This is the 19th post in a blog series to be published in 2023 by the APET Secretariat on behalf of the AU High-Level Panel on Emerging Technologies (APET) and the Calestous Juma Executive Dialogues (CJED)

Occurrences of foot and mouth disease (FMD) outbreaks in African countries have been significantly harmful to the cattle industry for the past decade. A prediction of FMD outbreaks based on relevant risk factors with a high prediction accuracy is, therefore, important for authorities to develop a plan for preventing the outbreaks. Machine Learning such as data-driven tools are widely accepted for their prediction abilities, but an application of these techniques to FMD outbreak prediction is very limited. 

The livestock industry has played a crucial role in reducing poverty, and ensuring food security, thereby making it a significant contributor to the African Union’s (AU) Agenda 2063 goals. To achieve economic growth and eradicate poverty, the Framework places increased priority on the livestock industry. Aspiration 1 of Agenda 2063 envisions a prosperous Africa through inclusive growth and sustainable development, with goal number 5 emphasising modern agriculture to boost productivity and production.[1] This goal recognises the importance of the livestock industry to the African economy and the need to develop it sustainably.

Cattle hold immense significance in Africa, with the continent having an estimated 370 million heads of cattle in 2020. Beef plays a crucial role as a daily source of food and nourishment, while also providing essential cash income, nitrogen-rich manure for soil replenishment, draught power, milk, and meat.[2] Additionally, cattle farming offers employment opportunities for many individuals on the continent and is symbolic of wealth and success in numerous tribes. However, cattle farming in Africa faces various challenges that impede production. Climate stress, nutritional deficiencies, illnesses, limited access to land and water, insufficient market channels, ineffective rangeland management, and inadequate feed supplies all contribute to these difficulties. Consequently, cattle are exposed to various stresses that negatively impact fertility, growth rate, and mortality, ultimately affecting the output of cattle farming on the continent.[3]

Foot and mouth disease (FMD) is an important economic transboundary animal disease caused by a virus of the genus Apthovirus from the Picornaviridae family.[4]  FMD is a highly contagious viral illness affecting cloven-hoofed animals such as cattle, pigs, sheep, and goats. It is characterised by fever, mouth, and foot blisters, leading to lameness. FMD spreads rapidly through contact with infected animals, fluids, or contaminated surfaces.

FMD is responsible for significant losses in production and productivity to cattle farmers, as well as resulting in trade embargoes and substantial overarching economic losses for the broader cattle industry. Globally, the disease is a cause of major concern as it is an important transboundary disease found in several regions of the world including Africa.  Hence, among the diseases that significantly affect cattle in the African region, transboundary foot-and-mouth disease (FMD), stands out as a highly contagious threat.

Although there is no cure, treatments can relieve symptoms and prevent transmission. Therefore, controlling FMD necessitates biosecurity, vaccination, early detection, and reporting. These preventive measures are critical due to the disease's devastating impact on livestock populations and economies. FMD presents significant economic and social challenges, leading to the culling of infected animals.  However, it does not pose a threat to human health.

The disease poses significant restrictions on production and market access, acting as a major barrier to the cattle industry and severely affecting the livelihoods and resilience of rural communities that rely on cattle.[5] The disease results in substantial income losses due to reduced milk yield, stunted growth in affected animals, and loss of livestock markets and trade disruptions.[6] Therefore, addressing the impact of foot and mouth disease is crucial in ensuring the sustainability of the cattle industry and the well-being of the communities dependent on it.

There is much interest in the use of prediction models for infectious diseases, as prediction models for livestock disease outbreaks based on classical statistical and mathematical techniques have been widely demonstrated.  To this end, the African Union High-Level Panel on Emerging Technologies (APET) recommends the utilisation of machine learning (ML) technologies to predict and control the spread of FMD. Machine learning is a branch of artificial intelligence (AI) and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy. Machine learning is an important component of the growing field of data science. Through the use of statistical methods, algorithms are trained to make classifications or predictions, and to uncover key insights in data mining projects. These insights subsequently drive decision making within applications and businesses, ideally impacting key growth metrics.

APET acknowledges that ML is still a relatively new technology globally, particularly in Africa.  However, the panel recognises its potential and the progress to be made in its development. APET opines that ML can be harnessed to combat FMD across Africa, including in the areas of biosecurity, vaccination, early detection, and reporting.  In 2019, for example, the African Union Inter-African Bureau for Animal Resources (AU-IBAR) implemented a machine learning-based system to monitor biosecurity measures in Kenyan livestock facilities.[7] By leveraging data from sensors and cameras, the system can identify and report potential biosecurity risks, such as animal movement and the presence of contaminated materials.

