Abstract
BACKGROUND: The mosquito landing rate measured by human landing catches (HLC) is the conventional endpoint used to evaluate the impact of vector control interventions on human-vector exposure. Non-exposure based alternatives to the HLC are desirable to minimize the risk of accidental mosquito bites. One such alternative is the human-baited double net trap (HDN), but the estimated personal protection of interventions using the HDN has not been compared to the efficacy estimated using HLC. This semi-field study in Sai Yok District, Kanchanaburi Province, Thailand, evaluates the performance of the HLC and the HDN for estimating the effect on Anopheles minimus landing rates of two intervention types characterized by contrasting modes of action, a volatile pyrethroid spatial repellent (VSPR) and insecticide-treated clothing (ITC). METHODS: Two experiments to evaluate the protective efficacy of (1) a VPSR and (2) ITC, were performed. A block randomized cross-over design over 32 nights was carried out with both the HLC or HDN. Eight replicates per combination of collection method and intervention or control arm were conducted. For each replicate, 100 An. minimus were released and were collected for 6 h. The odds ratio (OR) of the released An. minimus mosquitoes landing in the intervention compared to the control arm was estimated using logistic regression, including collection method, treatment, and experimental day as fixed effects. RESULTS: For the VPSR, the protective efficacy was similar for the two methods: 99.3%, 95% CI (99.5-99.0) when measured by HLC, and 100% (100, Inf) when measured by HDN where no mosquitoes were caught (interaction test p = 0.99). For the ITC, the protective efficacy was 70% (60-77%) measured by HLC but there was no evidence of protection when measured by HDN [4% increase (15-27%)] (interaction test p < 0.001). CONCLUSIONS: Interactions between mosquitoes, bite prevention tools and the sampling method may impact the estimated intervention protective efficacy. Consequently, the sampling method must be considered when evaluating these interventions. The HDN is a valid alternative trapping method (relative to the HLC) for evaluating the impact of bite prevention methods that affect mosquito behaviour at a distance (e.g. VPSR), but not for interventions that operate through tarsal contact (e.g., ITC).