Using state space models to monitor and estimate the effects of interventions on treatment risk and milk yield in dairy farms.

Fast, flexible, and internally valid analytical tools are needed to evaluate the effects of management interventions made on dairy farms to support decisions about which interventions to continue or discontinue. The objective of this observational study was to demonstrate the use of state space models (SSM) to monitor and estimate the effect of interventions on 2 specific outcomes: a dynamic linear model (DLM) evaluating herd-level milk yield and a dynamic generalized linear model evaluating treatment risk in a pragmatic pretest/posttest design under field conditions. This demonstration study is part of a Danish common learning project that ran from March 2020 to May 2021 within the framework of veterinary herd health consultancy in relation to reducing antimicrobial use and improving herd health. Specific interventions for 2 commercial herds were suggested by 4 visiting farmers and were implemented during the project period. The intervention for herd 1 was the application of teat sealers, implemented in August 2020. For herd 2, the intervention was an adjustment of cubicles for cows of parity 2 and above, implemented from November 2020. A shift to an automatic milking system in October 2020 was also modeled as an intervention for herd 1 because the 2 interventions coincided. Data available from the Danish Cattle Database on obligatory registrations for individual cow movements and treatments, as well as test day information on milk yield, were used for model building and testing. Data from a 3-yr period before the project were used to calibrate the SSM to herd conditions, and data from the study period (March 2020 to May 2021) were used for monitoring and intervention testing based on application of the SSM. Herd bulk tank milk recordings were added to the data set during the study period to increase the precision of the estimates in the DLM. The developed SSM monitored herd-level milk yield and the overall probability of treatment throughout the study period in both herds. Furthermore, at the time of intervention, the SSM estimated the effect on herd-level milk yield and treatment risk associated with the implemented intervention in each herd. The SSM were used because they can be calibrated to herd conditions and they take into account herd dynamics and autocorrelation and provide standard deviations of estimates. For herd 1, the intervention effect of applying teat sealers was inconclusive with the current SSM application. For herd 2, no statistically significant changes in cow treatment risk or milk production were identified following the adjustment of cubicles. The use of SSM on observational data under field conditions shows that in this case, the interventions had a nonspecific onset of effect, were implemented during unstable times, and had varying coherence with the measured outcomes, making fully automated SSM analysis difficult. However, similar or expanded SSM with both monitoring and effect estimation functions could, if applied under the right conditions, serve as improved data-based decision support tools for farmers (and veterinarians) to minimize the risk of misinterpreting data due to confounding bias related to dynamics in dairy herds.

Identifying Risk Factors for Low Piglet Birth Weight, High Within-Litter Variation and Occurrence of Intrauterine Growth-Restricted Piglets in Hyperprolific Sows.

This study aimed to identify risk factors affecting PBW, high CVPBW and the occurrence of IUGR piglets in 12 commercial Danish herds with hyperprolific sows using free-access stalls, floor or electronic sow feeding systems in the gestation unit. The following factors were investigated: the duration of previous lactation, the length of the interval from weaning to insemination, the length of gestation, litter size, parity, sow backfat thickness in late gestation and the type of feeding system in the gestation unit. The study included newborn piglets from 452 litters with the following production indicator averages: 21.3 piglets/L, 1235 g PBW, 22.9% CVPBW and 10.9% and 11.8% within-litter occurrence of severe and mild IUGR piglets, respectively. Increasing length of weaning-to-insemination interval decreased PBW by 25.8 g/day. For 2nd to 9th parity sows, each additional piglet in the litter increased CVPBW by 0.38%, the occurrence of severe IUGR piglets by 0.68% and mild IUGR piglets by 0.50%. Sows of 5th parity and older had a 1.39% higher CVPBW and 49.1 g lighter piglets compared with sows of 2nd to 4th parity. PBW was lower in one ESF herd, suggesting complex interactions that need to be further elucidated. The main critical risk factor observed was litter size.

The Effect of Different Feeding Systems on Salivary Cortisol Levels during Gestation in Sows on Herd Level.

