Pub Date : 2024-10-23DOI: 10.1016/j.prevetmed.2024.106363
Saulius Petkevičius , Patricija Klibavičė , Algirdas Šalomskas , Tomas Kupčinskas , Agata Moroz-Fik , Kinga Biernacka , Marcin Mickiewicz , Zofia Nowek , László Ózsvári , Krisztina Bárdos , Snorre Stuen , Carlos Eduardo Abril , Giuseppe Bertoni , Jarosław Kaba , Michał Czopowicz
Caprine arthritis-encephalitis (CAE) is a progressive disease of goats caused by small ruminant lentivirus (SRLV) and is considered as one of the most important threats for goat farming in developed countries. The disease prevalence has never been investigated in the Lithuanian goat population. Therefore, a descriptive cross-sectional study was carried out in 2021–2022 to determine if SRLV infection was present in the Lithuanian goat population and, in the case of a positive result, to estimate the true herd-level prevalence of SRLV infection and specify genotypes and subtypes of SRLV responsible for the infection. Thirty goat herds counting >5 adult goats were randomly selected and, in each herd, a representative sample of adult goats was blood-sampled and tested serologically for SRLV infection using a commercial ELISA. The herd was considered infected if at least one goat tested positive and the true herd-level prevalence of SRLV infection was estimated using the Bayesian approach. Seropositive animals were found in 17 / 30 herds (57 %; 95 % confidence interval: 39 %, 73 %). The true herd-level prevalence was 56 % (95 % credible interval: 36 %, 76 %). In 10 / 17 seropositive herds whose owners consented for resampling of seropositive goats, 1–5 seropositive goats were tested using the nested real-time PCR (nRT-PCR). Goats from 9 seropositive herds tested positive in the nRT-PCR: in 4 herds for genotype A, in 4 herds for genotype B, and in 1 herd – 2 goats for genotype B and 1 goat for genotype A. From each of 9 nRT-PCR-positive herds, 1 PCR product of each genotype was sequenced using Sanger method and the phylogenic tree was constructed using the neighbor-joining method in the Molecular Evolutionary Genetics Analysis software. Four herds turned out to be infected with B1 subtype (91 % identity with the prototypic strain), 3 herds with A2 subtype (90 %-92 % identity), and a herd with mixed infection was infected with B1 (91 % identity) and A2 subtype (90 % identity). In one herd, the only seropositive goat was found to be infected with the strain most closely related to the A1 subtype (80 % identity). This study shows for the first time that SRLV infection is present and widespread in the Lithuanian goat population and both classical SRLV genotypes, represented by quite typical subtypes A2 and B1, appear to be responsible for the infection.
{"title":"The herd-level prevalence of caprine arthritis-encephalitis and genetic characteristics of small ruminant lentivirus in the Lithuanian goat population","authors":"Saulius Petkevičius , Patricija Klibavičė , Algirdas Šalomskas , Tomas Kupčinskas , Agata Moroz-Fik , Kinga Biernacka , Marcin Mickiewicz , Zofia Nowek , László Ózsvári , Krisztina Bárdos , Snorre Stuen , Carlos Eduardo Abril , Giuseppe Bertoni , Jarosław Kaba , Michał Czopowicz","doi":"10.1016/j.prevetmed.2024.106363","DOIUrl":"10.1016/j.prevetmed.2024.106363","url":null,"abstract":"<div><div>Caprine arthritis-encephalitis (CAE) is a progressive disease of goats caused by small ruminant lentivirus (SRLV) and is considered as one of the most important threats for goat farming in developed countries. The disease prevalence has never been investigated in the Lithuanian goat population. Therefore, a descriptive cross-sectional study was carried out in 2021–2022 to determine if SRLV infection was present in the Lithuanian goat population and, in the case of a positive result, to estimate the true herd-level prevalence of SRLV infection and specify genotypes and subtypes of SRLV responsible for the infection. Thirty goat herds counting >5 adult goats were randomly selected and, in each herd, a representative sample of adult goats was blood-sampled and tested serologically for SRLV infection using a commercial ELISA. The herd was considered infected if at least one goat tested positive and the true herd-level prevalence of SRLV infection was estimated using the Bayesian approach. Seropositive animals were found in 17 / 30 herds (57 %; 95 % confidence interval: 39 %, 73 %). The true herd-level prevalence was 56 % (95 % credible interval: 36 %, 76 %). In 10 / 17 seropositive herds whose owners consented for resampling of seropositive goats, 1–5 seropositive goats were tested using the nested real-time PCR (nRT-PCR). Goats from 9 seropositive herds tested positive in the nRT-PCR: in 4 herds for genotype A, in 4 herds for genotype B, and in 1 herd – 2 goats for genotype B and 1 goat for genotype A. From each of 9 nRT-PCR-positive herds, 1 PCR product of each genotype was sequenced using Sanger method and the phylogenic tree was constructed using the neighbor-joining method in the Molecular Evolutionary Genetics Analysis software. Four herds turned out to be infected with B1 subtype (91 % identity with the prototypic strain), 3 herds with A2 subtype (90 %-92 % identity), and a herd with mixed infection was infected with B1 (91 % identity) and A2 subtype (90 % identity). In one herd, the only seropositive goat was found to be infected with the strain most closely related to the A1 subtype (80 % identity). This study shows for the first time that SRLV infection is present and widespread in the Lithuanian goat population and both classical SRLV genotypes, represented by quite typical subtypes A2 and B1, appear to be responsible for the infection.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106363"},"PeriodicalIF":2.2,"publicationDate":"2024-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142560795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sulfachloropyridazine (SCP) is an antimicrobial (AM) commonly used in the poultry industry. This drug is excreted as the original compound, which may accumulate in litter. This work was done to assess whether SCP residues from droppings of broiler chickens that were treated with therapeutic doses of this drug spread into the production environment and to determine if these events were associated with the selection of resistant bacteria. To this end, broiler chickens were raised under controlled conditions, and their droppings and litter were processed to detect and identify SCP residues using an HPLC-MS/MS technique. This study selected Escherichia coli as an indicator bacterium for AM resistance. Its phenotypic resistance was determined using the Kirby-Bauer disk diffusion method, and its genotypic resistance was determined by performing a conventional PCR test. Our results showed that SCP residues did spread from the treated group to untreated sentinel groups because SCP residues in the litter reached levels up to 43.05 µg·kg−1 in a group placed immediately adjoining to the treated group, while another group placed 30 cm away showed a concentration of 29.79 µg·kg−1. Meanwhile, only trace concentrations were detected in droppings collected from sentinel groups. Of 239 strains of E. coli isolated from droppings, 12.13 % were resistant to sulfonamides, whereas 23.91 % of 92 E. coli isolated from broiler litter were resistant. The most prevalent resistance gene was the sul2 gene, both in droppings and litter, followed by the sul1 gene. The SCP concentrations were associated with the probability of E. coli being resistant to sulfonamides (p-value = 0.01). A Multiple Correspondence Analysis (MCA) also showed that phenotypic and genotypic resistances were associated and that both genes sul1 and sul2 would determine phenotypic resistance to sulfonamides in E. coli. The results presented in this study show that inedible by-products of the poultry industry are potentially a source of drug resistance that can spread from the animal production line to the environment, so awareness of the correct use of antimicrobials is essential to combat antimicrobial resistance.
