Evaluation and optimization of surveillance systems for rare and emerging infectious diseases

Abstract

Surveillance for rare and emerging infectious diseases poses a special challenge to veterinary services. Most emerging infectious diseases like bovine tuberculosis (bTB) are zoonoses, affecting both human and animal populations. Despite the low prevalence of such an emerging infectious disease at time of incursion, the surveillance system should be able to detect the presence of the disease as early as possible. Because passive surveillance is a relatively cost-effective and therefore commonly used process, it is the basic tool for infectious disease surveillance. Because of under-reporting in passive surveillance, cost-intensive active surveillance is often required to increase the sensitivity of the surveillance system. Using scenario tree modelling, the sensitivity of passive and active surveillance system components (SSC) can be quantified and an optimal, cost-effective surveillance system developed considering the contributions of each SSC. We illustrate this approach with the example of bTB surveillance in Switzerland where the surveillance system for bTB consists of meat inspection at the slaughterhouse (SLI), passive clinical surveillance on farm (CLIN) and human surveillance (HS). While the sensitivities for CLIN and HS were both negligible (<1%), SLI was assessed to be 55.6%. The scenario tree model showed that SLI is increasable up to 80.4% when the disease awareness of meat inspectors in Switzerland is enhanced. A hypothetical random survey (RS) was also compared with a targeted survey (TS) in high-risk strata of the cattle population, and the sensitivity of TS was 1.17-fold better than in RS but with 50% of the sample size.