Development and Validation of ‘findtb’, A Novel Digital Tool for Tuberculosis Triage in Children - Using Innovative Methodologies and an Interdisciplinary Approach

Introduction

About 50-65% of children suffering from tuberculosis (TB) across all age groups are never detected. This contributes to the 250,000 deaths from TB in children annually, making the disease one of the top ten killers of children under the age of five.

Project Aim

To improve TB detection in children and reduce TB-related morbidity and mortality, novel diagnostic tests are urgently needed. The Find-TB project aims to address this urgent need by improving case finding at all levels of care in resource-limited, high TB burden settings through the development and validation of an innovative, digital TB triage tool.

Tool Development

This tool will provide individualized TB disease risk predictions coupled with clinical recommendations to guide next steps of care. The project will capitalize on digital and scientific innovations that have been successfully applied to the COVID-19 pandemic.

Methodology

Starting with large, existing geographically and demographically representative data sets, my team and I will:

1. Develop and validate two predictive models using highly innovative data science methods (machine learning and Bayesian techniques).
2. Create an app for use by health personnel at all levels of the health care system based on the best performing model.

Stakeholder Engagement

To maximize the app’s implementation potential, I will engage key stakeholders to assess preferences, acceptability, and feasibility prior to and throughout app development. Furthermore, I will apply human-centered design principles.

Evaluation

We will then evaluate the design-locked digital tool in a real-world triage use case in a prospective, cross-sectional diagnostic accuracy study. A cost-effectiveness analysis of population-level impact will round out a dossier that will inform a WHO review.

Expected Outcomes

This efficient, effective, and scalable digital tool for TB triage of children in resource-limited settings is expected to:

• Improve TB diagnosis on an individual level, thus reducing TB-related morbidity and mortality.
• Likely improve TB control on a population level.