{"title":"Assessing the spread of sulfachloropyridazine in poultry environment and its impact on Escherichia coli resistance","authors":"María Belén Vargas , Ekaterina Pokrant , Isidora García , Rocío Cadena , Francisco Mena , Karina Yévenes , Catalina Fuentes , Sebastián Zavala , Andrés Flores , Matías Maturana , Aldo Maddaleno , Héctor Hidalgo , Lisette Lapierre , Javiera Cornejo","doi":"10.1016/j.prevetmed.2024.106362","DOIUrl":"10.1016/j.prevetmed.2024.106362","url":null,"abstract":"<div><div>Sulfachloropyridazine (SCP) is an antimicrobial (AM) commonly used in the poultry industry. This drug is excreted as the original compound, which may accumulate in litter. This work was done to assess whether SCP residues from droppings of broiler chickens that were treated with therapeutic doses of this drug spread into the production environment and to determine if these events were associated with the selection of resistant bacteria. To this end, broiler chickens were raised under controlled conditions, and their droppings and litter were processed to detect and identify SCP residues using an HPLC-MS/MS technique. This study selected <em>Escherichia coli</em> as an indicator bacterium for AM resistance. Its phenotypic resistance was determined using the Kirby-Bauer disk diffusion method, and its genotypic resistance was determined by performing a conventional PCR test. Our results showed that SCP residues did spread from the treated group to untreated sentinel groups because SCP residues in the litter reached levels up to 43.05 µg·kg<sup>−1</sup> in a group placed immediately adjoining to the treated group, while another group placed 30 cm away showed a concentration of 29.79 µg·kg<sup>−1</sup>. Meanwhile, only trace concentrations were detected in droppings collected from sentinel groups. Of 239 strains of <em>E. coli</em> isolated from droppings, 12.13 % were resistant to sulfonamides, whereas 23.91 % of 92 <em>E. coli</em> isolated from broiler litter were resistant. The most prevalent resistance gene was the <em>sul</em>2 gene, both in droppings and litter, followed by the <em>sul</em>1 gene. The SCP concentrations were associated with the probability of <em>E. coli</em> being resistant to sulfonamides (p-value = 0.01). A Multiple Correspondence Analysis (MCA) also showed that phenotypic and genotypic resistances were associated and that both genes <em>sul</em>1 and <em>sul</em>2 would determine phenotypic resistance to sulfonamides in <em>E. coli.</em> The results presented in this study show that inedible by-products of the poultry industry are potentially a source of drug resistance that can spread from the animal production line to the environment, so awareness of the correct use of antimicrobials is essential to combat antimicrobial resistance.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106362"},"PeriodicalIF":2.2,"publicationDate":"2024-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142530198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-19DOI: 10.1016/j.prevetmed.2024.106361
Natalie Anne Steen , Karianne Muri , Magnhild Oust Torske
The One Welfare approach acknowledges the interrelationships between human wellbeing and animal welfare. Early research has suggested associations between stockperson wellbeing and livestock welfare, however these scenarios are complex and challenging to untangle.
In this study, we utilised merged data from over 700 farms to explore associations between farmer wellbeing and livestock welfare. The farms were engaged in cattle, sheep, and/or swine production in Norway between 2017 and 2020. The farmers participated in a general population-based health survey, and livestock welfare was measured using routinely collected, animal-based abattoir observations of over 480,000 animals. We determined a farm’s overall livestock welfare relative to the other farms and calculated within-farm differences in this relative welfare level over time. A subset of enterprises (n=328) with sufficient and non-ambiguous farmer wellbeing information were then used to explore differences in these within-farm differences by farmer wellbeing status. We found that poor farmer wellbeing — whether it was defined by anxiety symptoms, depression symptoms, symptoms of psychological distress, or life satisfaction — was associated with a deterioration in overall livestock welfare level (in terms of the mean of the farm’s abattoir observed welfare indicators). There was evidence that this association persisted for at least two years.
Given societal concerns regarding sustainable food production, farmer wellbeing, and livestock welfare, further research is indicated to explore the complex farmer-livestock relationship within the One Welfare framework. This study suggests that using within-farm changes in relative livestock welfare derived from routinely collected information can be a useful approach.
