Factors affecting feeding and activity behaviours in healthy pre-weaned artificially reared calves

Application: An understanding of factors affecting healthy calf behaviour will guide the development of behaviour-based tools for disease detection. Introduction: Behaviour is increasingly being researched as a tool for early disease detection in artificially reared calves. However, it is important to understand how other factors such as sex, and breed may affect healthy calf behaviour. This is to improve the sensitivity and specificity of behaviour-based tools, as alert accuracy is a major concern for animal keepers when discussing technology-based disease detection. Materials and methods: One hundred and fourteen dairy bred calves from two herds (herd A = 43 calves, herd B = 71 calves) entered group pens of 12-14 calves at approximately seven days of age and left the study when weaning began at 40 days of age. Calves were bedded on straw with ad-libitum access to racks of straw and concentrate and a single automatic milk feeder with an automated weighing platform (Biocontrol). Seven litres of acidified milk replacer (mixed at a rate of 150g/L) were available daily. Activity behaviours were recorded using a leg-mounted 3-axis accelerometer (Iceqube). Health was scored daily using the Wisconsin calf health score method. To ensure only days where calves were healthy were included, Data was removed for days when medication was given, health scores were intermediate or high, and for 3 days either side. Calves with at least 10 consecutive healthy days were included and the middle 10 days of the healthy period were taken to achieve a balanced dataset. For parameters with a normal distribution a generalised linear mixed model was constructed using REML in R with animal number nested within group as random effects. Fixed effects tested were live weight, age, sex, herd, season of birth, age of inclusion into the group, dam parity, birth weight and sire breed type (beef or dairy). Model selection was carried out using the Akaike information criterion. Results: Thirty-one calves were included in the analysis. Live weight, age, sex, herd, season of birth and age of inclusion into the group had significant effects on activity or milk feeding behaviours. Dam parity, birth weight and sire breed type had no significant effects on the behaviours studied. The results of the final models are presented in Table 1. Conclusion: When building behaviour-based disease detection tools, models should account for other calf factors including weight, age, and sex to improve accuracy. While there has been much research into changes in calf behaviour with social housing and greater milk allowances but there is very little literature on other factors that affect behaviour in pre-weaned calves.

Table 1. Factors affecting behaviour in healthy pre-weaned calves with P < 0.1 in the final model.

Behaviour Fixed effect Level Effect size Confidence interval P-value
Lying bouts /day Season Autumn Reference Reference Reference
Summer 6.8 2.6–10.8 0.01866
Winter −0.4 −3.4–2.7 0.78657
Herd B Reference Reference Reference
A 4.0 1.1–6.5 0.00816
Sex Female Reference Reference Reference
Male −3.2 −6.4–0.1 0.07107
Motion index/day Age −61.5 −96.2–−25.8 0.000883
Sex Female Reference Reference Reference
Male −653.98 −1360.5–3.3 0.065207
Motion index units/ standing bout Age −3.1 −4.9–−1.0 0.00143
Age at inclusion into group pen −11.2 −18.88–−2.2 0.00652
Visits to milk/day Weight 0.08 0.02–0.14 0.00727
Volume milk/ visit Weight −5.4 −11.1–−0.3 0.0454
Acknowledgements
This study was funded by Innovate-UK, EASTBIO DTP and AHDB.