Statistical Design and Analysis of Cluster Randomized Test-Negative Designs: Studies to Eliminate Dengue Fever

University Of California Berkeley, Berkeley CA

Investigators

Linked publications, trials & patents

Abstract

PROJECT SUMMARY/ABSTRACT The health/economic impacts of arboviral diseases transmitted by Aedes aegyti mosquitoes are escalating. The World Health Organization (WHO) considers dengue to be the fastest spreading mosquito-borne disease with now almost 4 billion people at risk of infection. In 2013, the estimated global burden of dengue was 390 million infections per year, with almost 100 million infections manifesting clinical disease. Further, the chikungunya virus, came to global attention in 2004 with epidemics on several Indian Ocean islands and the Zika virus emerged in epidemic fashion in the Western Pacific in 2013 and in Latin America in 2015. There is consensus that Ae. aegypti mosquitoes are the primary vectors of dengue, chikungunya and Zika. The Eliminate Dengue Program is an international research collaboration that is delivering a paradigm shift in the control of arboviral diseases transmitted by Ae. aegypti mosquitoes. The method utilizes the bacterium Wolbachia, common in insect species but not present in Ae. aegypti mosquitoes until they were stably transinfected in the laboratory. In insects Wolbachia is maternally transmitted via the egg and rapidly enters into naïve mosquito populations in a self-sustaining, durable manner. Strikingly, the presence of Wolbachia in Ae. aegypti mosquitoes renders them more resistant to disseminated arbovirus infection, including Dengue, Zika, and Chikungunya. There is an immediate need to design studies to test the efficacy of Wolbachia interventions in urban settings where cohort approaches are unrealistic due to ethical and logistic challenges associated with extensive continual disease monitoring. The difficulty of making evidence-based policy in relation to vector control has resulted in calls for improved trial methods. A field trial is planned from 2017?2020 in Yogyakarta, Indonesia, yielding data over the second half of this proposal period. We propose to study a variant of case-control design sampling, known as the test-negative design that has traditionally been used to assess the efficacy of seasonal influenza vaccines. Here we propose randomization of the intervention but in a cluster randomized fashion since intervention vector deployment must be done in contiguous areas. We need to assess properties of this design and also analytic techniques for the field data generated. Our proposed research will develop statistical methods for designing and analyzing cluster randomized test- negative studies. The methods will provide appropriate statistical tests and efficacy estimates, and inform future designs. Such methods are urgently needed to assess future Wolbachia efficacy trials in order to reduce human infections due to the dengue and Zika viruses.