One Health and AI approach aims to boost data insights  

A public health transformation can be achieved in Eastern Uganda using health and demographic surveillance from a sentinel site, preliminary analysis of two decades of data has revealed.  

Statistical analysis of 21 years of data from the country reveals a noticeable fall in mortality in the past decade.  

Even in this predominantly rural setting, people are living longer, and since 2018 the leading causes of death have become lifestyle diseases, such as diabetes, heart disease and cancer.  

This shift has likely been driven by improved control of infectious diseases, coupled with lifestyle choices. 

Where individuals live, their age and their economic status have in recent years become key to their health outcomes.  

Applying artificial intelligence (AI) to the data, from a population cohort numbering 110,000 people from 22,000 households in 65 communities, in a database with 1.2 million entries, would further unravel correlations and reveal the structure of drivers behind this health outcome shift, scientists say. 

For this Ugandan cohort, the Iganga-Mayuge Health & Demographic Surveillance Site, the shift towards non-communicable diseases as a leading cause of death was detectable in 2017.  

Recognition of this change in public health would have allowed a seven-year planning and interventional head start, and enabled actions to help the population age well.  

“With the right approach, this becomes the perfect early warning system,” says Dr Muwonge. “AI can allow you to unpick all the potential contributors to the trend, and understand what is driving the shift.” 

The vision offered by AI on long-term health data could be strengthened to include details of livestock and environmental health, to understand public health challenges in a more holistic way. 

Dr Muwonge is adapting this approach, with a pioneering framework called the One Health Living Lab, bringing together human, animal and environmental perspectives on health.  

By building on the Ugandan health cohort, the Living Lab aims to address pressing challenges such as antimicrobial resistance (AMR).  

Dr Muwonge defines a One Health living lab: “This cohort becomes a dedicated site and ecosystem where we can co-design and iteratively test solutions relevant to such communities.  

“Central to this is a One Health laboratory in this community, which allows monitoring of drivers of health outcomes across the One Health interfaces.” 

Roslin scientists will work with colleagues Dr Dan Kajungu, Professor Moses Joloba and Dr Clovice Kankya at of Makerere University to bolster data from the cohort, currently focused on health and demographics, by integrating information from veterinary sources and the environment.  

This approach will aim to create a holistic understanding of disease, which could inform policy. 

It has potential to unravel the infectious disease landscape, understand the risk factors linked to infection in a wider context, and to enable appropriate use of treatments for infection, such as antibiotics. 

A series of studies by Dr Muwonge and others has shown the scale of the antimicrobial challenge across sectors in Uganda.  

They have shown that a move to intensive farming, driven by growing demand, is likely fuelling a rise in antimicrobial resistance and use of antimicrobial treatments in farming. 

Antimicrobial resistance is also increasing among people, according to data from referral hospitals across the country, while a study of commercial poultry shows that almost two-thirds of imported antibiotics are administered to 30 per cent of birds. 

“Evidence is generated one piece at a time, and a vantage point is needed to make sense of the dynamics of a problem such as AMR – and communities of practice, including knowledge brokers, private sector and policymakers and regulators are essential for the One Health Living Lab idea to become a reality,” says Dr Muwonge. 

The approach will investigate cases of infections by considering exposure to human, animal and environmental diseases.  

Samples taken from people in the community may be tested for diseases common in animals or for exposure to environmental harms, such as pesticide. The results are intended to enable optimum treatments.

Dr Muwonge explains: “For example, if a person has fever from brucellosis – an animal infection caught by drinking unpasteurised milk – they may be presumed to have malaria or something else that is not zoonotic.  

“Human health differential diagnostics often have limited oversight of such infections in such settings, which not only chronically affect humans but also severely impact the productivity of their animals,“ he adds. 

“A One Health laboratory not only guides clinical management but also serves as a foundation for preparedness. Through diagnostic guided treatment we can reduce the overuse of drugs, particularly antimicrobials – the backbone of healthcare.  

“We are expanding the information that is collected from the community to include livestock and environment factors in addition to health and population. Our framework means prospectively when someone has a clinical sign of illness, we use the breadth of information to inform diagnosis and management as well as retrospectively examine how infectious and non-communicable diseases have emerged in a community. 

In future, the team hopes the potential of data sets can be further boosted by weaving in relevant publicly available information, such as weather records, economic data, or agricultural yields.  

Integrating this data could reveal how seasonal changes influence disease outbreaks, giving valuable context on health trends.  

Most countries in the Global South have long-term health studies involving cohorts. Expanding the remit of these studies, looking at factors influencing public health outcomes, has potential to generate a new understanding of health across life, Dr Muwonge says.