Abstract

National vital registration systems that record causes of death do not exist in most African countries. This makes estimating incidence of mortality from injuries in African regions particularly challenging. We discuss the viability of using retrospective and prospective data gathered from death registration sites that issue death certificates for medicolegal purposes in four African cities (Lusaka, Zambia; Maputo, Mozambique; Addis Ababa, Ethiopia; and Kampala, Uganda). We describe the flow of information in the form of administrative registers and documents maintained and issued by various institutions involved with deaths (including medical institutions, forensic, police and judicial investigation records, and funeral grounds and cemeteries). In each case, we show the implications of varying social, cultural and administrative record keeping practices and identify the most complete site for collecting cause-of-death information.

Next, we discuss methods for converting such data into reliable estimates of injury incidence. This requires addressing data quality issues, and estimating completeness and coverage of these data systems. We demonstrate how cause of death attribution can be improved in retrospective data gathering by following back a stratified sample of registered cases. For prospective data gathering, we recommend the adoption of a standardised surveillance instrument (currently being developed by WHO) for prospective data gathering. Finally, we demonstrate the method of estimating completeness and coverage by comparing recorded deaths with deaths predicted in the population using alternate sources.