{"title":"Exploring longitudinal associations between farmer wellbeing and the welfare of their livestock. The HUNT Study, Norway","authors":"Natalie Anne Steen , Karianne Muri , Magnhild Oust Torske","doi":"10.1016/j.prevetmed.2024.106361","DOIUrl":"10.1016/j.prevetmed.2024.106361","url":null,"abstract":"<div><div>The One Welfare approach acknowledges the interrelationships between human wellbeing and animal welfare. Early research has suggested associations between stockperson wellbeing and livestock welfare, however these scenarios are complex and challenging to untangle.</div><div>In this study, we utilised merged data from over 700 farms to explore associations between farmer wellbeing and livestock welfare. The farms were engaged in cattle, sheep, and/or swine production in Norway between 2017 and 2020. The farmers participated in a general population-based health survey, and livestock welfare was measured using routinely collected, animal-based abattoir observations of over 480,000 animals. We determined a farm’s overall livestock welfare relative to the other farms and calculated within-farm differences in this relative welfare level over time. A subset of enterprises (n=328) with sufficient and non-ambiguous farmer wellbeing information were then used to explore differences in these within-farm differences by farmer wellbeing status. We found that poor farmer wellbeing — whether it was defined by anxiety symptoms, depression symptoms, symptoms of psychological distress, or life satisfaction — was associated with a deterioration in overall livestock welfare level (in terms of the mean of the farm’s abattoir observed welfare indicators). There was evidence that this association persisted for at least two years.</div><div>Given societal concerns regarding sustainable food production, farmer wellbeing, and livestock welfare, further research is indicated to explore the complex farmer-livestock relationship within the One Welfare framework. This study suggests that using within-farm changes in relative livestock welfare derived from routinely collected information can be a useful approach.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106361"},"PeriodicalIF":2.2,"publicationDate":"2024-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142506520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Antimicrobial resistance (AMR) is a growing challenge for the successful treatment of bacterial infections in both human and veterinary medicine. Despite the need to mitigate AMR, food-producing animal species lack adequate information on bacterial susceptibility to support antimicrobial stewardship for conditions that drive antimicrobial usage, such as ovine respiratory disease (ORD). In an effort to help address this gap, the upper respiratory tract (URT) from 620 apparently healthy sheep carcasses was sampled at slaughter in California (CA), from April through September 2021, to determine the antimicrobial susceptibility profiles of commensal Mannheimia haemolytica and Pasteurella multocida, two bacteria commonly associated with ORD. Sheep sampled in the study were selected based on origin prior to processing (out-of-state or CA) and marketing status (antibiotic-free or conventional management), two factors hypothesized to influence antimicrobial susceptibility profiles. Of the total 620 carcasses sampled, 343 had at least one isolate of M. haemolytica or P. multocida recovered, for a recovery rate of 55.3 %. The recovery rate among sampled carcasses was 46.8 % (290/620) for M. haemolytica and 15.8 % (98/620) for P. multocida. Utilizing Clinical and Laboratory Standards Institute (CLSI) clinical breakpoints as interpretive criteria, all 98 P. multocida isolates were pansusceptible to the antimicrobials tested that are labeled for use in sheep, with the exception of one isolate that classified as intermediate to tilmicosin. Of the 290 M. haemolytica isolates, the greatest resistance was found to penicillin, with 51.0 % (148) of isolates classified as intermediate and 25.2 % (73) resistant, while nine isolates were resistant to chlortetracycline and oxytetracycline and one isolate was intermediate to spectinomycin. Multidrug resistance (resistance to three or more antimicrobial classes) was not found in any isolate of either target organism. Very low to no AMR was observed across both bacterial species tested, demonstrating that M. haemolytica and P. multocida in this apparently healthy sheep population remain broadly susceptible to the antimicrobials licensed for the treatment of ORD. In addition, antimicrobial susceptibility did not differ based on origin or marketing status, as defined for this study. Sampling at slaughter provides an opportunity to collect baseline information about antimicrobial susceptibility of the commensal flora of the sheep URT that can give rise to ORD, but may not be generalizable to sheep with clinical respiratory disease or to sheep reared by non-commercial producers.
{"title":"Antimicrobial susceptibility profiles among commensal Mannheimia haemolytica and Pasteurella multocida isolated from apparently healthy sheep processed in California: Results from a cross-sectional pilot study","authors":"Wendi Jackson , Jenna Tucker , Heather Fritz , Craig Bross , Jaymes Adams , Marissa Silva , Catherine Lorenz , Edith Marshall","doi":"10.1016/j.prevetmed.2024.106360","DOIUrl":"10.1016/j.prevetmed.2024.106360","url":null,"abstract":"<div><div>Antimicrobial resistance (AMR) is a growing challenge for the successful treatment of bacterial infections in both human and veterinary medicine. Despite the need to mitigate AMR, food-producing animal species lack adequate information on bacterial susceptibility to support antimicrobial stewardship for conditions that drive antimicrobial usage, such as ovine respiratory disease (ORD). In an effort to help address this gap, the upper respiratory tract (URT) from 620 apparently healthy sheep carcasses was sampled at slaughter in California (CA), from April through September 2021, to determine the antimicrobial susceptibility profiles of commensal <em>Mannheimia haemolytica</em> and <em>Pasteurella multocida,</em> two bacteria commonly associated with ORD. Sheep sampled in the study were selected based on origin prior to processing (out-of-state or CA) and marketing status (antibiotic-free or conventional management), two factors hypothesized to influence antimicrobial susceptibility profiles. Of the total 620 carcasses sampled, 343 had at least one isolate of <em>M. haemolytica or P. multocida</em> recovered, for a recovery rate of 55.3 %. The recovery rate among sampled carcasses was 46.8 % (290/620) for <em>M. haemolytica</em> and 15.8 % (98/620) for <em>P. multocida.</em> Utilizing Clinical and Laboratory Standards Institute (CLSI) clinical breakpoints as interpretive criteria, all 98 <em>P. multocida</em> isolates were pansusceptible to the antimicrobials tested that are labeled for use in sheep, with the exception of one isolate that classified as intermediate to tilmicosin. Of the 290 <em>M. haemolytica</em> isolates, the greatest resistance was found to penicillin, with 51.0 % (148) of isolates classified as intermediate and 25.2 % (73) resistant<em>,</em> while nine isolates were resistant to chlortetracycline and oxytetracycline and one isolate was intermediate to spectinomycin. Multidrug resistance (resistance to three or more antimicrobial classes) was not found in any isolate of either target organism. Very low to no AMR was observed across both bacterial species tested, demonstrating that <em>M. haemolytica</em> and <em>P. multocida</em> in this apparently healthy sheep population remain broadly susceptible to the antimicrobials licensed for the treatment of ORD. In addition, antimicrobial susceptibility did not differ based on origin or marketing status, as defined for this study. Sampling at slaughter provides an opportunity to collect baseline information about antimicrobial susceptibility of the commensal flora of the sheep URT that can give rise to ORD, but may not be generalizable to sheep with clinical respiratory disease or to sheep reared by non-commercial producers.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106360"},"PeriodicalIF":2.2,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142506518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-17DOI: 10.1016/j.prevetmed.2024.106358
