Using health and demographic surveillance to understand the burden of disease in populations: The case of tuberculosis in rural South Africa1

Abstract

To utilize the Agincourt health and demographic surveillance system (HDSS) platform to assess the burden of pulmonary tuberculosis (PTB) in a rural South African sub-district. During 1999, data from three sources were combined to estimate disease prevalence amongst a non-migrant adult population: (1) passive case-finding (PCF) through hospital register data; (2) active case finding (ACF) using a systematic household assessment of chronic coughers; and (3) verbal autopsy (VA) data on cause of death. Of 66,840 residents, 38,251 permanent adult residents were included in the analysis. A total of 102 cases of PTB were detected through PCF. ACF sweep detected 366 chronic coughers with 6 cases of confirmed PTB. Among 28 PTB deaths detected by VA, 13 (46%) were not previously identified by the health service. The total PTB prevalence was 157/100,000; 110/100,000 of prevalent cases were detected by PCF. Among undetected cases, 24/100,000 were identified through ACF, while 23/100,000 were detected by the VA process. Amongst prevalent PTB cases in the permanent adult population, 70% were detected by the health service; 15% of cases were undiagnosed in the community, while an equal proportion died of PTB prior to diagnosis. The latter groups contributed disproportionately to infectiousness in the community through prolonged duration of symptoms. As most of these cases presented to the health service on a number of occasions, strengthening early case detection should remain the cornerstone of TB control efforts. Strategies to strengthen the application of health & demographic surveillance systems to disease surveillance are discussed.