Counting TB Cases in Kids: Mechanistic Modeling to Quantify the Underreporting of Pediatric Tuberculosis

Abstract

Tuberculosis (TB) represents one of the major issues of global health, causing millions of deaths each year. Although countries are required to notify the World Health Organization (WHO) of all TB cases, many of them go undetected, and the gaps in case detection are especially large for children and young adolescents due to the poor performance of current TB diagnostics in these age groups. These gaps lead to delays in treatment initiation, likely contributing to many deaths. In this thesis, we estimate and compare the extent of childhood TB underreporting in 168 countries over the period 2013 to 2018. To do so, we formulate a mechanistic model of child TB incidence and utilize a Bayesian approach to fit this model’s parameters. The analysis compares our modeled incidence estimates to observed pediatric TB case notifications to estimate the level of underreporting in each country. We find that the majority of pediatric TB cases go unreported, though country-level proportions of undetected cases differ widely both within and across regions and income levels. Our results may provide guidance to health officials in individual countries and at institutions like the WHO in their efforts to mitigate TB in the pediatric population.