Similarly, autopsies among deaths due to external causes of undetermined intent increased from 79% to 84%. The frequency of autopsy among homicide deaths was consistently high over this period (between 96% and 97%). The number of autopsies for deaths of unknown or unspecified cause fluctuated between 28% and 32%.

The distribution of cause of death for all autopsies has changed. In 1980, natural deaths accounted for 70% of all autopsies. By 1985, natural deaths accounted for 66% of all autopsies.

Autopsies for natural deaths or deaths occurring among patients under the care of a physician are usually performed at the hospital where the death occurred and with the permission of the decedent's next of kin. If the death is sudden, unexpected, or due to external causes, local statutes may require an autopsy. This autopsy is either requested by a coroner or performed by a medical examiner, depending upon the local medicolegal system. Since deaths due to other than natural causes require medicolegal investigation in most states, the number of autopsies performed was examined by type of medicolegal jurisdiction in the state.

In 1980, 15 states had coroner systems; 18 states and the District of Columbia had medical examiner systems, and 17 states had both medical examiner systems and coroner systems (2). Approximately 44% of all deaths during the period 1980-1985 occurred in states with both medical examiners and coroners (a mixed medicolegal system); 29% of deaths occurred in states with a medical examiner system; the remaining 27% occurred in states with a coroner system. The percentage of deaths in which an autopsy was performed during this 6-year period was greatest among states with a mixed medicolegal system, 16% (Table 2). States with a medical examiner system had autopsies performed in 15% of deaths and states with coroners, 14%. States with a coroner system had the highest proportion of death records that did not indicate whether an autopsy was performed (16%), and states with mixed systems had the smallest (10%).

FIGURE 1. Percentage of deaths involving autopsy, by cause of death United States, 1980-1985

Autopsy Frequency - Continued

When autopsy frequency was examined by medicolegal system and cause of death, states with a medical examiner system had the highest autopsy frequency for deaths due to unintentional injuries and poisoning, homicide, suicide, undetermined intent, and unknown causes (Figure 2). States with mixed systems had higher autopsy frequencies for deaths due to unintentional injuries and poisoning, suicide, and undetermined intent than did states with coroners. The same pattern of annual trends was observed for each medicolegal system (Figure 1).

Reported by: Surveillance and Programs Br, Div of Environmental Hazards and Health Effects, Center for Environmental Health and Injury Control.

Editorial Note: Death certificates are the principal source of mortality statistics for the United States. Several studies, however, have raised questions concerning the accuracy of the recorded cause of death (3,4), and some investigators have advocated improving these statistics by performing more autopsies. Current data show a decline in the proportion of autopsy for natural causes of death (1) and an increase in autopsy proportions for medicolegal deaths (homicides, suicides, and deaths caused by unintentional injuries and poisoning). As a result, 34% of autopsies performed in 1985 involved deaths due to other than natural causes, compared to 30% of autopsies performed in 1980.

State and local laws vary, but medical examiners and coroners typically have the legal authority to order autopsies for traumatic, sudden, or unexpected deaths. A more accurate picture of the frequency of autopsy among deaths outside of the medicolegal system would require separating the sudden or unexpected deaths from other natural deaths. Such an analysis might reveal even lower autopsy frequency for natural deaths occurring in health-care facilities.

Death certificates that fail to indicate whether an autopsy was performed represent another potential source of bias. Larger autopsy percentages were associated with more complete data (Table 1). If deaths with missing data involved autopsy less

TABLE 2. Percentage of deaths involving autopsy, by medicolegal death investigation systems United States, 1980-1985

*Death certificate does not indicate whether an autopsy was performed. 'Deaths with unknown autopsy status not included in denominator.

Autopsy Frequency - Continued

frequently than deaths with complete data, the overall proportion of deaths with autopsy would be lower than the results reported here. This bias would have the greatest effect on natural deaths because that group had the highest proportion of missing data.

States with mixed medicolegal systems have some counties with coroners and others with medical examiners. Jurisdictions with medical examiners are often densely populated urban counties, whereas jurisdictions with coroners are often less populated rural counties. To determine if the higher autopsy frequencies are related to the urban settings or to the presence of medical examiners, each medicolegal jurisdiction and the accompanying population characteristics must be identified.

Finally, the quality and detail of the autopsy protocols - some are only partial and may not include toxicological testing - should be considered when assessing whether autopsy alone is indicative of a more accurate determination of the cause of death. FIGURE 2. Percentage of deaths involving autopsy in states with medical examiners or coroner jurisdictions, by cause of death, United States, 1980-1985

1. Kircher T, Nelson J, Burdo H. The autopsy as a measure of accuracy of the death certificate. N Eng J Med 1985;313:1263-9.

2. Health Services Administration. Death investigation, a synopsis and analysis of laws (including sudden infant death legislation) in 56 US jurisdictions, 1980. 2 vols. Washington, DC: US Department of Health and Human Services, Public Health Service, 1981. Final report for contract HSA240-80-0027.

3. Scottolini AG, Weinstein SR. The autopsy in clinical quality control. JAMA 1983;250:1192-4. 4. Battle RM, Pathak D, Humble CG, et al. Factors influencing discrepancies between premortem and postmortem diagnoses. JAMA 1987; 258:339-44.

5. Council on Scientific Affairs, American Medical Association. Autopsy: a comprehensive review of current issues. JAMA 1987;258:364-9.

Notice to Readers

Short Course in Field Epidemiology and Biostatistics

The second annual Short Course in Field Epidemiology and Biostatistics for physicians will be held in Thailand from June 6 through July 29, 1988. The course is sponsored by the Thailand Field Epidemiology Training Program (FETP). The first 4 weeks will include classroom lectures on basic epidemiological and statistical skills and practical exercises in disease surveillance, epidemic investigation and control, and applied field research. The last 4 weeks will consist of supervised field work in disease surveillance.

The course will be conducted in English. Registration should be made by April 30. For information, contact Dr. Khanchit Limpakarnjanarat, FETP, Division of Epidemiology, Ministry of Public Health, Bangkok 10200 Thailand (telephone: 66+2+2811479; cable: MINHELTH BK 10200; Dialcom: 132: PHF49808; U.S. telex: 4909945409 PHN UI).

The Morbidity and Mortality Weekly Report is prepared by the Centers for Disease Control, Atlanta, Georgia, and available on a paid subscription basis from the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402, (202) 783-3238.

The data in this report are provisional, based on weekly reports to CDC by state health departments. The reporting week concludes at close of business on Friday; compiled data on a national basis are officially released to the public on the succeeding Friday. The editor welcomes accounts of interesting cases, outbreaks, environmental hazards, or other public health problems of current interest to health officials. Such reports and any other matters pertaining to editorial or other textual considerations should be addressed to: Editor, Morbidity and Mortality Weekly Report, Centers for Disease Control, Atlanta, Georgia 30333.