Revolutionizing TB Diagnosis: AI-Powered Mobile App Development at Kenya Medical Research Institute

Led by Dr. Videlis Nduba and his team, the research involves recording coughs from individuals with respiratory ailments, including TB, as well as those without such diseases. By collecting cough data using various microphones, including smartphone microphones, the team aims to create software capable of accurately distinguishing between different types of coughs associated with TB and other serious respiratory conditions.

The recorded coughs are analyzed using AI algorithms, specifically the ResNet 18 computer software system developed by the University of Washington. This software utilizes mathematical modeling to analyze and differentiate cough spectral grams, enabling the identification of cough patterns indicative of TB infection.

Dr. Nduba emphasizes the potential impact of this technology in reducing the time taken for TB diagnosis and treatment initiation, thereby curbing the spread of the disease. He highlights the significance of early diagnosis in preventing TB transmission within communities, emphasizing the urgency of implementing efficient diagnostic tools.

Despite promising results, the software has yet to meet the stringent accuracy standards set by the World Health Organization (WHO). While achieving 80% accuracy in detecting TB and 70% accuracy in ruling out TB, the software falls short of the WHO's requirements of 90% accuracy for TB detection and 80% accuracy for ruling out the disease.

Former TB patient Johnson Munori, who participated in the research, underscores the potential benefits of the app in facilitating timely diagnosis and treatment initiation, based on his own experience.

Jarim Omogi, a public health specialist at Amref International University, emphasizes the growing role of AI in medicine and its potential to enhance healthcare delivery by expediting diagnosis and improving access to care.

While the trial has received funding from the National Institutes of Health, regulatory approval for the application is pending. Once approved, this innovative technology could revolutionize TB diagnosis and contribute significantly to global efforts to combat the disease.