Verbal autopsies: learning from reviewing deaths in the community

The Safe Motherhood Programme of the World Health Organization convened a workshop in 1994 so that investigators from around the world could share their experience in the use of verbal autopsies (postmortem interviews of the relatives and/or neighbors of the deceased) to determine the cause of maternal death and move towards a consensus of what should constitute verbal autopsy methods. Workshop outcomes included the following: 1) a determination that the postmortem interview should consist of verbal autopsy, verbal determination of the nonclinical causes of death, and verbal reporting of background characteristics; 2) agreement that the classification of direct causes of maternal death may be single, dual, or a combination of single and multiple; 3) development of a comprehensive list of the signs and symptoms of obstetric/medical causes of maternal death (direct and indirect); 4) agreement that verbal autopsies should be conducted for all deaths of women of reproductive age to determine early pregnancy deaths; 5) determination that the future development of flowcharts would be valuable for identifying and coding causes of maternal death; 6) identification of information to be included in all questionnaires; and 7) consideration of specific aspects of data collection (requirements for interviewers, respondents, getting information from health providers, the recall period, and validation of specific aspects). Appended to this report are 1) a list of participants, 2) flowcharts for causes of maternal death, 3) suggestions for questions to be asked, 4) suggestions for a questionnaire, and 5) a summary of selected studies that used lay reporting to identify causes of death.