tions may be too raw to conduct an extensive interview prior to 2 months after the death. Longer than 6 months after the death, many informants want closure on the suicide and no longer are willing to open up and discuss emotionally difficult topics. The quality of information, measured by the number of diagnoses generated, did not vary as a function of the amount of time since the death (Brent et al., 1988). Caution should be used when interpreting information gathered from friends and relatives; one experimental study found that subjects’ descriptions of psychological distress varied with characteristics of the deceased and aspects of the manner of death (Telcser, 1996). Use of a comparison group of individuals who died accidentally by similar means could strengthen validity of findings. In general, when case-control methods are used in psychological autopsy studies, the comparisons are made to individuals who died by natural causes matched on demographic variables or psychiatric diagnoses.
The psychological autopsy has many similarities to the Family History-Research Diagnostic Criteria or any other indirect interview. The interview is less informative than a direct interview (Andreasen et al., 1977; 1986) but improves with the number of informants. Certain informants may provide specific information that may not be available from others. For example, friends of adolescent suicide victims may be more aware of substance use and abuse than parents (Brent et al., 1988). Employers and co-workers may be able to describe the victim’s functional ability on the job; for younger victims, interview of teachers and review of school records may play an analogous role.
Certain types of information are very difficult, or even impossible, to obtain with a psychological autopsy approach. For example, sexual orientation is information that the victim may have been subliminally aware of, or may not have confided to a friend or parent. Information processing style, or other laboratory-based measures obviously cannot be obtained without the victim’s self-report. However, psychological autopsy studies can help to identify living individuals whose characteristics closely resemble suicide victims who can then be studied using more dynamic assessments.
Combining biological findings with information about psychopathology, personality, family history, treatment history, and history of family adversity may provide a much more complete picture about the neurobiology of suicidal behavior. For example, altered serotonin in the brain may be a consequence of adverse rearing environments (Kaufman et al., 1998; Kraemer et al., 1989; Pine et al., 1997), and may very well be a consistent finding across different mental disorders. Conversely, psychological autopsy data may allow for the selection of relatively homoge-