Despite the efforts in reducing vector densities, outbreaks of dengue fever have become a frequent event in Sri Lanka. As explained by dengue transmission dynamics, vector control activities intensified at peak or near peak outbreak situations would not be successful in controlling the outbreaks. A reliable method of outbreak prediction is always warranted for early preparedness. Relationships between the monthly Breteau indices of the two vector species (Aedes aegypti L. and Ae. albopictus Skuse) and the monthly dengue incidence in a selected high-risk health division (Kaduwela) in the Colombo District, Sri Lanka were determined for three consecutive years, 2009 to 2011. The same procedure was extended for the whole Colombo District from 2013 to 2016. Cross correlation functions were used to determine the relationships with the corresponding lag-periods. Receiver Operating Characteristic Curves (ROC) were constructed to determine the performance of the Breteau indices as predictors of impending dengue outbreaks and to establish the threshold values. The pronounced performance with >80% areas under ROC curves and >75% sensitivity and >70% specificity of threshold values with defined lag-periods in correlations emphasize the importance of the Breteau index as a promising early warning signal for dengue outbreaks. The results indicate the importance of the carefully planned routine vector larval surveillance in dengue control programs which would make reliable outbreak predictions with a sufficient window period for early preparedness.
Keywords: Breteau index; Dengue incidence; Mosquitoes; Outbreak; Prediction.
Published by Elsevier B.V.