Housed dairy cattle feeding in shed

Using dairy cow behaviour for early detection of disease

Take home messages:

  • Certain behaviour changes can be linked to specific illnesses - for example, cattle with acute lameness spend less time at the feeder in the week before diagnosis
  • Changes in behaviour for some illnesses can be detected before there are any clinical signs
  • Wearable technology for livestock has the potential to develop robust early-warning systems

Disease can have significant effects on farm economics, from payment for treatments and loss of production from the infected animals, as well as on animal welfare. So, what if you could detectthe onset of common diseases before any symptoms are even visible? With the use of wearable technology for cattle and an understanding of dairy cow behaviour, we are beginning to be able to do just that. 
What do we know about disease behaviour?

As in humans, illness affects animal behaviour. Sick animals will normally be more lethargic, isolate themselves and lose their appetite. 

Most of our knowledge about how sickness affects behaviour of dairy cows is based around their activity and feeding behaviour.

Lame cows are known to spend longer lying down, change their weight distribution and walk at a slower speed compared to healthy animals. 

Dairy cows with mastitis idle more and spend less time lying down, feeding, ruminating and grooming. It is also suggested that, as chronic illnesses appear to instigate more behavioural changes, monitoring of this can distinguish whether an animal has an acute or chronic illness.
Does behaviour change before diseases are visually detected?

In most cases, research has found that feeding behaviour seems to be a relatively reliable predictor for disease onset.

Cattle with acute lameness have shown a reduction of time spent at the feeder of around 19 minutes every day for the week before lameness was visualised, along with a reduction in the number of visits to the feeder. 

Cows who develop metritis 7-9 days after calving, spend less time at the feed bunker before they have even calved. 

For animals with ketosis, it is clear that feeding and activity behaviour changes, with a reduction in feed intake in the 3 days before diagnosis and a 20% increase in standing time in the week before calving. 

However, high variability has been noted for feeding behaviour linked with udder diseases, and so would not be a good predictor of disease development for mastitis. 

In all cases, it is unknown if these changes in behaviour are a cause or effect of a developing illness, is it the reduced feed intake that causes the disease, or is this just an early effect of the illness? Further work is needed to understand this, and an understanding may influence selection traits in the future.

What should I look out for?

Changes in feeding behaviours, like a reduction in time spent eating, are good indicators that an animal is developing an illness. 

However these changes are often quite subtle, or require animals to be watched and recorded all day, which is why the development of wearable technology for cattle is becoming so useful.
How can we detect these changes?

With the rise of precision livestock farming, there are many wearable devices available on the market that detect changes in behaviour amongst other variables. 

Most of these measure activity through accelerometers fitted to the leg or neck, or rumen function via electronic boluses. Farm devices such as walkover force sensors can also be utilised to assess gait consistency. 

Many of these devices are programmed to detect changes such as coming into oestrous, although most are less well known for their ability to detect changes in behaviour for early detection of diseases. 

Many devices are able to detect potential illness at a basic level, and by flagging animals with a warning someone can be alerted to go and investigate what may be wrong.

To improve on the ability to detect a specific disease and the accuracy of these devices, the behavioural research completed to date must be utilised to develop mathematical predictor algorithms. 

These algorithms will then be able to use and simplify the data that the sensors collect to provide a simple notification system to alert that a cow is likely to develop a certain illness. 

This should mean that the cow in question will be closely monitored or examined further to catch the disease before any visual symptoms develop. 

Although systems have been researched to detect diseases such as lameness and mastitis, no system will be truly useful to everyday practice until a simplified answer can be provided from the depth of data that will be needed to determine the status of the cow.

Read the full article

Article published with the kind permission of Farming Connect

By Dr Ruth Wonfor: IBERS, Aberystwyth University