Tasneem Imam , Sara Horsman , Ben Wood , John D. Grewar , Charlotte Langhorne , Rochelle Price , Caitlin Wood , Joerg Henning , Justine S. Gibson
<div><div>Mastitis, a prevalent and economically important disease in the dairy industry, poses substantial challenges to dairy cow health, milk quality, and farm profitability worldwide. Mastitis is predominantly caused by bacterial infections. The objective of this study was to estimate the sensitivity (Se) and specificity (Sp) of bacterial culture and the VetMAX™ MastiType Multi Kit PCR in identified clinical mastitis pathogens. A total of 396 quarter-level milk samples were collected from 396 cows with clinical mastitis on 29 farms in the subtropical dairy region of Australia between March and December 2021. These samples were cultured and tested by PCR, and analysed using Bayesian latent class analysis under the assumption of <em>one population two tests</em> and also of <em>three populations two tests,</em> by dividing the population into subpopulations based on regions. Informative priors used in the analysis were calculated from published evidence. Models were compared using the Deviance Information Criterion (DIC). Sensitivity analysis was performed to evaluate the impact of changes in priors. The most common isolates cultured and detected by PCR were <em>Streptococcus uberis</em> (17.4 % and 27.3 %, respectively) and <em>Escherichia coli</em> (12.6 % and 25.0 %, respectively). Under the assumption of <em>one population two tests</em>, the Se of PCR (at cycle threshold (Ct) ≤ 37) was higher than that of bacterial culture for both pathogens: for <em>E. coli</em>, the Se was 50.2 % (95 % posterior probability interval (PPI): 37.4; 74.1) for bacterial culture, and 93.7 % (95 % PPI: 85.5; 98.4) for PCR. For <em>S. uberis</em>, the Se was 50.4 % (95 % PPI: 40.9; 61.3) for bacterial culture, and 81.5 % (73.0; 88.9) for PCR. Conversely, the Sp of bacterial culture was higher than that of PCR for both pathogens: for <em>E. coli</em>, the Sp was 99.2 % (97.8; 100) for bacterial culture, and 95.1 % (87.8; 99.4) for PCR. For <em>S. uberis</em>, the Sp was 99.2 % (95 % PPI: 97.6; 100) for bacterial culture, and 96.7 % (95 % PPI: 92.1; 99.2) for PCR. Bayesian latent class analysis with <em>three populations two tests</em> was only performed for <em>S. uberis</em>. For <em>E. coli,</em> this could not be performed because there were no PCR-positive results in one subpopulation. Under the assumption of <em>three populations two tests</em>, for <em>S. uberis,</em> the Se was 49.6 % (40.6; 59.4) for bacterial culture, and 81.1 % (72.6; 88.6) for PCR; and the Sp for bacterial culture was 99.1 % (97.7; 100), and for PCR was 96.9 % (93.0; 99.3). The DIC for the <em>one population two tests</em> model was lower than the DIC for the <em>three populations two tests</em> model<em>.</em> The sensitivity analysis for the <em>one population two tests</em> model demonstrated that a 10 % reduction in priors led to substantial changes in Se of both bacterial culture and PCR tests for <em>E. coli</em> and <em>S. uberis</em>, with overlap percentages ranging from 80.6
{"title":"Assessment of sensitivity and specificity of bacterial culture and the VetMAX™ MastiType Multi Kit in detecting Streptococcus uberis and Escherichia coli in milk samples from dairy cows with clinical mastitis in subtropical Australia","authors":"Tasneem Imam , Sara Horsman , Ben Wood , John D. Grewar , Charlotte Langhorne , Rochelle Price , Caitlin Wood , Joerg Henning , Justine S. Gibson","doi":"10.1016/j.prevetmed.2024.106358","DOIUrl":"10.1016/j.prevetmed.2024.106358","url":null,"abstract":"<div><div>Mastitis, a prevalent and economically important disease in the dairy industry, poses substantial challenges to dairy cow health, milk quality, and farm profitability worldwide. Mastitis is predominantly caused by bacterial infections. The objective of this study was to estimate the sensitivity (Se) and specificity (Sp) of bacterial culture and the VetMAX™ MastiType Multi Kit PCR in identified clinical mastitis pathogens. A total of 396 quarter-level milk samples were collected from 396 cows with clinical mastitis on 29 farms in the subtropical dairy region of Australia between March and December 2021. These samples were cultured and tested by PCR, and analysed using Bayesian latent class analysis under the assumption of <em>one population two tests</em> and also of <em>three populations two tests,</em> by dividing the population into subpopulations based on regions. Informative priors used in the analysis were calculated from published evidence. Models were compared using the Deviance Information Criterion (DIC). Sensitivity analysis was performed to evaluate the impact of changes in priors. The most common isolates cultured and detected by PCR were <em>Streptococcus uberis</em> (17.4 % and 27.3 %, respectively) and <em>Escherichia coli</em> (12.6 % and 25.0 %, respectively). Under the assumption of <em>one population two tests</em>, the Se of PCR (at cycle threshold (Ct) ≤ 37) was higher than that of bacterial culture for both pathogens: for <em>E. coli</em>, the Se was 50.2 % (95 % posterior probability interval (PPI): 37.4; 74.1) for bacterial culture, and 93.7 % (95 % PPI: 85.5; 98.4) for PCR. For <em>S. uberis</em>, the Se was 50.4 % (95 % PPI: 40.9; 61.3) for bacterial culture, and 81.5 % (73.0; 88.9) for PCR. Conversely, the Sp of bacterial culture was higher than that of PCR for both pathogens: for <em>E. coli</em>, the Sp was 99.2 % (97.8; 100) for bacterial culture, and 95.1 % (87.8; 99.4) for PCR. For <em>S. uberis</em>, the Sp was 99.2 % (95 % PPI: 97.6; 100) for bacterial culture, and 96.7 % (95 % PPI: 92.1; 99.2) for PCR. Bayesian latent class analysis with <em>three populations two tests</em> was only performed for <em>S. uberis</em>. For <em>E. coli,</em> this could not be performed because there were no PCR-positive results in one subpopulation. Under the assumption of <em>three populations two tests</em>, for <em>S. uberis,</em> the Se was 49.6 % (40.6; 59.4) for bacterial culture, and 81.1 % (72.6; 88.6) for PCR; and the Sp for bacterial culture was 99.1 % (97.7; 100), and for PCR was 96.9 % (93.0; 99.3). The DIC for the <em>one population two tests</em> model was lower than the DIC for the <em>three populations two tests</em> model<em>.</em> The sensitivity analysis for the <em>one population two tests</em> model demonstrated that a 10 % reduction in priors led to substantial changes in Se of both bacterial culture and PCR tests for <em>E. coli</em> and <em>S. uberis</em>, with overlap percentages ranging from 80.6 ","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106358"},"PeriodicalIF":2.2,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142506519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-17DOI: 10.1016/j.prevetmed.2024.106356
