From Data to Impact: AI-Powered TB Hotspot Mapping in Uganda

We are proud to share that our latest research on AI-driven tuberculosis hotspot mapping in Uganda has been published as a preprint:“AI-Driven Tuberculosis Hotspot Mapping to Optimize Active Case-Finding: Implementing the Epi-Control Platform in Uganda.” 

This publication represents an important milestone for EPCON and our partners. It demonstrates how artificial intelligence can help national TB programs identify where undiagnosed patients are most likely to be found — enabling more targeted and effective interventions.

Why This Matters

Tuberculosis remains a major public health challenge in Uganda, with hundreds of new cases reported every day. Despite significant progress by the National TB and Leprosy Control Program, many cases remain undetected due to limited access to diagnostics, asymptomatic infections, and uneven healthcare access across the country.

Active case finding — proactively screening communities rather than waiting for patients to present at health facilities — has proven highly effective. But it is also resource-intensive. Mobile X-ray vans, community outreach teams, and diagnostic tools must be deployed strategically to maximize impact.

This is where data and artificial intelligence can make a real difference.

Using AI to Predict TB Hotspots

In this project, the Epi-control platform, developed by EPCON, was implemented in Uganda to predict TB hotspots at the community level.

The platform combines multiple data sources, including:

• Chest X-ray screening data from community interventions

• Demographic and socioeconomic indicators

• Environmental and infrastructure data

• Access to healthcare and vulnerability indicators

Using a Bayesian AI modelling approach, the platform created a digital epidemiological “twin” of TB risk across Uganda. This allowed us to estimate TB positivity rates at the sub-parish level, even in areas where screening had not yet taken place. Publication Uganda

The results were then visualized on an interactive geoportal, enabling program managers and field teams to identify high-risk areas and plan targeted screening activities.

The Results: Finding More Patients Where It Matters Most

The findings clearly demonstrate the value of predictive modelling.

Screening activities conducted in areas predicted as hotspots had significantly higher TB detection rates than those conducted in non-hotspot areas. Across the country, the model showed a risk ratio of 1.69, meaning screening in predicted hotspots identified substantially more TB cases.

In practical terms, this means that data-driven targeting could increase case detection by approximately 68% compared to routine screening strategies. This is a powerful example of precision public health: using data to ensure that limited resources are deployed where they can save the most lives.

A Strong Partnership

This work would not have been possible without close collaboration with our partners in Uganda.

We would like to sincerely thank:

• The National TB and Leprosy Control Program (NTLP) team at the Ministry of Health Uganda

• The incredible team at the Infectious Diseases Institute (IDI)

• All the implementing partners and community health teams who contributed to the active case finding interventions

Their expertise, dedication, and operational insights were essential not only for implementing the interventions but also for shaping and validating the modelling approach.

This project is a testament to what can be achieved when local expertise, programmatic data, and advanced analytics come together.

Looking Ahead

The results from Uganda reinforce a broader trend we are observing across multiple countries: AI-driven epidemiological modelling can significantly improve the efficiency of disease control programs.

By combining screening data with contextual indicators, health programs can move from reactive responses to proactive, data-driven strategies.

At EPCON, our mission is to support governments and partners in strengthening public health systems through smarter use of data.

We are proud that the Epi-control platform is helping to identify undiagnosed TB patients and guide more efficient interventions — bringing us one step closer to ending tuberculosis.