Plain-English translation of NCT05317247 on ClinicalTrials.gov ↗ ·
CAGE-TB is testing a smartphone application that listens to your cough sound to see if it can help quickly identify whether you might have tuberculosis. Researchers will collect cough audio recordings from people with persistent coughs and use machine learning—a type of artificial intelligence—to teach the app to recognize patterns in cough sounds that suggest tuberculosis. This study will help determine whether the app is accurate enough to become a useful, low-cost screening tool in clinics.
Tuberculosis is a serious infection that spreads in high-burden communities, and early detection saves lives. Current screening methods can be slow, expensive, and hard to access in resource-limited settings. This trial exists to see if a simple, cheap smartphone app could help clinics quickly identify patients who might have tuberculosis so they can get tested and treated faster.
You likely qualify if…
You likely don't qualify if…
You will visit a clinic once to participate. During that visit, you will record your cough sounds using a smartphone app, and you will provide a sputum sample so doctors can test whether you have tuberculosis. The study has two phases: first, researchers will use recordings from about 473 people to teach the app to recognize tuberculosis cough patterns; then, they will test the app on about 1,278 new participants from South Africa and Uganda to see how well it actually works. Your participation is observational only—no medication or treatment is being given as part of this study.
AI-generated summary from trial data · Jun 3, 2026 · Not medical advice
South Africa
Uganda