Janice Y. Park , Magnus R. Campler , Ting-Yu Cheng , Justin D. Kieffer , Andréia G. Arruda , Madonna E. Benjamin , Dale W. Rozeboom , Andrew S. Bowman
Emergency contingency plans for the U.S. swine industry, including depopulation protocols, that are essential to limit the spread of sudden outbreaks of infectious diseases is currently lacking. Although novel depopulation methods such as water-based foam (WBF) and high-expansion nitrogen foam (N2F) are being investigated, carbon dioxide gas (CO2) is currently the only American Veterinary Medical Association (AVMA)-preferred method capable of depopulation of groups of swine. The AVMA’s assessment of depopulation methods evaluates efficacy, animal welfare implications, and caretaker physical and mental health, in addition to logistical aspects of equipment acquisition and use. These criteria are best gauged using input from individuals familiar with the operations within the swine industry. Therefore, the aim of this study was to describe swine industry stakeholders’ perceptions of WBF, N2F and CO2 depopulation after a large-scale field demonstration. A survey was created based on the criteria outlined in the AVMA Depopulation Guidelines to determine respondent perceptions of each method. Swine industry stakeholders of various backgrounds (N=32) were recruited and invited to observe demonstrations of each method. Mixed linear regression models were built to investigate the association between survey question scores and depopulation method. Respondents varied in occupation, with 37.5 % (12/32) belonging to an academic institution or veterinary medical association, 21.9 % (7/32) to a pork organization, and 18.8 % (6/32) to state or federal regulatory agencies. The remaining 21.8 % (7/32) was a group consisting of one producer (3.1 %), one individual in swine industry retail (3.1 %), one veterinarian in a private practice setting (3.1 %) and four (12.5 %) who did not disclose their affiliation. Average experience (±SD) in the swine industry was 14.4 (±12.4) years, and 40.6 % (13/32) had previous experience of any type in swine depopulation. The overall method impression scores revealed that WBF and N2F were perceived as better options compared to CO2 (P < 0.001). WBF and N2F scored higher on pig distress mitigation, protecting emotional and psychological health of personnel, and equipment accessibility compared to CO2 (P < 0.001). Stakeholders without a clear affiliation provided higher scores compared those affiliated with regulatory agencies, pork associations or academic or veterinary institutions, regarding minimizing pig distress, method safety/accessibility, and overall method impression. Few demographic differences were observed, suggesting similar perceptions of the three depopulation methods during the demonstrations. This industry feedback is valuable for future considerations, method improvements and facilitation for possible implementations into future response plans.
{"title":"Perceptions of swine industry stakeholders on the use of water-based foam, high-expansion nitrogen foam, and carbon dioxide gas as methods of swine depopulation","authors":"Janice Y. Park , Magnus R. Campler , Ting-Yu Cheng , Justin D. Kieffer , Andréia G. Arruda , Madonna E. Benjamin , Dale W. Rozeboom , Andrew S. Bowman","doi":"10.1016/j.prevetmed.2024.106356","DOIUrl":"10.1016/j.prevetmed.2024.106356","url":null,"abstract":"<div><div>Emergency contingency plans for the U.S. swine industry, including depopulation protocols, that are essential to limit the spread of sudden outbreaks of infectious diseases is currently lacking. Although novel depopulation methods such as water-based foam (WBF) and high-expansion nitrogen foam (N<sub>2</sub>F) are being investigated, carbon dioxide gas (CO<sub>2</sub>) is currently the only American Veterinary Medical Association (AVMA)-preferred method capable of depopulation of groups of swine. The AVMA’s assessment of depopulation methods evaluates efficacy, animal welfare implications, and caretaker physical and mental health, in addition to logistical aspects of equipment acquisition and use. These criteria are best gauged using input from individuals familiar with the operations within the swine industry. Therefore, the aim of this study was to describe swine industry stakeholders’ perceptions of WBF, N<sub>2</sub>F and CO<sub>2</sub> depopulation after a large-scale field demonstration. A survey was created based on the criteria outlined in the AVMA Depopulation Guidelines to determine respondent perceptions of each method. Swine industry stakeholders of various backgrounds (N=32) were recruited and invited to observe demonstrations of each method. Mixed linear regression models were built to investigate the association between survey question scores and depopulation method. Respondents varied in occupation, with 37.5 % (12/32) belonging to an academic institution or veterinary medical association, 21.9 % (7/32) to a pork organization, and 18.8 % (6/32) to state or federal regulatory agencies. The remaining 21.8 % (7/32) was a group consisting of one producer (3.1 %), one individual in swine industry retail (3.1 %), one veterinarian in a private practice setting (3.1 %) and four (12.5 %) who did not disclose their affiliation. Average experience (±SD) in the swine industry was 14.4 (±12.4) years, and 40.6 % (13/32) had previous experience of any type in swine depopulation. The overall method impression scores revealed that WBF and N<sub>2</sub>F were perceived as better options compared to CO<sub>2</sub> (<em>P</em> < 0.001). WBF and N<sub>2</sub>F scored higher on pig distress mitigation, protecting emotional and psychological health of personnel, and equipment accessibility compared to CO<sub>2</sub> (<em>P</em> < 0.001). Stakeholders without a clear affiliation provided higher scores compared those affiliated with regulatory agencies, pork associations or academic or veterinary institutions, regarding minimizing pig distress, method safety/accessibility, and overall method impression. Few demographic differences were observed, suggesting similar perceptions of the three depopulation methods during the demonstrations. This industry feedback is valuable for future considerations, method improvements and facilitation for possible implementations into future response plans.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106356"},"PeriodicalIF":2.2,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142472984","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-16DOI: 10.1016/j.prevetmed.2024.106359
Thanicha Chanchaidechachai , Egil A.J. Fischer , Helmut W. Saatkamp , Mart C.M. de Jong , Henk Hogeveen
