Current and lagged associations of meteorological variables andAedesmosquito indices with dengue incidence in the Philippines

Abstract

BackgroundDengue is an increasing health burden that has spread throughout the tropics and sub-tropics. There is currently no effective vaccine and control is only possible through integrated vector management. Early warning systems to alert potential dengue outbreaks are currently being explored but despite showing promise are yet to come to fruition. This study addresses the use of meteorological variables for predicting both entomological indices and dengue incidences and assesses the added value of additionally using entomological indices for predicting dengue incidences.Methodology/Principal FindingsEntomological surveys were carried out monthly for 14 months in six sites spread across three environmentally different cities of the Philippines. Meteorological and dengue data were acquired. Non-linear generalized additive models were fitted to test associations with the meteorological variables and both entomological indices and dengue cases. Rain and the diurnal temperature range (DTR) contributed most to explaining the variation in both entomological indices and number of dengue cases. DTR and minimum temperature also explained variation in dengue cases occurring one and two months later. The number of adult mosquitoes did associate with the number of dengue cases, but contributed no additional value for predicting dengue cases.Conclusions/SignificanceThe use of meteorological variables to predict future risk of dengue holds promise. The lack of added value of using entomological indices confirms several previous studies and given the onerous nature of obtaining such information, more effort should be placed on improving meteorological information at a finer scale to evaluate efficacy in early warning of dengue outbreaks.Author summaryDengue is a widespread mosquito-borne disease. Mosquitoes are sensitive to temperature and rainfall and hence there have been efforts to identify such variables for predicting dengue outbreaks. Several mosquito indices are measured routinely by national surveillance systems, but which vary considerably in their success of predicting dengue outbreaks. This study explored the current and lagged associations of meteorological variables with mosquito indices and dengue incidence. Associations of mosquito indices with dengue were also explored. Rain and the diurnal temperature range (DTR) contributed most to explaining the variation in both entomological indices and number of dengue cases. DTR and minimum temperature also explained variation in dengue cases occurring one and two months later. Mosquito indices did not provide any additional predictive power of dengue incidences. Given the onerous nature of measuring mosquito indices, advanced warning systems might be improved using meteorological variables measured at finer scales than that traditionally available.