Comparative evaluation of anopheline sampling methods in three localities in Indonesia

BACKGROUND: The effectiveness of vector control efforts can vary based on the interventions used and local mosquito behaviour and adaptability. In many settings, biting patterns of Anopheles mosquitoes can shift in response to interventions targeting indoor-biting mosquitoes, often resulting in higher proportions of mosquitoes feeding outside or at times when people are not protected. These behaviourally resistant mosquitoes have been shown to sustain residual malaria transmission and limit control efforts. Therefore, it is important to accurately sample mosquitoes to understand their behaviour. METHODS: A variety of traps were evaluated in three geographically diverse sites in malaria-endemic Indonesia to investigate local mosquito feeding behaviour and determine effective traps for surveillance. RESULTS: Eight traps were evaluated in three sites: Canti village, Lampung, Kaliharjo village, Purworejo, and Saketa village, Halmahera, Indonesia, including the gold standard human landing collection (HLC) and a variety of traps targeting host-seeking and resting mosquitoes both indoors and outdoors. Trapping, using indoor and outdoor HLC, the Ifakara tent trap C, goat and human-occupied tents, resting pots and boxes, and CDC miniature light traps was conducted for 16 nights in two sites and 8 nights in a third site, using a Latin square design. Trap efficacy varied by site, with outdoor HLC yielding the highest catch rates in Canti and Kaliharjo and a goat-baited tent trap proving most effective in Saketa. In Canti village, anthropophilic Anopheles sundaicus were caught indoors and outdoors using HLCs, peaking in the early morning. In Kaliharjo, a variety of mosquitoes were caught, mostly outdoors throughout the night. HLC was ineffective in Saketa, the only site where a goat-baited tent trap was tested. This trap was effective in catching zoophilic vectors outdoors before midnight. CONCLUSIONS: Different trapping methods were suitable for different species, likely reflecting differences in behaviour among species. The three villages, each located on a different island in the Indonesian archipelago, contained mosquito populations with unique behaviours. These data suggest that the effectiveness of specific vector monitoring and control measures may vary by location. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12936-017-2161-9) contains supplementary material, which is available to authorized users.