All-cause mortality trends in Dikgale, rural South Africa, 1996-2003

Aims: The Dikgale Demographic Surveillance System (DDSS) site, established in 1995, is one of three rural surveillance sites in South Africa. This paper describes detailed mortality patterns of a rural African population in the central region of Limpopo Province. Methods: These data were based on yearly household visits to collect data on vital events, covering 63, 873 person-years of observation over eight years. Results: Crude mortality was 7.5 per 1,000 person-years (females: 6.9, males: 8.1). Under-1 year and under-5 years mortality was 15.1 and 5.8 per 1,000 person-years, respectively. Life expectancy at birth was 64.3 years ( females: 68.1, males: 60.0). For the two four-year periods (1996-9 and 2000-3) under-20 years mortality risk decreased (rate ratio=0.45, 95% CI: 0.25 to 0.80) while 20-49 years mortality risk increased (rate ratio=1.55, 95% CI: 1.10 to 2.20). Multivariate mortality risk for migrants remained relatively constant (0.71, 95% CI: 0.54 to 0.94) across the two four-year periods, but has increased 2.5-fold in all DDSS +50 year-old adults across the two four-year periods. Conclusions: The DDSS mortality estimates appear to have remained relatively constant while recent mortality estimates for the Agincourt Demographic and Health Surveillance System (ADHSS) site suggest that mortality risk is higher and life expectancy is lower in ADHSS residents. Moreover, DDSS mortality estimates are substantially more favourable compared with provincial and national mortality estimates.