Cultural imperatives and the ethics of verbal autopsies in rural Ghana

Due to a paucity of statistics from vital registration systems in developing countries, the verbal autopsy (VA) approach has been used to obtain cause-specific mortality data by interviewing lay respondents on the signs and symptoms experienced by the deceased prior to death. In societies where the culture of mourning is adhered to, the use of VA could clash with traditional norms, thus warranting ethical consideration by researchers.

The study was designed to explore the ethics and cultural context of collecting VA information through a demographic and health surveillance system in the Kassena-Nankana District (KND) of Ghana.

Data were collected through qualitative in-depth interviews (IDIs) with four field staff involved in the routine conduct of VAs, four physicians who code VAs, 20 selected respondents to the VA tool, and eight opinion leaders in the KND. The interviews were supplemented with observation by the researchers and with the field notes of field workers. Interviews were audio-recorded, and local language versions transcribed into English. Thematic analysis was performed using QSR NVivo 8 software.

The data indicate that cultural sensitivities in VA procedures at both the individual and family levels need greater consideration not only for ethical reasons but also to ensure the quality of the data. Discussions of some deaths are culturally prohibited and therefore lead to refusal of interviews. Families were also concerned about the confidentiality of information because of the potential of blame for the death. VA teams do not necessarily engage in culturally appropriate bereavement practices such as the presentation of tokens. The desire by families for feedback on the cause of death, which is currently not provided by researchers, was frequently expressed. Finally, no standard exists on the culturally acceptable time interval between death and VA interviews.

Ethical issues need to be given greater consideration in the collection of cause of death data, and this can be achieved through the establishment of processes that allow active engagement with communities, authorities of civil registrations, and Institutional Review Boards to take greater account of local contexts.