Aga Khan University has launched the Kikohozi Classifier Project, an artificial intelligence-powered initiative aimed at enhancing early detection of tuberculosis and other respiratory diseases across five Tanzanian regions.
The project will be implemented in Iringa, Dar es Salaam, Kilimanjaro, Shinyanga and Dodoma regions, with coverage in Dodoma extending to Bahi District and Dodoma City.
Speaking at the launch on Feb. 23, Prof. Eunice Ndirangu, head of the School of Nursing and Midwifery for East Africa at Aga Khan University, said the project has received approval from the Ministry of Health and the National Tuberculosis and Leprosy Control Programme.
“We must undertake our own research with professionalism and produce results that are beyond doubt,” she said. “By doing so, Tanzania can become a benchmark for other nations.”
The Ministry of Health’s National Tuberculosis and Leprosy Control Programme Manager, Dr. Kisonga Riziki, said TB remains a major public health challenge in Tanzania and warned that the situation could worsen due to increased mining activity following the discovery of new mineral deposits, which are expected to draw more miners to the country.
Riziki called the project potentially transformative and urged the Aga Khan University team to implement it carefully to ensure positive outcomes. He noted that mining areas could become hotspots for respiratory diseases, underscoring the need for targeted health interventions and education in affected communities.
The initiative seeks to develop and evaluate a noninvasive, AI-powered cough audio classifier capable of detecting tuberculosis, asthma and chronic obstructive pulmonary disease in outpatient settings.
Specialized cough sound recording devices will be used to collect diverse cough samples from patients at health care facilities, with the data then analyzed using AI signal processing and machine learning techniques to generate an algorithm capable of automated disease identification.
