The Challenge
With nearly one-fourth of the world's tuberculosis (TB) cases occurring in India1, TB remains a key public health priority. The disease is preventable and curable, and outcomes improve substantially when identified early.
Since traditional screening relies mainly on visible symptoms, asymptomatic or minimally symptomatic individuals are often missed, a critical gap, since undiagnosed individuals can unknowingly transmit the disease to 10–15 people each year2.
Our AI Solution
Cough Against TB (CATB) is an AI-powered, non-invasive screening application that analyzes cough sounds and symptoms to identify individuals with presumptive pulmonary TB, enabling early detection in both facility and community settings, even offline.
Designed for India's public health ecosystem, CATB works on Android smartphones and extends the reach of India's TB program to the last mile.
How it works
01
App setup
the healthcare worker installs the app and begins the screening procedure with consent from the individual.
App setup
the healthcare worker installs the app and begins the screening procedure with consent from the individual.
02
Individual registration
a digital record is created for screening and follow-up (age, gender, and other details).
Individual registration
a digital record is created for screening and follow-up (age, gender, and other details).
03
Symptom assessment and cough recordings
health workers record TB-related symptoms and capture three cough recordings following all NTEP infection and prevention control measures.
Symptom assessment and cough recordings
health workers record TB-related symptoms and capture three cough recordings following all NTEP infection and prevention control measures.
04
CATB analyzes cough sounds
an AI model trained on India-specific data analyzes cough sounds alongside symptoms.
CATB analyzes cough sounds
an AI model trained on India-specific data analyzes cough sounds alongside symptoms.
05
Instant risk flagging and referral
the app identifies presumptive or non-presumptive pulmonary TB; presumptive cases are referred for further diagnostic testing.
Instant risk flagging and referral
the app identifies presumptive or non-presumptive pulmonary TB; presumptive cases are referred for further diagnostic testing.
06
Ni-kshay ID generation and update
a Ni-kshay ID is generated or updated for referred cases to ensure continuity within the national TB program.
Ni-kshay ID generation and update
a Ni-kshay ID is generated or updated for referred cases to ensure continuity within the national TB program.
Where It's Deployed
Operational since March 2023, developed with India's Central TB Division, Ministry of Health and Family Welfare. Active in 17 states and 103 districts.
Impact
CATB demonstrates how AI-powered screening, designed around field realities, can strengthen TB detection within public health systems. By identifying presumptive TB cases—including 14–16% of individuals who may not present with typical symptoms—it extends screening beyond conventional symptom-based approaches, supporting earlier detection and stronger linkage to confirmatory diagnosis. With 90% sensitivity, 69% specificity, and an AUROC of 0.887, CATB demonstrates strong screening performance across community and facility settings. By strengthening the reach and efficiency of TB screening, CATB contributes to India’s efforts under the Pradhan Mantri TB Mukt Bharat Abhiyan and SDG 3, including the goal of ending epidemics and advancing universal health coverage.
Our Other Solutions
Partners
