Comparing laboratory surveillance with the notifiable diseases surveillance system in South Africa

•The South African laboratory and notification surveillance systems were compared.•For the three tracer diseases, fewer cases were notified than laboratory confirmed.•Completeness, stability, and representativeness were higher for the laboratory system.•The sensitivity of the notification system was low for one of the tracer diseases.•The positive predictive value of the notification system was acceptable for the three tracer diseases.

SummaryObjectiveThe aim of this study was to compare laboratory surveillance with the notifiable diseases surveillance system (NDSS) in South Africa.MethodsData on three tracer notifiable diseases - measles, meningococcal meningitis, and typhoid - were compared to assess data quality, stability, representativeness, sensitivity and positive predictive value (PPV), using the Wilcoxon and Chi-square tests, at the 5% significance level.ResultsFor all three diseases, fewer cases were notified than confirmed in the laboratory. Completeness for the laboratory system was higher for measles (63% vs. 47%, p < 0.001) and meningococcal meningitis (63% vs. 57%, p < 0.001), but not for typhoid (60% vs. 63%, p = 0.082). Stability was higher for the laboratory (all 100%) compared to notified measles (24%, p < 0.001), meningococcal meningitis (74%, p < 0.001), and typhoid (36%, p < 0.001). Representativeness was also higher for the laboratory (all 100%) than for notified measles (67%, p = 0.058), meningococcal meningitis (56%, p = 0.023), and typhoid (44%, p = 0.009). The sensitivity of the NDSS was 50%, 98%, and 93%, and the PPV was 20%, 57%, and 81% for measles, meningococcal meningitis, and typhoid, respectively.ConclusionsCompared to laboratory surveillance, the NDSS performed poorly on most system attributes. Revitalization of the NDSS in South Africa is recommended to address the completeness, stability, and representativeness of the system.