HIV remains a major public health issue in Cameroon. The recent shortage of testing kits, linked to the reduction in international funding, limits access to tests and compromises the early identification of infected individuals. In this context, optimising the use of screening resources becomes a priority.

The objectives of this work are to identify the main sociodemographic, economic and behavioural determinants of HIV by region, to use these determinants to develop a prediction tool based on a model for predicting serological status, and to evaluate the tool’s performance.

Our method detects 129 positive cases compared with 11 using the WHO rule, representing nearly 12 times more cases identified. This gain, observed across all ten regions of Cameroon, highlights the value of a regionalised approach based on AI. The proposed approach makes it possible to prioritise HIV screening at the regional level in a context of limited resources. It provides a simple, interpretable decision-support tool that can be used directly by healthcare professionals.