Information on the epidemiological and economic consequences of control measures is fundamental to design effective foot and mouth disease (FMD) control measures. One approach to obtaining this information is through bioeconomic modelling. In this study, a bioeconomic model was used to evaluate FMD control in two different study areas in Thailand: a high farm density area predominantly consisting of dairy farms and a low farm density area with mixed farm types. The bioeconomic model consists of an epidemiological part and an economic part. For the epidemiological part, a stochastic between-farm transmission model was constructed with transmission parameters estimated from FMD outbreaks in Thailand. The outputs from the epidemiological model, i.e. the number of infected farms, the number of affected farms and the outbreak duration, are used as inputs for economic model to calculate the economic consequences. We applied the simulation model with four FMD control measures: culling the animals of infected farms, ring vaccination, animal movement restrictions and isolation of infected farms. Furthermore, we included effect of farmers' compliance to asses its effect on control measures. The simulated FMD outbreaks in the low farm density area were small, thus control measures did not greatly affect the size of outbreaks and, therefore, did not have a positive economic return. In contrast, in the high farm density area, FMD outbreaks were large without control measures. All measures reduced the size of the outbreaks but resulted in different total costs. In terms of outbreak control, culling infectious farms was the best option, but its total cost was higher than ring vaccination or isolation of infected farms. In terms of cost-effectiveness, ring vaccination was the best measure. If farmers' compliance were low, all control measures would be ineffective, resulting in high total cost of the outbreak. The cost distribution between compliant and non-compliant farms showed that non-compliant farms paid more than compliant farms, except for the ring vaccination scenario. The results emphasize the economic significance to customize control measures specific to the area's conditions and highlight the importance of farmers' compliance when designing control measures.
{"title":"One-size measures do not fit all areas: Evaluation of area-specific control of foot and mouth disease in Thailand using bioeconomic modelling","authors":"Thanicha Chanchaidechachai , Egil A.J. Fischer , Helmut W. Saatkamp , Mart C.M. de Jong , Henk Hogeveen","doi":"10.1016/j.prevetmed.2024.106359","DOIUrl":"10.1016/j.prevetmed.2024.106359","url":null,"abstract":"<div><div>Information on the epidemiological and economic consequences of control measures is fundamental to design effective foot and mouth disease (FMD) control measures. One approach to obtaining this information is through bioeconomic modelling. In this study, a bioeconomic model was used to evaluate FMD control in two different study areas in Thailand: a high farm density area predominantly consisting of dairy farms and a low farm density area with mixed farm types. The bioeconomic model consists of an epidemiological part and an economic part. For the epidemiological part, a stochastic between-farm transmission model was constructed with transmission parameters estimated from FMD outbreaks in Thailand. The outputs from the epidemiological model, i.e. the number of infected farms, the number of affected farms and the outbreak duration, are used as inputs for economic model to calculate the economic consequences. We applied the simulation model with four FMD control measures: culling the animals of infected farms, ring vaccination, animal movement restrictions and isolation of infected farms. Furthermore, we included effect of farmers' compliance to asses its effect on control measures. The simulated FMD outbreaks in the low farm density area were small, thus control measures did not greatly affect the size of outbreaks and, therefore, did not have a positive economic return. In contrast, in the high farm density area, FMD outbreaks were large without control measures. All measures reduced the size of the outbreaks but resulted in different total costs. In terms of outbreak control, culling infectious farms was the best option, but its total cost was higher than ring vaccination or isolation of infected farms. In terms of cost-effectiveness, ring vaccination was the best measure. If farmers' compliance were low, all control measures would be ineffective, resulting in high total cost of the outbreak. The cost distribution between compliant and non-compliant farms showed that non-compliant farms paid more than compliant farms, except for the ring vaccination scenario. The results emphasize the economic significance to customize control measures specific to the area's conditions and highlight the importance of farmers' compliance when designing control measures.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106359"},"PeriodicalIF":2.2,"publicationDate":"2024-10-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142506521","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-15DOI: 10.1016/j.prevetmed.2024.106357
Matthew Thomas, Francesca Occhiuto, Martin Green, Jorge A Vázquez-Diosdado, Jasmeet Kaler
Subtle changes in behaviour, which may be useful as early indicators of disease or stress in farm animals, can be detected using precision livestock technologies. However, as animals often display differences in their baseline behaviours from one another, individual variation needs to be measured and accounted for if we want to successfully detect abnormalities. In addition to consistent individual differences in their mean behaviour, which are called behavioural types and measured by repeatability, animals may differ in the amount of day-to-day variation around their own mean. This is known as predictability and it is very rarely measured in animals as it requires data with high granularity for extended periods of time, which is hard to obtain without precision technologies. Here we aim to quantify the repeatability and predictability of lying and feeding, two behaviours which have been investigated as potential indicators of disease with inconsistent results, in dairy cows. We used data on daily lying behaviours from leg-mounted sensors for 1439 cows and data on daily feeding behaviours from neck-mounted sensors for 2584 cows belonging to multiple herds to quantify the variation between individuals, the individual differences in predictability and any correlations between individual behaviour type and predictability of the same behaviour. Using multivariate double hierarchical generalised linear models, we were able to quantify individual variation in feeding and lying behaviours of adult cows for the first time. Consistent variation between individuals was present for all measures of lying and feeding with repeatability values of 0.38 for daily feeding time and 0.34 for daily lying time. Individuals also varied in their levels of predictability with coefficients of 0.46 and 0.29 for the daily lying and feeding time respectively. Lastly, there were significant positive correlations between several behaviours and the amount of residual intra-individual variation for the same behaviour (mean lying bout length: 0.93, mean feeding bout length: 0.88), indicating that cows with the longest bouts of the two behaviours were also more variable and unpredictable, perhaps due to an opportunistic strategy. These results highlight the importance of measuring individual variation in behaviour, supporting the evidence of personality traits in cattle, as well as differences in predictability, which must be taken into account in any use of behavioural indicators for early disease diagnosis.