The aim of this study was to investigate herd cortisol levels as an indicator of stress during gestation in three different feeding systems. Twelve commercial Danish herds with 800 to 3050 sows were included, with either free-access feeding stall (Stall), floor feeding (Floor), or electronic sow feeding (ESF; n = 4 herds per system). Saliva samples were collected from 30 sows/herd in the gestation unit for cortisol analysis with an average of 67.2 gestation days for ESF, 72.4 days for Floor, and 68.6 days for Stall. Data on piglet birth weight (PBW) and the percentage of intrauterine growth restricted (IUGR) piglets from 452 litters (9652 piglets, 8677 liveborn) from all 12 herds were obtained on the saliva collection days. The cortisol levels in saliva increased throughout gestation (p < 0.01), and lower concentrations were observed among sows belonging to Stall (4.80 nmol/L), compared to Floor (7.03 nmol/L) and ESF (7.87 nmol/L), and that difference was significant as an independent effect in the case of ESF (p < 0.01). There was no difference between Floor and ESF or Stall and Floor (p > 0.05). An interaction was observed between parity and feeding system, with parities 4-5 in ESF herds having lower levels than other parities within the ESF system (p = 0.02).

Epidemiology of Salmonella enterica Serovar Dublin in Cattle and Humans in Denmark, 1996 to 2016: a Retrospective Whole-Genome-Based Study.

Salmonella enterica serovar Dublin is a cattle-adapted S. enterica serovar causing both intestinal and systemic infection in its bovine host, and it is also a serious threat to human health. The present study aimed to determine the population structure of S Dublin isolates obtained from Danish cattle herds and to investigate how cattle isolates relate to Danish human isolates, as well as to non-Danish human and bovine isolates. Phylogenetic analysis of 197 Danish cattle isolates from 1996 to 2016 identified three major clades corresponding to distinct geographical regions of cattle herds. Persistence of closely related isolates within the same herd and their circulation between epidemiologically linked herds for a period of more than 20 years were demonstrated. These findings suggest that a lack of internal biosecurity and, to some extent, also a lack of external biosecurity in the herds have played an important role in the long-term persistence of S Dublin in Danish cattle herds in the period investigated. Global population analysis revealed that Danish cattle isolates clustered separately from bovine isolates from other countries, whereas human isolates were geographically spread. Resistance genes were not commonly demonstrated in Danish bovine isolates; only the isolates within one Danish clade were found to often harbor two plasmids of IncFII/IncFIB and IncN types, the latter plasmid carrying blaTEM-1, tetA, strA, and strB antibiotic resistance genes.IMPORTANCES Dublin causes economic losses in cattle production, and the bacterium is a public health concern. A surveillance and control program has been in place in Denmark since 2002 with the ultimate goal to eradicate S Dublin from Danish cattle herds; however, a small proportion of herds have remained positive for many years. In this study, we demonstrate that herds with persistent infection often were infected with the same strain for many years, indicating that internal biosecurity has to be improved to curb the infection. Further, domestic cases of S Dublin infection in humans were found to be caused both by Danish cattle isolates and by isolates acquired abroad. This study shows the strength of whole-genome sequencing to obtain detailed information on epidemiology of S Dublin and allows us to suggest internal biosecurity as a main way to control this bacterium in Danish cattle herds.

Application of Molecular Typing Results in Source Attribution Models: The Case of Multiple Locus Variable Number Tandem Repeat Analysis (MLVA) of Salmonella Isolates Obtained from Integrated Surveillance in Denmark.

Salmonella is an important cause of bacterial foodborne infections in Denmark. To identify the main animal-food sources of human salmonellosis, risk managers have relied on a routine application of a microbial subtyping-based source attribution model since 1995. In 2013, multiple locus variable number tandem repeat analysis (MLVA) substituted phage typing as the subtyping method for surveillance of S. Enteritidis and S. Typhimurium isolated from animals, food, and humans in Denmark. The purpose of this study was to develop a modeling approach applying a combination of serovars, MLVA types, and antibiotic resistance profiles for the Salmonella source attribution, and assess the utility of the results for the food safety decisionmakers. Full and simplified MLVA schemes from surveillance data were tested, and model fit and consistency of results were assessed using statistical measures. We conclude that loci schemes STTR5/STTR10/STTR3 for S. Typhimurium and SE9/SE5/SE2/SE1/SE3 for S. Enteritidis can be used in microbial subtyping-based source attribution models. Based on the results, we discuss that an adjustment of the discriminatory level of the subtyping method applied often will be required to fit the purpose of the study and the available data. The issues discussed are also considered highly relevant when applying, e.g., extended multi-locus sequence typing or next-generation sequencing techniques.