Emerging methods and applications for test-negative studies of of infectious disease interventions

PROJECT SUMMARY/ABSTRACT The need for longitudinal follow-up of individuals exposed to interventions against rare infectious disease endpoints poses a barrier to prospective efficacy and effectiveness studies. Studies using the test-negative design (TND) have become a popular alternative. TNDs represent a variant on traditional case-control designs: studies enroll subjects who seek care for a clinical syndrome, defining those who test positive and negative for a pathogen of interest as “cases” and “controls”, respectively. To facilitate rigorous and reproducible assessments of vaccine performance, the project will re-assess emerging applications of the TND, and contribute methods to measure intervention effects from data collected by TND studies. The first two aims revisit estimation strategies for vaccine-conferred protection against infection and against the progression of infection to diseaseâ€Â"the two components of the vaccine “direct effect” that TND studies aim to measure. We propose novel frameworks to estimate each effect through extensions of the TND: one through comparisons of symptomatic and asymptomatic persons, and another leveraging the age distribution of cases. We will apply these methods to estimate pneumococcal conjugate vaccine effectiveness against vaccine-serotype pneumococcal pneumonia and carriage, and to re-assess reported differences in rotavirus vaccine effectiveness across high, middle, and low-income countries. In Aim 3, we will continue development of statistical procedures for cluster randomized test- negative designs. A major field trial of this nature for a vector intervention utilizing the bacterium Wolbachia to reduce dengue fever transmission is nearing completion. Permutation- based inference is planned because of small numbers of clusters. We will extend such methods (i) to allow for individual measures of intervention exposure (based on human and mosquito mobility), and (ii) to include design variants such as the stepped wedge design and interrupted time series where data collection has either recently ended or is in process. We will also consider a novel application to assess a new vaccine against Ebola Virus Disease (EVD) in the Democratic Republic of the Congo. This project will thus contribute methods and computational routines to allow valid inferences from TND study data, permitting novel assessments of interventions against rotavirus, pneumococcal disease, EVD, and dengue etc.