{"title":"Repeatability and predictability of lying and feeding behaviours in dairy cattle","authors":"Matthew Thomas, Francesca Occhiuto, Martin Green, Jorge A Vázquez-Diosdado, Jasmeet Kaler","doi":"10.1016/j.prevetmed.2024.106357","DOIUrl":"10.1016/j.prevetmed.2024.106357","url":null,"abstract":"<div><div>Subtle changes in behaviour, which may be useful as early indicators of disease or stress in farm animals, can be detected using precision livestock technologies. However, as animals often display differences in their baseline behaviours from one another, individual variation needs to be measured and accounted for if we want to successfully detect abnormalities. In addition to consistent individual differences in their mean behaviour, which are called behavioural types and measured by repeatability, animals may differ in the amount of day-to-day variation around their own mean. This is known as predictability and it is very rarely measured in animals as it requires data with high granularity for extended periods of time, which is hard to obtain without precision technologies. Here we aim to quantify the repeatability and predictability of lying and feeding, two behaviours which have been investigated as potential indicators of disease with inconsistent results, in dairy cows. We used data on daily lying behaviours from leg-mounted sensors for 1439 cows and data on daily feeding behaviours from neck-mounted sensors for 2584 cows belonging to multiple herds to quantify the variation between individuals, the individual differences in predictability and any correlations between individual behaviour type and predictability of the same behaviour. Using multivariate double hierarchical generalised linear models, we were able to quantify individual variation in feeding and lying behaviours of adult cows for the first time. Consistent variation between individuals was present for all measures of lying and feeding with repeatability values of 0.38 for daily feeding time and 0.34 for daily lying time. Individuals also varied in their levels of predictability with coefficients of 0.46 and 0.29 for the daily lying and feeding time respectively. Lastly, there were significant positive correlations between several behaviours and the amount of residual intra-individual variation for the same behaviour (mean lying bout length: 0.93, mean feeding bout length: 0.88), indicating that cows with the longest bouts of the two behaviours were also more variable and unpredictable, perhaps due to an opportunistic strategy. These results highlight the importance of measuring individual variation in behaviour, supporting the evidence of personality traits in cattle, as well as differences in predictability, which must be taken into account in any use of behavioural indicators for early disease diagnosis.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106357"},"PeriodicalIF":2.2,"publicationDate":"2024-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142442933","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-10DOI: 10.1016/j.prevetmed.2024.106355
Adrián Beato-Benítez , David Cano-Terriza , Moisés Gonzálvez , Iratxe Pérez-Cobo , Rosa Martínez-Valverde , Javier Martínez , Andrea Carretero , Andrea Ferreiro-Prado , Rafael Guerra , Miguel Ángel Quevedo-Muñoz , Ignacio García-Bocanegra
Leptospirosis is a worldwide zoonotic disease caused by serovars of Leptospira spp. that can infect a wide range of wild and domestic species, highlighting non-human primates (NHPs) as one of the most susceptible taxonomic groups. The aim of the present study was to determine the seroprevalence and potential risk factors associated with exposure to Leptospira spp. in captive NHPs in Spain. Between 2007 and 2021, sera were collected from 258 NHPs in 16 zoos and wildlife rescue centers (WRCs), and tested for antibodies to Leptospira spp. using the modified microagglutination test (MAT). Anti-Leptospira spp. antibodies were detected in 73 (28.3 %) of the 258 NHPs evaluated. Seropositivity was found in 61.0 % (25/41) of the species analyzed and in 87.5 % (14/16) of the sampled centers. Sera seropositive for six different serovars of Leptospira spp. were detected, with L. Grippotyphosa being the most prevalent. Seroprevalence was found to be significantly higher in Hominidae (61.8 %; P < 0.001) compared to other NHP families tested. To the author’s knowledge, the present study is the largest serosurvey of Leptospira spp. conducted in NHPs in Europe and also reports for the first time exposure to Leptospira spp. in nine NHP species, expanding the host range for this zoonotic bacterium. Our results indicate high and widespread seropositivity of Leptospira spp. in NHPs kept in captivity in Spain, which may be of conservation and animal health concern. This study supports the need to include captive NHPs in monitoring programs to evaluate the exposure of these species to Leptospira spp. in captive centers.
