Over the past two decades, pet nutrition has become big business. Grand View Research says that the global pet food market size $83 billion in 2018 and is expected to grow by 4.5% from 2019 to 2025. The move from synthetic to natural ingredients in dog food has been a primary driver of this trend.
In 2017, Wild Earth, a technology company, launched a clean dog food brand based on sustainable plant-based protein. Today, the vegan dog food market is expected to grow at a growth rate of 12.0% in the forecast period 2021 to 2028.
According to Dr. Darren Logan, Head of Research, Waltham Petcare Science Institute, the pet care industry is nearing a moment when it is focusing on extending pet’s health versus addressing health concerns as they arise.
“If you look at osteoarthritis and joint pain, for example, which are key health concerns often experienced by aging pets – being able to detect the signs and symptoms earlier will assist veterinarians and pet owners in helping to prevent the impact on a pet’s quality of life,” said Logan. “Unlike humans, pets can’t tell us directly their likes, needs, or problems. That means to improve their lives; we need to obtain this information.”
Logan says that has been historically achieved through observation. “For example, an observant pet owner might notice that their dog is scratching a bit more frequently, which could lead to a visit to the veterinarian and result in a diagnosis of a skin disorder. But many owners may not notice the increased scratching until it is quite severe, leaving the dog suffering discomfort in the meantime,” said Logan.
MORE FOR YOU
It’s all in the data
“Today, we collect an array of data about our pets, from their veterinary medical records to their genetic make-up, from their diets to their daily activity profiles. Within that data, there are patterns [..] that can be connected to understand the health and behavior of pets better,” said Logan.
“Artificial intelligence (AI) helps us review these patterns efficiently and effectively by unlocking the latent power of this pet data,” said Logan. “In the case of dogs scratching, we use AI to find patterns of movement in dogs diagnosed with a skin condition, and now we can spot earlier when dogs begin to scratch more frequently, simply by using an activity monitor attached to their collar.”
“Another application of the same technology can be used in spotting when our dogs first begin to develop joint issues like osteoarthritis,” said Logan. “We can use AI to spot subtle variations in gait from activity tracking devices, to identify arthritis in dogs much earlier.”
Logan says that once this is made available to pet owners, it could enable faster access to treatment and reduce the number of dogs in joint pain.
“Ultimately, a move towards preventive care means more healthy and happy time together for pet owners and pets,” added Logan.
Predictive models for animals
Logan says that the common use of AI in human health is developing predictive models where a computer can spot subtle patterns in data that are an early indicator of a health condition or disease.
“Because so many diseases have better outcomes if they are picked up and treated earlier, predictive models have the potential to be transformative for healthcare.”
Logan uses his own cat Morgana as an example. “We combined data from hundreds of thousands of anonymized historical medical records from pets, and we set out to compare the blood and urine profiles of cats that went on to develop a common form of kidney disease and compare these to the profiles from healthy cats.”
“We then developed an AI algorithm that can predict when a cat will develop this disease, up to two years ahead of a traditional clinical diagnosis, simply by reviewing routine annual blood and urine tests,” said Logan. “Unlike traditional diagnostic tests, no additional samples need to be taken from the cat, and the AI can generate the result in seconds.”
“Today, chronic kidney disease in cats isn’t considered curable after diagnosis,” added Logan. “But this AI-based approach now gives us a new two-year window to aid veterinary professionals and pet owners in treating the disease potentially before the cats begin to show clinical symptoms.”
“Time will tell whether this will completely prevent the disease, but we hope it will give us more time with our pets, which we think is well worth the effort, so we are also working on applying similar AI approaches to other diseases in pets.”
Paulo de Castro, CFO global Petcare of Mars Petcare, said that AI is all about preventative care for him. “We’re using the power of AI not only to understand how a pet’s genetics and behavior can reveal powerful insights about their overall health but also to help predict disease in pets so veterinarians and pet owners can better partner to keep pets healthy and happy.”
According to the Association for Pet Obesity Prevention, an estimated 60% of cats and 56% of dogs in the United States are overweight or obese, leading to a range of health issues for pets.
“The company has taken a data-driven approach to examine obesity in pets which can help predict illness in pets early, so they live healthier happier lives,” added de Castro
de Castro says that when pets are brought into the vet, their weight is recorded, but how that information is recorded can vary from practice to practice.
“Sometimes it is recorded as a number, sometimes written in the medical notes and sometimes done as a body condition score – which can also vary between the different types of scores being used,” said de Castro. “This means it can be hard to get a proper set of standards for the weight and body condition score of pets across a large sample.”
To address this, Mars Petcare wrote a Natural Language Processing (NLP) process algorithm for veterinary hospitals that would efficiently pull out this information.
“This approach helped us get a better picture of the scale of the problem, locating 30% more cases of obese pets and discovering a particular issue with overweight and obesity in cats,” said de Castro. “This then enables veterinary professionals to explore the pathways available for pet owners after diagnosis, evaluate which interventions were most successful and support recommended action for pet owners.”
De Casto says the company found that pets on an active weight management program with frequent veterinary examinations within the three to four months after diagnosis could reach a healthy weight faster than those who saw the vet infrequently. “These insights can also influence more tailored reminders for pet owners to help veterinarians start to influence pet and pet owner behavior outside of the traditional clinical setting,” said de Castro.
Pet wearables and AI innovation
Logan believes we will see an explosion of AI tools in the pet industry over the coming years, such as a new generation of pet wearables using AI to interpret the data they collect from pets.
“Activity trackers are already moving beyond tracking the number of steps a dog has taken to identify all kinds of health-related behaviors. For example, tracking scratching and licking to help owners and veterinarians assess if there is an underlying cause when these behaviors are elevated or severe,” said Logan. “AI tools will also be increasingly integrated into routine veterinary practice.”
Logan adds that the Covid-19 pandemic has accelerated video consults’ growth over the last year, relying on the remote collection of data that lends itself to AI-assisted triage.
“AI is well suited to analyze rich data sources like video – and we know that people love to shoot video of their pets,” said Logan. “I expect to see AI-based instruments and apps that can, responsibly and legally, interpret and “translate” the behavior of your pet.”