Estimation of country-level incidence of early-onset invasive Group B Streptococcus disease in infants using Bayesian methods.

Neonatal invasive disease caused by Group B Streptococcus (GBS) is responsible for much acute mortality and long-term morbidity. To guide development of better prevention strategies, including maternal vaccines that protect neonates against GBS, it is necessary to estimate the burden of this condition globally and in different regions. Here, we present a Bayesian model that estimates country-specific invasive GBS (iGBS) disease incidence in children aged 0 to 6 days. The model combines different types of epidemiological data, each of which has its own limitations: GBS colonization prevalence in pregnant women, risk of iGBS disease in children born to GBS-colonized mothers and direct estimates of iGBS disease incidence where available. In our analysis, we present country-specific maternal GBS colonization prevalence after adjustment for GBS detection assay used in epidemiological studies. We then integrate these results with other epidemiological data and estimate country-level incidence of iGBS disease including in countries with no studies that directly estimate incidence. We are able to simultaneously estimate two key epidemiological quantities: the country-specific incidence of early-onset iGBS disease, and the risk of iGBS disease in babies born to GBS-colonized women. Overall, we believe our method will contribute to a more comprehensive quantification of the global burden of this disease, inform cost-effectiveness assessments of potential maternal GBS vaccines and identify key areas where data are necessary. Author summary: Invasive disease caused by Group B Streptococcus (GBS) in young infants continues to be a major public health problem in both developed and developing countries. However, data on the incidence of this infection during the first week of life are only available for a small number of countries, which has complicated the quantification of the burden of this disease globally. In this paper, we develop a Bayesian framework to estimate the incidence of invasive GBS infection that combines data from multiple types of epidemiological studies, with adjustment for relevant factors such as diagnostic methods used in the studies. We present estimates from a series of models, and our results highlight the potential weaknesses of different types of studies and the importance to consider the entire evidence when estimating global burden of invasive neonatal infections. We believe this model is a step toward better quantification of the number of cases in different regions. [ABSTRACT FROM AUTHOR]