ShapeShapeauthorShapecrossShapeShapeShapeGrouphamburgerhomeGroupmagnifyShapeShapeShapeShape

How will artificial intelligence change farm practice?

After 14 years in practice, Phil Elkins tells us why he made the move to developing innovative farm health solutions

21 October 2019, at 12:00pm

Phil Elkins qualified as a veterinary surgeon in 2005; he went into mixed practice and then farm practice in the UK and New Zealand, before settling at Westpoint Farm Vets in Cornwall in 2009. Phil opened a second branch in the West Country, where he worked for 10 years, growing the branch to a team of nine vets. Phil was director of Westpoint Farm Vets for his last three years with the practice and Chair of the Clinical Governance Board for the last two. In May 2019, he made the decision to leave Westpoint and join the team at Prognostix – an innovation-driven animal health company that aims to improve disease prevention and boost performance of farm animals.

What were the key drivers in your move from veterinary practice to animal health innovation?

The number one factor was the direction that Prognostix is looking to head in. We’re trying to take veterinary knowledge and put a different spin on it, and trying to adapt and drive solutions to future challenges.

The clinical governance within Westpoint is something I am incredibly proud of and I am really happy to be associated with the business. I was just looking at the clinical veterinary practice and seeing a lack of investment in innovation; corporatisation can bring a lack of willingness to take risks and a view towards short-term returns.

Innovation always comes with a risk and the returns are more long term. You need to see things through and accept the fact that you might not get any money out of it for the next 12, 24 or 36 months. This can be a hard sell to a business that is ultimately based on giving returns to its investors.

You’ve also got a situation where farmers are looking to change their relationships with vets. The economics of a vet treating a sick animal generally don’t fare as well as lots of other investments on the farm. There’s a lot of rhetoric about it being more about the prevention of disease, which was probably the new thing 10 years ago. Now, it’s much more about optimisation of performance – with prevention of disease forming part of that.

My concern is that traditional veterinary practice is putting itself in a difficult place by not innovating quickly enough and allowing associated industries a head start in these areas.

How will Prognostix improve farm animal practice?

Prognostix is taking data from multiple sources, some of which are our own hardware and software, and analysing it to give vets, consultants and farmers the tools to predict when diseases are likely to happen. We will define early warning signs for disease – even for things that are going to be precursors for disease – and, taking that to the next stage, early warning signs for things that are going to challenge performance. This should provide the tools on the farm basis to optimise the systems.

For the main project at the moment – which is about respiratory disease in calves – on a farm basis, we can predict which calves are going to get sick because their feeding behaviour changes a couple of days beforehand and their temperature starts to peak (which we see through the constant temperature monitoring via rumen boluses). We can tie that in with environmental sensors and say that we know that on a particular farm, when the temperature in the shed goes above a threshold, you’re likely to get disease in a week’s time. You can then put a flag up to the farmer and say, “get the temperature down quickly to stop the animals getting sick in the first place”.

We can throw in some weight data, and add, “we know your animals grow best when your temperature and humidity are at this level, or light intensity is at this level”, and then start controlling those factors, and getting the animals performing as well as possible.

Is data collection and benchmarking the basis of the product?

There are a number of tools for collecting data; if it was just that, I don’t think it would have been enough to drag me away from what I was doing. The big thing is about incorporating artificial intelligence and machine learning into the software so that you can actually turn this into useful advice at the farm level.

There are papers out there that say that, for example, if the temperature gets below 10ºC, with the humidity over 85 percent, cows are more likely to get pneumonia. The reality is that the work has been done on a relatively small number of farms and individual scenarios, and each building behaves differently. If you incorporate machine learning and artificial intelligence, you can start to make those cut-offs and thresholds farm-specific.

Will you be expanding the focus from poultry to cattle? Can you tell us about the Y-ware project?

The poultry side is definitely more advanced than the ruminant, but my role is entirely ruminant. The Y-ware project is an Innovate-funded project looking specifically at respiratory diseases. And for me that is the starting point. The initial point is to start looking at data and thinking: how far can we go with this? What algorithms can we build with this? Let’s define as wide a range of normal as possible so we can start defining abnormal – and looking at what factors contribute to that. That funding has allowed us to take the initial push, but the project is far bigger and far wider than that for me.

Are your products going to be negatively impacted by the low strength of mobile signal in rural areas?

We are ideally not looking to limit ourselves to 4G cellular networks. We’re going to be using different communication pathways, which shouldn’t be affected by mobile signal. Rural broadband connectivity may well play a role in how much we can achieve on certain farms. I suspect, given personal experience, that the cases of abysmal broadband are fewer and further between than the media likes to make out.

How will you use the big data collected?

We’re looking at being as open as possible with the data. We don’t see ourselves as being in competition with those who have management software or vet analysis programmes. We’d like to work in a way that has two-way communication between the programmes, and allow open access to that data.

How do you think the tech solutions are going to change large animal practice?

think we’ll find a number of businesses embracing data and technology, and they will be providing a higher level of service to their clients. That doesn’t mean that there won’t be vets going out to do calvings, caesareans, sick animals, TB testing, etc. It means there will be a cohort of vets within those practices that will be using this data to advise farmers much better on how to prevent disease and how to optimise returns.

Did you change any of your practices as a result of your time working in New Zealand?

We’re all products of our experiences; at that time I was about two to three years graduated and I think we subconsciously adapt what we’re doing based on the experiences we have along the way. There are some real positives that I picked up from working in New Zealand. And there are some real concerns – which is probably the reason why I’m not still there.

Working in New Zealand gave me a much better ability to look at things through multiple and different viewpoints, and to try and force myself to have an open mind about things rather than assuming that things need to be done a certain way. I think this can help you grow as a person as well as an advisor and a communicator.

Do you have any tips for what better communication between farmers and vets would look like?

The biggest challenge I had as a manager of a branch – a practice principal as we would call it – is matching personalities of vets with personalities of clients. It comes down to relationships and it takes time to work out those relationships. Some people communicate better in some ways than other ways. And that works equally for vets and farmers.

I think the best thing you can do around communication is use multi-modal communication with your farmers, and then talk to them about what works best for them. Is it emailing, texts, phone calls, a sheet of paper, stuff that’s laminated, pdf reports?

For some farmers, all they want is to be told what to do. They don’t necessarily want to understand the reasons behind it. But there are other people who need to be taken on a journey before they make a decision. They need to be convinced with every little bit of evidence out there. You’ll discuss something and then a month later they’ll show up with a Farmers Weekly report from six years ago that says something slightly contradictory and they’ll want you to respond to it, and then four years down the line you’ve found that something has happened.

We’ve got some of the Prognostix kit going in at one of the farms I used to work at. I spent eight years trying to convince them to put water buckets down for their calves from day one. I went back there after not being on the farm for a couple of months, and every cow has a water bucket in front of it... What happened to instigate that change? I don’t know! Maybe it was me not telling him to do it.

To me, the only difference between a good vet and a great vet is the ability to enact change on a client. And that’s far deeper than any article will ever cover!