What if we could identify the biosecurity practices most likely to be adopted on a dairy farm to improve animal health and welfare? This is the question guiding the work of Faustin Farison, a doctoral student in veterinary science at the Université de Montréal. To answer it, he is studying how recommendations can be better adapted to the realities of dairy farms, using a tool that is still rarely used in this context: artificial intelligence.
By Mélanie Lagacé
Originally from France, Faustin developed an interest in animal health at an early age, influenced by his grandfather, a rural veterinarian. “I often accompanied him, and I’ve always maintained a connection with the dairy sector,” he says.
After studying agronomy at Bordeaux Sciences Agro and then epidemiology at the l’École nationale vétérinaire de Maisons-Alfort, he gradually shifted his focus toward disease prevention. “I liked the idea of taking a step back to better understand what’s happening at the herd level,” he explains. His experiences, particularly with bovine tuberculosis and his collaborations with farmers, reinforced this interest.
Then, an internship at the Office français de la biodiversité served as a turning point. “I discovered that I could help develop tools useful for fieldwork.”
It was ultimately in Quebec that his career took a new turn. Following discussions with professors Juan Carlos Arrango Sabogal and Simon Dufour, he began a Ph.D. at the Faculty of Veterinary Medicine (FMV) at the Université de Montréal, within the Research Chair in Biosecurity in Dairy Production, co-directed by the two professors. The project’s supervisory team is rounded out by Professor Pablo Valdes-Donoso.
His arrival in Montreal required him to adapt to different ways of doing things and to a new professional and cultural environment. “The methods are different, but the transition went very smoothly,” he notes. He quickly found his place in an ecosystem where there is plenty of interaction and strong ties to the field.
Alongside his research, Faustin laces up his cleats with the Montreal Carabins, a top-tier university sports team. Rugby is an integral part of his daily life. “It gives structure to my days and helps me unwind,” he says. It’s a discipline he also applies to his approach to scientific work.
Simple actions… with big impact
On dairy farms, biosecurity relies on a series of concrete actions aimed at limiting the introduction and spread of pathogens. Visitor management, screening of incoming animals, hygiene: all these measures, when combined, help protect the health of the herd.
“Biosecurity is a bit like the preventive measures we take in human health, but applied to livestock,”
Faustin Farison
These measures have very real effects. The outbreak of a disease in a herd can lead to significant losses, both in terms of animal welfare and economically. Hence the importance of taking action early on.
To better understand current practices in the industry, Faustin draws on the risk assessment questionnaires from the Canadian proAction* program, which are completed jointly by veterinarians and producers. These questionnaires provide a detailed picture of current practices on dairy farms across the country. As part of his doctoral research, he is analyzing the responses collected in Quebec.
But one challenge remains: How to prioritize recommendations?
In practice, veterinarians can only propose a limited number of measures at a time, usually one to three. It therefore becomes crucial to target those most likely to be implemented, based on the specific context of each farm.
That’s where artificial intelligence comes into play.
When practices interact
Rather than considering each measure in isolation, Faustin focuses on how they relate to one another. In the field, interventions aren’t adopted at random: some naturally go hand in hand.
“The idea is to identify practices that are frequently implemented together and to use these combinations to guide the actions to be taken,” he says.
To achieve this, he uses a method called association rule learning, based on algorithms. Well-known in marketing, particularly for analyzing grocery shopping baskets—for example, when a consumer chooses milk and flour and naturally adds eggs—it allows us here to highlight connections between different biosecurity measures.
In other words, when certain practices are already in place, it becomes possible to estimate which ones are most likely to be adopted in the future.
The study is based on 3,825 questionnaires from Quebec livestock farms. Millions of configurations were examined and then filtered to retain only the most relevant associations. Ultimately, 63 relationships between measures were identified. Some confirm what stakeholders observe in the field. Others, less intuitive, shed new light on the situation. In several cases, the probability of adoption exceeds 70%, allowing for more precise guidance in recommendations.
“What’s already in place provides good clues as to what a producer is more likely to adopt,” explains Faustin. The goal is not to replace veterinarians’ expertise, but to support it. “We’re adding complementary data to guide the guidelines.”
By combining on-the-ground expertise with the analysis of available information, it becomes possible to propose actions that are both relevant… and realistic.
Toward data-driven agriculture
Faustin’s research is part of a broader trend: that of an agricultural sector where data analysis plays an increasingly important role in decision-making.
Here, the benefit is clear. The data is there, but its potential remains underutilized.
“The challenge is to make full use of it,” he emphasizes.
Ultimately, this type of approach could help make biosecurity measures more effective by better accounting for the specific context of each farm. The idea is not to add constraints, but to propose actions that integrate naturally into existing practices.
His work continues in this direction, with the goal of deepening the analyses and putting these results to practical use.
“What interests me is that it can be useful in real-world contexts,” he concludes.
After completing his PhD, Faustin hopes to continue working at the intersection of research and application, contributing to the development of tailored tools to support farmers and industry professionals.
* ProAction® : a mandatory Canadian quality assurance program in the dairy sector, established by Dairy Farmers of Canada. It governs on-farm practices across six areas: milk quality, food safety, animal welfare, traceability, biosecurity, and the environment.
To learn more about the project
- View the Research Chair in Biosecurity
- Visit the Chair’s website
- Read the popular science article written by Faustin and the research team