{"title":"Serosurvey of Leptospira spp. in captive non-human primates in Spain","authors":"Adrián Beato-Benítez , David Cano-Terriza , Moisés Gonzálvez , Iratxe Pérez-Cobo , Rosa Martínez-Valverde , Javier Martínez , Andrea Carretero , Andrea Ferreiro-Prado , Rafael Guerra , Miguel Ángel Quevedo-Muñoz , Ignacio García-Bocanegra","doi":"10.1016/j.prevetmed.2024.106355","DOIUrl":"10.1016/j.prevetmed.2024.106355","url":null,"abstract":"<div><div>Leptospirosis is a worldwide zoonotic disease caused by serovars of <em>Leptospira</em> spp. that can infect a wide range of wild and domestic species, highlighting non-human primates (NHPs) as one of the most susceptible taxonomic groups. The aim of the present study was to determine the seroprevalence and potential risk factors associated with exposure to <em>Leptospira</em> spp. in captive NHPs in Spain. Between 2007 and 2021, sera were collected from 258 NHPs in 16 zoos and wildlife rescue centers (WRCs), and tested for antibodies to <em>Leptospira</em> spp. using the modified microagglutination test (MAT). Anti-<em>Leptospira</em> spp. antibodies were detected in 73 (28.3 %) of the 258 NHPs evaluated. Seropositivity was found in 61.0 % (25/41) of the species analyzed and in 87.5 % (14/16) of the sampled centers. Sera seropositive for six different serovars of <em>Leptospira</em> spp. were detected, with <em>L</em>. Grippotyphosa being the most prevalent. Seroprevalence was found to be significantly higher in Hominidae (61.8 %; <em>P</em> < 0.001) compared to other NHP families tested. To the author’s knowledge, the present study is the largest serosurvey of <em>Leptospira</em> spp. conducted in NHPs in Europe and also reports for the first time exposure to <em>Leptospira</em> spp. in nine NHP species, expanding the host range for this zoonotic bacterium. Our results indicate high and widespread seropositivity of <em>Leptospira</em> spp. in NHPs kept in captivity in Spain, which may be of conservation and animal health concern. This study supports the need to include captive NHPs in monitoring programs to evaluate the exposure of these species to <em>Leptospira</em> spp. in captive centers.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106355"},"PeriodicalIF":2.2,"publicationDate":"2024-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142433962","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-09DOI: 10.1016/j.prevetmed.2024.106354
Théophile Ghislain Loïc Eyango Tabi , Maud Rouault , Victoria Potdevin , Xavier L’hostis , Sébastien Assié , Sébastien Picault , Nicolas Parisey
Bovine Respiratory Disease (BRD) is a prevalent infectious disease of respiratory tract in cattle, presenting challenges in accurate diagnosis and forecasting due to the complex interactions of multiple risk factors. Common methods, including mathematical epidemiological models and data-driven approaches such as machine learning models, face limitations such as difficult parameter estimation or the need for data across all potential outcomes, which is challenging given the scarcity and noise in observing BRD processes. In response to these challenges, we introduce a novel approach known as the Bayesian Deep Mechanistic method. This method couples a data-driven model with a mathematical epidemiological model while accounting for uncertainties within the processes. By utilising 265 lung ultrasound videos as sensor data from 163 animals across 9 farms in France, we trained a Bayesian deep learning model to predict the infection status (infected or non-infected) of an entire batch of 12 animals, also providing associated confidence levels. These predictions, coupled with their confidence levels, were used to filter out highly uncertain diagnoses and diffuse uncertainties into the parameterisation of a mathematical epidemiological model to forecast the progression of infected animals. Our findings highlight that considering the confidence levels (or uncertainties) of predictions enhances both the diagnosis and forecasting of BRD. Uncertainty-aware diagnosis reduced errors to 32 %, outperforming traditional automatic diagnosis. Forecast relying on veterinarian diagnoses, considered the most confident, had a 23 % error, whilst forecast taking into account diagnosis uncertainties had a close 27.2 % error. Building upon uncertainty-awareness, our future research could explore integrating multiple types of sensor data, such as video surveillance, audio recordings, and environmental parameters, to provide a comprehensive evaluation of animal health, employing multi-modal methods for processing this diverse data.
{"title":"Harnessing uncertainty: A deep mechanistic approach for cautious diagnostic and forecast of Bovine Respiratory Disease","authors":"Théophile Ghislain Loïc Eyango Tabi , Maud Rouault , Victoria Potdevin , Xavier L’hostis , Sébastien Assié , Sébastien Picault , Nicolas Parisey","doi":"10.1016/j.prevetmed.2024.106354","DOIUrl":"10.1016/j.prevetmed.2024.106354","url":null,"abstract":"<div><div>Bovine Respiratory Disease (BRD) is a prevalent infectious disease of respiratory tract in cattle, presenting challenges in accurate diagnosis and forecasting due to the complex interactions of multiple risk factors. Common methods, including mathematical epidemiological models and data-driven approaches such as machine learning models, face limitations such as difficult parameter estimation or the need for data across all potential outcomes, which is challenging given the scarcity and noise in observing BRD processes. In response to these challenges, we introduce a novel approach known as the Bayesian Deep Mechanistic method. This method couples a data-driven model with a mathematical epidemiological model while accounting for uncertainties within the processes. By utilising 265 lung ultrasound videos as sensor data from 163 animals across 9 farms in France, we trained a Bayesian deep learning model to predict the infection status (infected or non-infected) of an entire batch of 12 animals, also providing associated confidence levels. These predictions, coupled with their confidence levels, were used to filter out highly uncertain diagnoses and diffuse uncertainties into the parameterisation of a mathematical epidemiological model to forecast the progression of infected animals. Our findings highlight that considering the confidence levels (or uncertainties) of predictions enhances both the diagnosis and forecasting of BRD. Uncertainty-aware diagnosis reduced errors to 32 %, outperforming traditional automatic diagnosis. Forecast relying on veterinarian diagnoses, considered the most confident, had a 23 % error, whilst forecast taking into account diagnosis uncertainties had a close 27.2 % error. Building upon uncertainty-awareness, our future research could explore integrating multiple types of sensor data, such as video surveillance, audio recordings, and environmental parameters, to provide a comprehensive evaluation of animal health, employing multi-modal methods for processing this diverse data.</div></div>","PeriodicalId":20413,"journal":{"name":"Preventive veterinary medicine","volume":"233 ","pages":"Article 106354"},"PeriodicalIF":2.2,"publicationDate":"2024-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142546948","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号