Use of advanced technologies to enhance monitoring of dairy cow health
Lead Research Organisation:
University of Edinburgh
Department Name: Royal (Dick) School of Veterinary Scienc
Abstract
Mastitis and lameness are common conditions in dairy cows worldwide, resulting in significant economic losses and reduction in animal welfare. Recent studies suggest that the current UK clinical mastitis case rate is around 40-70 cases per 100 cows per year, with lameness prevalence around 50% based on mobility scoring data. Early detection and treatment of these diseases in dairy cows will reduce the economic losses, as well as hasten recovery and improve animal welfare.
Most cases of such endemic diseases on dairy farms are treated by the stockman (rather than by veterinarians), and it is widely acknowledged that early detection and diagnosis of such conditions by farm staff is poor. Such issues are exacerbated by the recent trends in expansion in UK dairy farms, with more cows and fewer staff resulting in more cows per staff member. There is simply not enough time during the day for the close inspection and examination required for the early detection of disease issues.
IceRobotics have developed automated remote behavioural monitoring systems for dairy cows which are commercially available for dairy herds (CowAlert). The system has been designed for health alerting but has so far been predominantly used by commercial dairy farmers for oestrus detection alone. The practical application of behavioural monitoring in specific disease conditions has not been studied in detail. For example, it has been shown that lameness and mastitis in dairy cows will affect cow behaviour and activity (Chapinal et al., 2010; Medrano-Galarza et al., 2012) when comparing the behaviour of affected and unaffected animals following farm detection. However no long-term prospective studies have looked at the ability of such systems to detect lameness and mastitis, what stage they might be able to detect such disease (for example prior to identification by farm staff), and the benefits of early intervention at this stage.
The University of Edinburgh's Langhill Farm has 240 adult Holstein milking cows averaging 10,000 litres of milk per lactation, and are mobility scored every 2 weeks by a trained technician, with weekly/fortnight foot trimming sessions to identify and treat lame cows. Based on data from 2014-15, there are approximately 100 new lameness cases and 80 new clinical mastitis cases arising annually.
The project will use the IceRobotics behavioural monitoring system CowAlert on all cows at Langhill Farm for the early detection and treatment of key production diseases in dairy cows (lameness and mastitis) with the objectives of:
1) Identifying cases of clinical and subclinical mastitis, and lameness in dairy cattle using the automated system in comparison with traditional methods such as mobility scoring, human observation and individual cow somatic cell counts.
2) Examining whether other less common disease conditions (metritis, endometritis, ketosis) can be accurately identified using the automated system.
3) Examining behavioural data relating to individual cows prior to farm detection, determining at what stage it is possible to diagnose disease conditions using cow behaviour alone, or in combination with other diagnostic criteria (for example daily milk yield).
4) Determining whether or not early diagnosis and treatment of affected cows based on behavioural monitoring results in significant positive benefits for production, cow health and animal welfare.
Most cases of such endemic diseases on dairy farms are treated by the stockman (rather than by veterinarians), and it is widely acknowledged that early detection and diagnosis of such conditions by farm staff is poor. Such issues are exacerbated by the recent trends in expansion in UK dairy farms, with more cows and fewer staff resulting in more cows per staff member. There is simply not enough time during the day for the close inspection and examination required for the early detection of disease issues.
IceRobotics have developed automated remote behavioural monitoring systems for dairy cows which are commercially available for dairy herds (CowAlert). The system has been designed for health alerting but has so far been predominantly used by commercial dairy farmers for oestrus detection alone. The practical application of behavioural monitoring in specific disease conditions has not been studied in detail. For example, it has been shown that lameness and mastitis in dairy cows will affect cow behaviour and activity (Chapinal et al., 2010; Medrano-Galarza et al., 2012) when comparing the behaviour of affected and unaffected animals following farm detection. However no long-term prospective studies have looked at the ability of such systems to detect lameness and mastitis, what stage they might be able to detect such disease (for example prior to identification by farm staff), and the benefits of early intervention at this stage.
