In veterinary medicine, patients can’t speak up to describe what is wrong and where it hurts. But new technologies such as big data and the animal Internet of Things (IoT) are emerging to help give pets and production animals a voice.
The IoT refers to the collection of physical devices—sensors, wearable devices, etc.—that connect via the internet and share large amounts of data. For humans, this includes a vast array of devices, from your smart phone to a smart doorbell. The application for animals may not be as widespread, but it can have an impact on veterinary medicine.
Veterinary medicine is in the early days of tapping the potential of data and new IoT monitoring devices such as wearables, says Aaron Massecar, assistant director of continuing education at Colorado State University’s Translational Medicine Institute and former executive director of the Veterinary Innovation Council.
“Veterinarians are always producing information, and it’d be amazing if you could collate all these medical records and generate actionable insights about animals,” says Massecar. “For instance, how many incidents of cancer in golden retrievers do we have? How often are they getting bladder stones? How often are animals on grain-free diets getting dilated cardiomyopathy?”
The Standardized Coding Barrier
The problem is that veterinary medicine does not have the standardized coding that exists in human medical records. As a result, you can’t always collate data from different doctors at the same practice, let alone across practices. There is potential to use advanced artificial intelligence (AI) techniques like natural language processing (NLP) to dig through records. “There’s a project called DeepTag that was taking reams of medical records, running some NLP algorithms, then distilling that into quantifiable data sets,” says Massecar. “If we could do that on a broad scale, you could drive insights from that, but nobody’s really doing that.”
Another interesting benefit from medical record analysis pertains to disease tracking, especially with zoonotic diseases. “A vet might look at an animal presenting with a particular condition and think it’s strange, it might be a rare disease,” says Massecar. “If you had 2 doctors saying that at the same time in closely related practices, you might be able to identify it’s a tick-borne disease, or a seasonal thing.”
Sourcing Data
One of the most promising ways to generate big data from animals, with the goal of doing what’s in the best interest of the animals, is IoT technologies like wearables. Wearable technology, such as smart collars, provides a way to monitor animal behavior and detect anything that’s out of the ordinary.
“Some of these wearables are doing a really good job of producing data sets and dashboards for their animals, everything from data around movements and scratching, to heart rates,” says Massecar. “If veterinarians saw these as opportunities for pre-op or post-op care, and get year-over-year information about the health of their animal, they could move away from the subjectivity associated with owners taking those measurements, breathing rate, heart rate, and activity level of their animals and build a reputation as a progressive, forward-looking practice.”
The Current State of Wearables
There are 2 ends of the spectrum: consumer-driven wearables and diagnostic devices. But some convergence is happening between them. Vetrax (vetrax.com), for instance, started as an activity tracker, but expanded to look at the relationship between pet activity and diet after forming a partnership with Hill’s Pet Nutrition.
Another wearable collar provider is Whistle (Mars Petcare, whistle.com). Whistle partnered with the Pet Insight Project and Banfield Animal Hospital to conduct a large-scale study of pet health that examined the links between pet behaviors, activity, and health.
“Whistle tracks pets’ health, location, and fitness, and provides access to on-demand televet services. By combining pet science and technology, Whistle trackers and its app help pet parents care for their best friend in ways not previously possible,” says Scott Lyle, head of the Pet Insight Project. “Tracking scratching and licking is one of the most important and often highly praised features, since they can be tied to more serious issues from unknown allergies to dermatological issues. We’ve made it easy for pet parents to seamlessly convey health updates to their local vet by easily emailing a 30-day report directly to their vet through the app.”
The Pet Insight Project has distributed approximately 100 000 devices and collected 16 billion minutes of accelerometer data. “The way we’re researching this is by correlating high-resolution Whistle accelerometer data with the pet’s health records to tell a health story. This combination allows us to map relevant patterns of pet behaviors and movement with illness, to aid in more precise detection of ailments so we can provide early warnings,” says Lyle.
Production Animal Wearables
In some ways, the IoT scene for large, food, or production animals is ahead of companion pets.
One reason is the considerable return on investment. “Food animal producers, specifically in the dairy industry, track key metrics that are important to the animal’s health. For instance, these metrics can measure rumination time, milk production, and even alert that a cow is approaching a targeted time for artificial insemination,” says Aaron Wallace, co-CEO at Lacuna Diagnostics, Inc, a digital pathology provider. “Dairy producers need to maximize the period an animal can produce milk while balancing optimal health and decreasing stress on the animal. Tracking key performance indicators from wearables helps maximize production and health while alerting producers if data suggests medical or other intervention is needed. In a sense, the data is helping producers and veterinarians monitor animals while also perfecting the reproductive cycle. When this cycle is off, the ROI goes down.” CowManager (cowmanager.com), an ear tag that has an accelerometer similar to a Fitbit, is one such device. It tracks temperature, activity, rumination, eating, and resting time in order to provide insight into health and fertility.
