A study of autopsy procedures in Ghana: implications for the use of autopsy data in epidemiological analyses

The study of cause of death certification remains a largely neglected field in many developing countries, including Ghana. Yet, mortality information is crucial for establishing mortality patterns over time and for estimating mortality attributed to specific causes. In Ghana, autopsies remain the appropriate option for determining the cause of deaths occurring in homes and those occurring within 48 hours after admission into health facilities. Although these organ-based autopsies may generate convincing results and are considered the gold standard tools for ascertainments of causes of death, procedural and practical constraints could limit the extent to which autopsy results can be accepted and/or trusted. The objective of our study was to identify and characterise the procedural and practical constraints as well as to assess their potential effects on autopsy outcomes in Ghana. We interviewed 10 Ghanaian pathologists and collected and evaluated procedural manuals and operational procedures for the conduct of autopsies. A characterisation of the operational constraints and the Delphi analysis of their potential influence on the quality of mortality data led to a quantification of the validity threats as moderate (average expert panel score = 1) in the generality of the autopsy operations in Ghana. On the basis of the impressions of the expert panel, it was concluded that mortality data generated from autopsies in urban settings in Ghana were of sufficiently high quality to guarantee valid use in health analysis.