Early Detection of Infection in Pigs through an Online Monitoring System

Abstract

Late detection of emergency diseases causes significant economic losses for pig producers and governments. As the first signs of animal infection are usually fever and reduced motion that lead to reduced consumption of water and feed, we developed a novel smart system to monitor body temperature and motion in real time, facilitating the early detection of infectious diseases. In this study, carried out within the framework of the European Union research project Rapidia Field, we tested the smart system on 10 pigs experimentally infected with two doses of an attenuated strain of African swine fever. Biosensors and an accelerometer embedded in an eartag captured data before and after infection, and video cameras were used to monitor the animals 24 h per day. The results showed that in 8 of 9 cases, the monitoring system detected infection onset as an increase in body temperature and decrease in movement before or simultaneously with fever detection based on rectal temperature measurement, observation of clinical signs, the decrease in water consumption or positive qPCR detection of virus. In addition, this decrease in movement was reliably detected using automatic analysis of video images therefore providing an inexpensive alternative to direct motion measurement. The system can be set up to alert staff when high fever, reduced motion or both are detected in one or more animals. This system may be useful for monitoring sentinel herds in real time, considerably reducing the financial and logistical costs of periodic sampling and increasing the chances of early detection of infection.

Early detection of health and welfare compromises through automated detection of behavioural changes in pigs

“…Accelerometers attached to pig ears combined with body temperature sensors automatically detected an infection 1–3 days prior to sampling techniques (Martínez-Avilés et al., 2015). …”

“…eYeNamic measures general activity from movement and can be interpreted for assessing aggression (Costa et al., 2007), responses to dust concentration (Costa et al., 2009), enrichments (Ismayilova et al., 2013), and climatic variation (Costa et al., 2014). A similar approach using commercially available, general-purpose motion detection software measured movement in hot spots of a pen to automatically detect behavioural changes (Martínez-Avilés et al., 2015). …”

“…In three cases, the data analysis methods were capable of automatically detecting behavioural changes in drinking behaviour from water flow sensors before diarrhoea (Madsen and Kristensen, 2005), in feeding visits and consumption with RFID feeding stations before tail biting (Wallenbeck and Keeling, 2013), and movement activity from video before clinical signs of swine fever (Martínez-Avilés et al., 2015). Monitoring behaviour measures and reports behaviour, but these three cases use more advanced approaches of applying data analysis to automatically detect behavioural changes.…”

Technologies for capturing and analysing animal health data in near real time

“…For example, infrared thermography has been demonstrated to detect FMD in cattle and mule deer based on increased foot temperatures that signalled the disease before the onset of clinical signs (44,45). Biosensors within ear tags in pigs were shown to detect African swine fever based on increased body temperature and decreased movement before or as clinical signs were first observed (46). Uses of radio-frequency identification (RFID) have extended beyond animal traceability and can now be injected/ ingested and connected via wireless networks and mobile devices for measuring animal biometrics, location and behaviour, including contact with other animals (47,48).…”

“…Still, the current limitations of wireless sensor technologies, such as the costs of their use per individual animal and the variable success of their use due to environmental conditions, has limited their broader application in animal health (50). For the early detection of disease events, such technologies may still have value for real-time data collection and monitoring within targeted disease surveillance strategies, such as the use of sentinel herds for early disease detection in an at-risk geographical area (46).…”