Farm Jenny (farmjenny.com) is another venture in the equine and bovine space. Its founders, Pennsylvania-based Rob and Tammy Crouthamel, built Farm Jenny after more than 25 years of experience in big data and wearable technology, along with personal experience raising and caring for horses.
“Tammy has a good gut sense of animals, but we were missing things when we couldn’t be on the farm,” says Rob Crouthamel. “When obvious signs of disease show up, the only option is to call an emergency vet, but if you detect those subtle signs earlier, you can take steps to avoid the crisis.”
Farm Jenny is designed to recognize those subtle changes in behavior as early and as accurately as possible. It monitors behaviors such as activity level and compares them to the other animals in the herd. It’s moving toward detecting more specific, subtle behaviors such as head motions, kicking, and biting the side.
Its wearable device is a sensor attached to the horse’s soft headband or a cow’s ear tag, and uses a gyroscope to detect motion. The sensors can run for months on battery without maintenance. They connect to solar-powered field receivers around the farm, then via cellular connection into the cloud. “Connectivity is particularly challenging on farms, where you can’t expect Wi-Fi. One of our unique plays is cellular connectivity and mesh connectivity using Thread,” says Rob. “Wherever the animal is on the farm, we continuously get that stream of data uploaded to the cloud.”
Farm Jenny takes that constant stream of each animal and analyzes the behavior. If it senses any behavior out of the ordinary, it can alert the farmer’s smartphone. The historical data is where the real value comes. The system is designed to meet the needs of small to midsize operations, from a single horse and stall up to hundreds of acres and thousands of animals. There are a couple of ways Farm Jenny provides data access to a veterinarian. Multiple people can sign in, so a single veterinary professional can have continuous behavioral tracking of all of their clients’ animals. Farm Jenny also has machine learning–based behavior detection that correlates sensor data and camera footage, and offers that as a standalone product to universities, research hospitals, and pharmaceutical companies.
A Transmitter in the Litter
The IoT goes beyond wearable collars for feline patients. “Since 2008, and possibly longer, feline patients simply do not routinely visit veterinary hospitals nearly as often as dogs do,” Wallace says. “One of the many exciting IoT advancements currently being developed are digital litter boxes. You can measure how frequently a cat is going to the litter box and how long they’re spending in the litter box. It also alerts the owner if the cat is going too frequently, which could be a sign of UTI or constipation. The boxes can also alert the owner if the cat is spending too much time in the litter box, which could be an early indication of urethral obstruction, a common life-threatening medical problem we treat in the veterinary E.R.”
In-clinic workflow is another area ripe for innovation with wearable technology, says Dr. Caleb Frankel, practicing veterinarian and founder of Instinct Science, an electronic medical records and workflow platform.
“I don’t think anyone is working on the in-clinic side,” says Frankel. “One of the problems for every practice is staffing. Smart devices that can automate vital sign gathering during the check-in process could help. You might even improve the accuracy of these measurements, since it’s often stressful for pets at that initial part of a visit. It seems like a huge opportunity for wearable devices that can augment staff, improve patient care, and enhance workflow.”
The Next Steps
“When we think about big data in veterinary medicine, our patients can’t tell us something is wrong,” says Frankel. “One of Lacuna’s mantras is ‘time is trauma.’ We believe if we can diagnose cancer or infection faster, our pets can get treatment faster, which leads to a better medical outcome with less suffering.”
If a wearable collar tells you a patient’s heart rate is faster than normal and activity is down, the client can be alerted to schedule an examination. Or if a million cats use digital litter boxes, the ability to examine that data to see what behavior and clinical signs serve as a precursor for emergency treatment. Taking that data and applying AI techniques could even allow veterinarians to diagnose emergencies before they emerge.
In the big picture of big data and the animal IoT, we’re in the early phases. The costs of AI are high. “Currently, everyone who obtains a degree or has experience with AI has several career opportunities out there from self-driving cars to human medicine, but I look forward to the time when we have more experts working on these neural networks in the veterinary space,” Wallace says. “In 5 to 10 years, veterinary medicine will have a whole different look through the eyes of data than we have now. When we look at IoT, eventually we will have a collar or device that accurately measures and constantly surveys patient activity, appetite, gait, respiration rate, and many other important data points that can assist pet parents before medical conditions progress to emergencies. It is truly exciting to think about how we can combine IoT devices with big data techniques, for it will most certainly lead to new ways of diagnosing. These are topics the veterinary industry needs to see.”