The University of Edinburgh's Langhill Farm has 240 adult Holstein milking cows averaging 10,000 litres of milk per lactation, and are mobility scored every 2 weeks by a trained technician, with weekly/fortnight foot trimming sessions to identify and treat lame cows. Based on data from 2014-15, there are approximately 100 new lameness cases and 80 new clinical mastitis cases arising annually.
The project will use the IceRobotics behavioural monitoring system CowAlert on all cows at Langhill Farm for the early detection and treatment of key production diseases in dairy cows (lameness and mastitis) with the objectives of:
1) Identifying cases of clinical and subclinical mastitis, and lameness in dairy cattle using the automated system in comparison with traditional methods such as mobility scoring, human observation and individual cow somatic cell counts.
2) Examining whether other less common disease conditions (metritis, endometritis, ketosis) can be accurately identified using the automated system.
3) Examining behavioural data relating to individual cows prior to farm detection, determining at what stage it is possible to diagnose disease conditions using cow behaviour alone, or in combination with other diagnostic criteria (for example daily milk yield).
4) Determining whether or not early diagnosis and treatment of affected cows based on behavioural monitoring results in significant positive benefits for production, cow health and animal welfare.
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| BB/M010996/1 | 01/10/2015 | 31/03/2024 | |||
| 1806611 | Studentship | BB/M010996/1 | 01/10/2016 | 30/09/2020 |
| Description | The objective of this study was to use automated behavioral monitoring under commercial farm conditions to describe the behavior of dairy cattle in late gestation, to quantify any behavioral differences between primiparous and multiparous cows, and to quantify any behavioral differences between assisted and unassisted calvings. Data were collected from 32 multiparous and 12 primiparous Holstein dairy cattle to describe normal calving behavior. To quantify behavior related to calving difficulty, the data from 14 animals that were assisted at calving were matched to cows that had an unassisted calving based on parity, locomotion score, calf breed, calf sex, month and year of calving. An IceQube (IceRobotics Ltd., South Queensferry, United Kingdom) was fitted to the right hind leg of cows 4 weeks prior to their expected calving date. Data for lying time, standing time, number of steps, motion index (total motion) and the total number of standing and lying bouts (postural transitions) were automatically collected and summed into 15-minute blocks. Behavioral variables were summarized into -2h and -24h periods prior to analyses. Mixed-effect models were used to analyze cow behavior in the last four days before calving (d -4 to d -1), and on the day of calving. In the 4 days prior to calving, primiparous cows lay down an average 2.8h less per day than multiparous cows, had an average of 9.1 more postural transitions per day (37.7 ± 1.2 vs. 27.6 ± 0.7), had an average of 172 more steps per day, and had a higher motion index (2743.5 ± 94.9 vs. 2071.9 ± 59.6). In the last 24 hours prior to calving, assisted cows had 16.5% more postural transitions on the day of calving compared to unassisted cows (51.7 ± 3.4 vs. 43.8 ± 48 3.6). Piecewise regression analyses concluded that the number of postural transitions occurred earlier on the day of calving for assisted cows (-11.0h) compared to unassisted cows (-8.5h).These findings indicate that the number of postural transitions could be used as an indicator of animals that are experiencing calving difficulty, and parity should be considered when predicting the day of calving. |
| Exploitation Route | As herd size and pressure on farm staff time increases, there is a need to apply user-friendly, automated technology to facilitate herd management. Remote sensing devices have the potential to improve animal behavior monitoring as they can continuously and automatically measure animal activity without altering the animal's natural behavior. The findings of this project could be used by IceRobotics to create a product that can be used on farm to automatically detect calving. |
| Sectors | Agriculture, Food and Drink |
| Description | Open Farm Sunday |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Public/other audiences |
| Results and Impact | Around 100 members of the general public attended a local farm, where I spoke about the use of technology to improve productivity, health and welfare on farms. |
| Year(s) Of Engagement Activity | 2018 |