Background: Surveillance is an essential component of global programs to eliminate infectious diseases, with surveillance systems often being considered the first line in averting epidemics for (re-)emerging diseases. As the numbers of cases decline, costs of treatment and control diminish, but those for surveillance remain high even after the ‘last’ case. Economies made by reducing surveillance risk missing persistent or (re-)emerging foci of disease, therefore it is vital that surveillance networks are adequately designed; however, guidance on how best to optimise disease surveillance in an elimination setting is lacking. Here, we use a simulation-based approach to determine the minimal number of passive surveillance sites required to ensure maximum coverage of the population at-risk (PAR) of an infectious disease.
Methodology and Principal Findings: For this study, we use Gambian human African trypanosomiasis (g-HAT) in north-western Uganda, a neglected tropical disease (NTD) which has been reduced to historically low levels (<1000 cases/year), as an example application. To quantify travel time to diagnostic facilities, a proxy for surveillance coverage, we produced a high spatial-resolution friction surface and performed cost-distance analyses. We simulated travel time for the PAR with different numbers (n = 1...170) and locations (170,000 total placement combinations) of diagnostic facilities, quantifying the percentage of the PAR within 1h and 5h travel of the facilities, as per in-country targets. Our simulations indicate that a minimum of 54 diagnostic centres are required to meet provisional targets, and we determine where best to place these facilities, enabling a minimal impact scale back from 170 facilities operational in 2017. Scaling back operational facilities ensures that costs can be reduced without impairing surveillance of the core areas with remaining cases.
Conclusions: Our results highlight that surveillance of g-HAT in north-western Uganda can be scaled back without reducing coverage of the PAR. The methodology described can contribute to cost-effective strategies for the surveillance of NTDs and other infectious diseases approaching elimination or applied to (re-)emerging diseases for which the design of a novel surveillance network is required.