Novel approaches to improving calf health: new technology and data analytics

Abstract

Calf disease significantly impacts both welfare and production. Bovine respiratory disease (BRD) and neonatal calf diarrhoea (NCD) commonly affect calves and are frequently missed by farm staff.

Activity and milk feeding variables have been associated with BRD and NCD and are being investigated for use in disease detection tools. Saliva may also hold some potential for NCD detection.

Group housed calves had their activity and milk feeding variables recorded by a tri-axial accelerometer and automatic milk feeders and were scored for BRD in two trials. In the second trial, calves were scored for NCD and dehydration daily and had blood and saliva samples taken weekly and when they developed NCD. Haematocrit was measured in all blood samples and electrolytes and serum proteins in a subset of thirty. All saliva samples had both pH and conductivity measured.

Four datasets were created. Healthy calves from the first trial were selected by excluding data three days either side of a diseased or intermediate BRD score or treatment event, giving a dataset of 31 calves with 10 days each. The association between intrinsic calf factors and activity and milk feeding variables was analysed. Calves with BRD were paired with a healthy calf on the same date, of the same sex, within seven days of age and ten kilograms of weight. Three days either side of peak disease were also taken for analysis. Mixed models analysed the association between BRD and activity and milk feeding variables.

The second trial’s dataset was cleaned to remove the confounder of BRD and split into laboratory and behaviour variables. The dehydrated calves were matched with a healthy calf and a NCD and normally hydrated calf. The association between NCD with or without dehydration and saliva pH and conductivity, haematocrit and total protein were analysed using mixed models. Lin’s correlation concordance was correlated between saliva pH and haematocrit in the larger dataset and saliva pH and strong ion difference in the smaller dataset. Mixed modelling was used to analyse the association of NCD with or without dehydration on milk feeding and activity variables.

Heavier healthy calves visited the milk more often for shorter visits and drank faster than their lighter counterparts. Older healthy calves had shorter lying bouts and were less active than younger calves.

Summer born healthy calves lay down for longer, had more lying bouts and, more shorter standing bouts each day, than healthy calves born in autumn or winter. Healthy male calves drank more slowly, were less likely to consume their milk allowance and had longer lying bouts than healthy female calves.

Calves with BRD visited the milk feeder for longer for the three days following peak disease and drank more milk per feed on days -3 and -2 relative to peak disease. No other comparisons between diseased and healthy calves on any day were significant.

Differences between these finding and those in the literature may be due to nutrition.

Saliva pH but not conductivity was associated with NCD, regardless of hydration status. Saliva pH was not associated with haematocrit or strong ion difference. Calves with NCD, regardless of hydration status were less active, less likely to consume their milk allowance and drank more slowly than healthy calves. Calves with NCD that were normally hydrated had fewer, longer lying bouts and visits to the milk feeder and more standing bouts than healthy calves.

Dehydrated calves with NCD had longer visits to milk than either their normally hydrated diarrhoeic or their healthy counterparts.

In conclusion, young calves do show differences in behaviour when they are sick. Saliva pH is associated with diarrhoea so may have potential for a disease detection tool but the mechanism is unclear.