most significant clinicopathologic correlations. In the community, pathology groups can identify an appropriate time for peer review of difficult autopsy cases. For autopsy services that perform autopsies frequently, departmental quality assurance activities usually include retrospective review of a random fraction of cases by qualified peers.

Quality Improvement of the Autopsy