The AU-IBAR, in collaboration with the University of Edinburgh in Scotland, has developed the FMD Weather Index, a predictive project used in Africa to assess the risks of FMD outbreaks. To facilitate reporting the AU-IBAR developed a machine learning-based system in 2019 to automate the reporting of FMD cases. This system utilises animal health records and weather data, the system identifies FMD cases and automatically reports them to relevant authorities, expediting responses to FMD outbreaks.[8]

By analysing weather data, including temperature, humidity, and rainfall, the system identifies areas at high risk of experiencing FMD outbreaks and generates a risk score for each region.[9] This information aids in prioritising vaccination campaigns and implementing biosecurity measures. Currently in use in Ethiopia, Kenya, and South Africa, the FMD Weather Index has proven to be effective in predicting and reducing FMD incidence. The system utilises machine learning algorithms trained on historical FMD outbreaks and weather data, with potential to expand to other African countries in the future.[10]

In 2018, Ethiopia partnered with the World Bank to launch a project that used machine learning to enhance vaccination coverage against FMD. The project targeted the needs of Ethiopia's livestock sector, which significantly contributes to the country's economy. By combining animal health records and weather data, the project identifies areas prone to FMD outbreaks, facilitating prioritised vaccination efforts and increased coverage. Through the project's efforts, vaccination coverage was increased, leading to a reduction in the incidence of FMD in Ethiopia.

Furthermore, in Kenya, a project led by the African Institute of Technology (AIT) in collaboration with the Kenya Agricultural and Livestock Research Organization (KALRO) developed a new FMD vaccine in 2021. By employing machine learning techniques, the project aims to identify the most effective vaccine strains and optimise the vaccine manufacturing process. Using machine learning, the project, aims to identify effective vaccine strains and optimise the manufacturing process by analysing animal health records and weather data.[11] The team seeks to detect patterns to determine the most potent FMD vaccine strains and enhance production quality while reducing costs. Though in its early stages, the project has the potential to revolutionise FMD vaccine development, potentially preventing outbreaks and safeguarding African livestock populations.

To enhance the early detection of FMD,  Kenya implemented the FMD Early Warning System (FEWS) in 2017, which actively detects FMD outbreaks in livestock populations. FEWS analyses data from animal health records and weather reports, including vaccination status, herd health, and weather conditions, to identify potential FMD outbreak patterns and issue timely alerts. The system's effectiveness in early detection enables swift action, curbing the disease's spread and reducing economic losses.[12]

FEWS is also under development for implementation in other African countries. This promising tool holds the potential to protect livestock populations and prevent FMD outbreaks across Africa. Developed through a collaborative project between KALRO and the Food and Agriculture Organisation (FAO) of the United Nations, FEWS relies on machine-learning algorithms trained on historical FMD outbreak data and is slated to be fully operational by 2023.

Another example is the FMD Early Detection System (FEMS), designed by the University of Pretoria, which serves as a machine learning-based tool to detect FMD outbreaks in South African dairy farms. It leverages sensor readings and animal health records to identify early indicators of an FMD outbreak, such as temperature, humidity, animal movement, and health status. Analysing this data with machine learning algorithms, FEMS identifies patterns signalling an FMD outbreak and generates timely alerts to prevent disease spread. The system's effectiveness in early detection has proven to be cost-effective and can save lives and safeguard livestock populations.[13]

Currently operational in South African dairy farms, FEMS is being developed for implementation in other African countries as well, making it a valuable tool for preventing FMD outbreaks and protecting livestock populations in Africa. Collaboratively developed with the South African Dairy Development Corporation (SDDC), the system is based on historical FMD outbreak data and is expected to be fully operational by 2023. In 2021, the FAO launched a machine learning-based system featuring a chatbot to assist farmers in reporting FMD cases. Simplifying the reporting process, the system enhances early detection of FMD outbreaks and facilitates prompt action. These innovative approaches demonstrate how machine learning is contributing to effective FMD control measures in Africa.

Policy recommendations for supporting FMD control using machine learning technologies encompass several key areas. Firstly, utilising ML to develop systems for monitoring biosecurity measures in livestock facilities can help identify and report potential risks, such as animal movement and contaminated materials.[14] Secondly, employing ML to predict FMD outbreaks based on weather and animal health data can aid in targeting vaccination efforts and biosecurity measures in high-risk areas. Thirdly, implementing ML-based systems for early detection of FMD outbreaks can expedite responses, preventing further disease spread.

