From Data to Impact: Building Smarter TB Programs in Kenya

At EPCON, we believe that the future of public health lies in transforming data and AI into actionable intelligence. In Kenya, we are putting this belief into practice, working closely with the National TB Program (NTP) and the Centre for Health Solutions, Kenya (CHS) to strengthen tuberculosis (TB) programs through advanced AI-driven modeling.

This work is part of a broader effort to move beyond fragmented data systems and toward integrated, intelligent decision-making that supports health workers on the ground as well as policymakers at the national level.

A New Generation of TB Intelligence

In Kenya, our work focuses on four interconnected components:

1. TB risk modeling

2. Predictive modeling for preventive treatment (TPT)

3. Treatment outcome optimization

4. Diagnostic and service network optimization (in progress)

Understanding Where TB Happens: Risk Modeling

Our TB risk model is designed to identify where the burden of disease is likely to be highest.

By integrating a wide range of data sources—including surveillance data, socio-economic indicators, environmental factors, and health system access—we uncover the complex drivers of TB transmission.

This includes variables such as:

• Population density and demographics

• HIV prevalence

• Poverty and deprivation indices

• Access to healthcare facilities

• Environmental and spatial factors

The result is a high-resolution TB vulnerability map that supports:

• Targeted screening strategies

• Better allocation of outreach teams

• More efficient use of limited resources

Moving Upstream: Predicting Preventive Treatment Needs

Finding TB cases is only part of the solution. Preventing TB before it develops is equally critical.

That’s why we are developing predictive models for TB Preventive Treatment (TPT), helping to identify:

• Where uptake is likely to be low

• Which population groups are most at risk

• What barriers are preventing successful implementation

By integrating data on index patients, contact tracing, and treatment uptake, the model enables programs to proactively intervene rather than react to missed opportunities.

This allows health systems to:

• Prioritize vulnerable population groups

• Improve adherence strategies

Improving Outcomes: Predicting Treatment Success

Ensuring that patients successfully complete TB treatment remains a major challenge.

In Kenya, we are developing models that analyze:

• Patient-level risk factors

• Contextual risk factors

The goal is to identify patients and communities at risk of poor treatment outcomes and provide actionable recommendations to improve success rates.

Optimizing the System: Smarter Networks

Beyond individual patients and communities, we are also working on optimizing the broader TB service network.

This includes:

• Mapping diagnostic capacity (GeneXpert, Truenat, CXR, microscopy)

• Identifying underserved areas

• Detecting underutilized resources

• Simulating improved referral and service delivery pathways

The objective is clear: align supply with demand, ensuring that services are available where they are needed most.

Looking Ahead

Our work in Kenya is still evolving. Models are being refined, new data streams are being integrated, and additional capabilities, such as real-time updates and network optimization, are under development.

But one thing is already clear: When data is integrated, contextualized, and translated into action, it becomes one of the most powerful tools we have to fight TB.