New study to evaluate Artificial Intelligence-assisted skin test for Tuberculosis prevention

The initiative will be implemented in Mozambique, Tanzania and Uganda.

The Manhiça Health Research Centre (CISM) will implement a research project to evaluate an innovative, low-cost tuberculosis skin testing approach supported by Artificial Intelligence (AI). The initiative aims to make tuberculosis prevention services faster, more accessible and more responsive to people’s needs.

The study will be implemented in Mozambique, Tanzania and Uganda. In Mozambique, it will take place in Manhiça District and will be led by CISM researcher and the study’s Principal Investigator in the country, Sozinho Acácio. Globally, the project is coordinated by the KNCV Tuberculosis Foundation in the Netherlands and funded by the Global Health EDCTP3 Joint Undertaking under the Horizon Europe Programme, co-funded by the European Union.

Known as TEST4TPT, the project will evaluate a strategy combining home-based tuberculosis skin testing using SIILTIBCY, the involvement of Community Health Workers (Agentes Polivalentes Elementares – APEs), and the use of AI to support the interpretation of test results. The initiative aims to bring tuberculosis prevention services closer to communities, reduce access barriers and promote a more people-centred approach. “To make the process more convenient and responsive to participants’ needs, they will be able to choose whether to have the test at home or at a health facility,” explains Sozinho Acácio.

Tuberculosis Preventive Treatment (TPT) is one of the main strategies for preventing people with tuberculosis infection from developing the disease. However, treatment initiation and completion rates remain below targets in many sub-Saharan African countries, partly due to limited access to feasible and affordable diagnostic methods for identifying tuberculosis infection.

According to the researcher, the main objective of the study is to determine whether integrating home-based testing with AI-assisted test interpretation increases the proportion of contacts of people recently diagnosed with tuberculosis who initiate and complete preventive treatment, compared with the model currently used by health services. “The study will also assess the reliability of test interpretation by APEs with the support of Artificial Intelligence, comparing their results with those obtained by healthcare professionals at health facilities. The feasibility, acceptability, clinical utility and cost-effectiveness of this strategy will also be assessed,” he adds.

The research is expected to involve approximately 330 participants in the reliability study and around 1,200 contacts of people recently diagnosed with bacteriologically confirmed pulmonary tuberculosis in the main trial.

The TEST4TPT results are expected to provide scientific evidence on the potential of this solution to strengthen tuberculosis prevention and support future decisions on the implementation of scalable, accessible and people-centred strategies in countries with a high burden of the disease, such as Mozambique, Tanzania and Uganda.

In addition to CISM, the project brings together the National Institute for Medical Research and Mwitikio wa Kudhibiti Kifua Kikuu na Ukimwi from Tanzania; the Makerere University Lung Institute and Mildmay Research Centre Uganda from Uganda; the Barcelona Institute for Global Health (ISGlobal), Spain; the London School of Hygiene & Tropical Medicine, United Kingdom; and the KNCV Tuberculosis Foundation, which coordinates the study globally.