APET urges that integrating ML in reporting processes can streamline FMD case identification and reporting to relevant authorities, enhancing response times. Therefore, to maximise the impact of ML in FMD control, the panel encourages African countries to invest in research and development for new tools, provide training on ML usage to government officials and livestock farmers, and foster a supportive regulatory environment for its implementation.[15]

In conclusion, the panel asserts that the adoption of machine learning technologies in FMD control represents a pivotal step towards eradicating the disease and protecting livestock populations in Africa. Embracing these policy recommendations will empower African countries to leverage ML for predicting and preventing FMD outbreaks effectively.

The panel concludes further that the implementation of machine learning-based techniques can bolster early detection of FMD indicators, enabling proactive measures to mitigate potential epidemics or pandemics. By harnessing the potential of machine learning, African countries can forge a path towards a healthier and more resilient livestock industry, ensuring the well-being of both animals and communities.

 

Featured Bloggers – APET Secretariat

Aggrey Ambali

Justina Dugbazah

Barbara Glover

Bhekani Mbuli

Chifundo Kungade

Nhlawulo Shikwambane

 

[1] https://au.int/en/agenda2063/goals

[2] https://link.springer.com/article/10.1007/bf02217292

[3] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10015193/

[4] https://au.int/sites/default/files/bids/40616-04_TOR_-_FMD_-_LAB_EXPERT_-_ANIMAL_HEALTH-1.pdf

[5] https://au.int/sites/default/files/bids/40616-04_TOR_-_FMD_-_LAB_EXPERT_-_ANIMAL_HEALTH-1.pdf

[6] Grubman, M. J., & Baxt, B. (2004). Foot-and-mouth disease. Clinical microbiology reviews, 17(2), 465–493. https://doi.org/10.1128/CMR.17.2.465-493.2004

[7] https://www.woah.org/app/uploads/2021/12/the-global-foot-and-mouth-disease-control-strategy.pdf

[8] Punyapornwithaya V, Klaharn K, Arjkumpa O, Sansamur C. Exploring the predictive capability of machine learning models in identifying foot and mouth disease outbreak occurrences in cattle farms in an endemic setting of Thailand. Prev Vet Med. 2022 Oct;207:105706. doi: 10.1016/j.prevetmed.2022.105706. Epub 2022 Jul 5. PMID: 35863259.

[9] https://theses.gla.ac.uk/7465/1/2016Casey-BryersPhD.pdf

[10] Punyapornwithaya V, Klaharn K, Arjkumpa O, Sansamur C. Exploring the predictive capability of machine learning models in identifying foot and mouth disease outbreak occurrences in cattle farms in an endemic setting of Thailand. Prev Vet Med. 2022 Oct;207:105706. doi: 10.1016/j.prevetmed.2022.105706. Epub 2022 Jul 5. PMID: 35863259.

[11] Hammond JM, Maulidi B, Henning N. Targeted FMD Vaccines for Eastern Africa: The AgResults Foot and Mouth Disease Vaccine Challenge Project. Viruses. 2021 Sep 14;13(9):1830. doi: 10.3390/v13091830. PMID: 34578411; PMCID: PMC8472200.

[12] REVIEW article, Front. Vet. Sci., 30 June 2023, Sec. Animal Behavior and Welfare, Volume 10, 2023, https://doi.org/10.3389/fvets.2023.1201578https://www.frontiersin.org/articles/10.3389/fvets.2023.1201578/full.

[13] Pacheco JM, Brito B, Hartwig E, Smoliga GR, Perez A, Arzt J, Rodriguez LL. Early Detection of Foot-And-Mouth Disease Virus from Infected Cattle Using A Dry Filter Air Sampling System. Transbound Emerg Dis. 2017 Apr;64(2):564-573. doi: 10.1111/tbed.12404. Epub 2015 Aug 25. PMID: 26303975.

[14] Marshall M, Roger P. Policy and science of FMD control: the stakeholders' contribution to decision making. A call for integrated animal disease management. J Clin Lab Immunol. 2004-2005;53:27-38. PMID: 16805323.

[15] https://rr-asia.woah.org/wp-content/uploads/2019/10/seacfmd-roadmap_2016-2020.pdf.