Signal capturing in participatory digital disease surveillance and its application for swine productivity improvement

Abstract

Nowadays, disease outbreaks in both humans and animals may have severe consequences for livestock industry and the world as a whole, making an efficient disease surveillance system necessary. Syndromic surveillance is used to detect abnormal signs and signals before an outbreak occurs. The capture of abnormal signals has been made easier with the help of digital technology, which involves a data input system (for information reporter and storage), an information transfer system, a processing system, and a notification system. For instance, the Participatory One Health Digital Disease Detection system, or PODD system, enables efficient and rapid detection of disease outbreaks and the control of their spread. The PODD system seeks cooperation from the public in submitting records of abnormalities occurring in people, animals, and the environment via a mobile phone application. The system detects clusters of illness symptoms. Between the years 2015 and 2016, 1,011 incidents of abnormalities were reported by users of the application, and 340 of those reports were continuous. These reports, which met the definition of an outbreak, were then sent to the government with the goal of developing enhanced control and prevention strategies for disease outbreaks. The objective of this thesis is to apply the concept of the participatory digital disease surveillance system, which helps to monitor abnormal signals before the spread of disease. In addition, digital technology has been developed to increase signal capture and improve productivity in pig farms. The technology used in this thesis studies includes use of artificial sucking sounds, development of sensors to monitor the body temperature of sow herds, classifying pig coughing sounds using machine learning for the detection of respiratory problems among fattening pigs. In the farrowing house, technology was applied to monitor and increase productivity on a pig farm. A total of 30 sows were divided into two groups: a treatment group (15 sows) that was exposed to artificial sucking sounds and a control group (15 sows) that was not. Both groups were exposed to the same management practices, and the housing areas were separated by a distance of about 270 meters. The objectives of this study were to compare farrowing indexes and to observe the sucking behavior of piglets using CCTV cameras. Moreover, Fecal samples of sows were collected daily to assess glucocorticoid hormone levels. The results indicate that a piglet approached a sow’s udder a significantly higher average number of times, while the sows were associated with a shorter onset time for the first piglet to come to the sow’s udder than for the control group (both p<0.05). The patterns and levels of FGM between the two groups were not different (both p>0.05), however; the treatment group produced better farrowing indexes than the control group, particularly in terms of litter weight gain and percent preweaning mortality rates. Moreover, the weaning to the first service interval of the treatment group was shorter than for the control group (p<0.05). Additionally, the stress hormone (glucocorticoid hormone, FGM) was activated in a sow during this process, indicating a state of stress. This would indicate that the artificial sucking sound could improve production index results and probably would have no adverse effect on the stress conditions of post-parturition sows. The body temperature of animals would need to be monitored for the prevention or progression of any disease within the herd. To replace a conventional rectal thermometer, an infrared sensor was used to detect the body temperature of sows. In this study, the automatic transfer of data was utilized, wherein data would flow from the infrared sensor to the point of data collection (cloud database). The data then undergo analysis through the web base, and responses would then be delivered (through notifications) to stakeholders. The body temperatures of 100 gestating sows were measured with the use of a standard thermometer (inserted into the rectum), while our device could be used on each part of the body of the sows. The results indicated that the valva or anus was an area that revealed a high correlation between the two measurements (R=0.78). In addition, this tool was employed for a full year in 2019 on commercial pig farms that were home to at least 300 sows. Production indexes were compared between 2018 (standard thermometer) and 2019 (infrared sensor). The results indicated that the majority of the production indexes in the after period (2019) were observed to be better than those of the before period (2018), especially in terms of the herd health status of the animals, productivity of fallowing house, and the welfare of those animals. Furthermore, this surveillance system could detect abnormalities as important signs of outbreaks in sow herds (via body temperature). Respiratory illness is an important problem that is commonly found on pig farms. Accordingly, coughing is an important sign that must be monitored. Many diseases are relevant as the cause of animal respiratory problems such as Mycoplasma hyopneumoniae (M.hyo), Streptococcus spp., Pasturella spp., Porcine reproductive and respiratory syndrome; PRRSv, and Porcine circovirus; PCV2. Machine learning can be used to classify pig coughing sounds on pig farms. Python was selected to convert the sound files to images (wave plots, spectrograms, and log power spectrograms). A recorder was used with a total of 45 healthy three-bred weaned piglets, wherein three replications of each were used with 15 weaning pigs per pen during different months. Blood samples and tonsil swabs were collected every month, but sounds were collected every week. Pig cough sounds were then distinguished from other sounds and a coughing index (CI) was established. Blood samples and tonsil swabs were utilized to determine respiratory diseases via laboratory tests that included ELISA, PCR, and bacterial cultures. According to our results, pig coughs sound were distinctly different from other sounds, as had been classified by python. Moreover, the laboratory results of the seroprofile of M.hyo, PRRSv, and PCV2, as had been established by the ELISA test, were employed in disease detection procedures during the fattening period. Spearman rank correlations and Kappa analysis were used to establish correlation values between coughing and the results of the laboratory tests. CI revealed a high correlation coefficient and agreement with the ELISA results of M.hyo, as well as the PCR results of PRRSv and PCV2 (p<0.05), while CI revealed a low correlation coefficient and agreement with the results of the Streptococcus spp. and Pasteurella spp. cultures (p>0.05). Therefore, the monitoring of coughing can be suited to detect respiratory problems and any potential relationships they may have with M.hyo, PRRSv, and PCV2 infections. In summary, the PODD system employs a syndromic surveillance approach that utilizes digital technology. The report was compiled with the help of reporters who were either villagers or workers of local government agencies. My study utilized digital technology to increase farm productivity by creating artificial sucking sounds. The PODD system could be applied to monitor endemic diseases and prevent outbreaks on pig farms by detecting early signs of abnormality, such as changes in body temperature and pig coughing. An automatic sensing system was employed which included infrared sensors and machine learning algorithms in the sound recorder. Those features make the surveillance system more user-friendly and accurate, especially in terms of data analysis and the potential to decrease the occurrence of human errors. After infrared sensors and the machine learning of the sound recorder were used in these studies, it was determined such system could be used for the early detection of certain abnormal signs such as for the detection of elevated body temperatures and the detection of coughing symptoms among pigs, respectively. The studies system is particularly effective for monitoring the body temperature of sows, as it can capture disease signals early, allowing rapid response to and prevent the spread of diseases that may impact productivity.