Use of rumination and activity monitoring for the identification of dairy cows with health disorders: Part II. Mastitis

Use of rumination time in health risk assessment of prepartum dairy cows

The objectives of this observational cohort study were to evaluate the associations of rumination time (RT) in the last week of pregnancy with transition cow metabolism, inflammation, health, and subsequent milk production, reproduction, and culling. Pregnant nulliparous (n = 199) and parous (n = 337) cows were enrolled 21 d before the expected calving. RT and physical activity were monitored automatically by sensors from d -21 to 15 relative to calving. Blood samples were collected on d -14, -5, 4, 8, and 12 ± 1 relative to calving. Diagnoses of clinical health problems were performed by researchers from calving to 15 d in milk (DIM). In classification 1, cows were ranked based on average daily RT in the last week of pregnancy and classified into terciles as short RT (SRT), moderate RT (MRT), or long RT (LRT) for association analyses. In classification 2, RT deviation from the parity average was used in a receiver operating characteristic curve to identify the best threshold to predict postpartum clinical disease. Cows were then classified as above the threshold (AT) or below the threshold (BT). Compared with cows with LRT, cows with SRT had greater serum concentrations of NEFA (0.47 vs 0.40 ± 0.01 mmol/L), BHB (0.58 vs 0.52 ± 0.01 mmol/L), and haptoglobin (0.22 vs 0.18 ± 0.008 g/L) throughout the transition period, and reduced concentrations of glucose, cholesterol, albumin, and magnesium in a time-dependent manner. Parous cows with SRT had higher odds of postpartum clinical disease (adjusted odds ratio [AOR]: 3.7; 95% confidence interval [CI]: 2.1-6.4), lower odds of pregnancy by 210 DIM (AOR: 0.34; CI: 0.15-0.75), and lower milk production (46.9 vs 48.6 ± 0.5 kg/d) than parous cows with LRT. Deviation in prepartum RT had good predictive value for clinical disease in parous cows (area under the curve [AUC]: 0.65; CI: 0.60-0.71) but not in nulliparous (AUC: 0.51; CI: 0.42-0.59). Separation of parous cows according to the identified threshold (≤-53 min from the parity average) resulted in differences in serum concentrations of NEFA (AT = 0.31 ± 0.006, BT = 0.38 ± 0.014 mmol/L), BHB (AT = 0.49 ± 0.008, BT = 0.53 ± 0.015 mmol/L), and globulin (AT = 32.3 ± 0.3, BT = 34.8 ± 0.5 g/L) throughout the transition period, as well as in serum cholesterol, urea, magnesium, albumin, and haptoglobin in a time-dependent manner. BT parous cows had higher odds of clinical disease (AOR: 3.7; CI: 2.1-6.4), reduced hazard of pregnancy (AHR: 0.64, CI: 0.47-0.89), greater hazard of culling (AHR: 2.1, CI: 1.2-3.6), and lower milk production (46.3 ± 0.7 vs 48.5 ± 0.3 kg/d). External validation using data from 153 parous cows from a different herd and the established threshold in RT deviation (≤-53 min) resulted in similar predictive value, with the odds of postpartum disease 2.4 times greater in BT than AT (37.5 vs 20.1%). In conclusion, RT in the wk preceding calving was a reasonable predictor of postpartum health and future milk production, reproduction, and culling in parous cows but not in nulliparous cows.

Behavioural response to gastrointestinal parasites of yearling dairy calves at pasture

AIMS

To investigate the association between gastrointestinal parasites (GIP) and animal behaviour in dairy calves under New Zealand pastoral conditions, using animal-mounted, accelerometer-based sensors.

METHODS

Thirty-six, 5-6-month-old, Friesian-Jersey, heifer calves fitted with animal activity sensors to track behaviour were randomly allocated to one of two treatment groups. Half the animals were challenged with an oral dose of 20,000 larvae of Ostertagia ostertagi and Cooperia oncophera once a week for 3 weeks and half were unchallenged. Five weeks after the last dose, seven infected and nine uninfected animals were treated with an oral anthelmintic (AHC) and data collected for a further week. Accelerometer data were classified into minutes per day eating, ruminating, in moderate-high activity or in low activity. Live weight and faecal egg counts (FEC) were recorded weekly over the study period. All animals co-grazed a newly sown pasture not previously grazed by ruminants and were moved every week to fresh grazing. Treatment status was blinded to those managing the animals which were otherwise treated identically.

RESULTS

Complete behavioural records were available from 30/36 calves, (13 challenged and 17 unchallenged). Before treatment with AHC, FEC increased in infected and un-treated calves over the study, while uninfected animals maintained a near zero FEC. There was no difference in live weight gain between the two groups over the study period. Bayesian, multinomial regression predicted differences in animal behaviour between infected and uninfected animals that were not treated with AHC over the 7 weeks following initial infection. Parasitised calves not treated with AHC were less active and spent up to 6 (95% highest density interval (HDI) = 1-11) minutes/day less in low level activity and up to 15 (95% HDI = 7-20) minutes/day less in moderate to high level activity. They ruminated up to 9 (95% HDI = 2-15) minutes/day more and ate up to 10 (95% HDI = 2-19) minutes/day more than control calves that were not treated with AHC. The effect of AHC on time spent in each behaviour differed between infected and uninfected calves and increased the coefficient of dispersion of the behavioural data.

CONCLUSIONS AND CLINICAL RELEVANCE

Small differences in animal behaviour can be measured in calves with GIP. However, to use this to target treatment, further validation studies are required to confirm the accuracy of behavioural classification and understand the complex drivers of animal behaviour in a dynamic and variable pasture-parasite-host environment.

The development and validation of a milk feeding behavior alert from automated feeder data to classify calves at risk for a diarrhea bout: A diagnostic accuracy study

The objective of this diagnostic accuracy study was to develop and validate an alert to identify calves at risk for a diarrhea bout using milk feeding behavior data (behavior) from automated milk feeders (AMF). We enrolled Holstein calves (n = 259) as a convenience sample size from 2 facilities that were health scored daily preweaning and offered either 10 or 15 L/d of milk replacer. For alert development, 132 calves were enrolled and the ability of milk intake, drinking speed, and rewarded visits collected from AMF to identify calves at risk for diarrhea was tested. Alerts that had high diagnostic accuracy in the alert development phase were validated using a holdout validation strategy of 127 different calves from the same facilities (all offered 15 L/d) for -3 to 1 d relative to diarrhea diagnosis. We enrolled calves that were either healthy or had a first diarrheal bout (loose feces ≥2 d or watery feces ≥1 d). Relative change and rolling dividends for each milk feeding behavior were calculated for each calf from the previous 2 d. Logistic regression models and receiver operator curves (ROC) were used to assess the diagnostic ability for relative change and rolling dividends behavior relative to alert d) to classify calves at risk for a diarrhea bout from -2 to 0 d relative to diagnosis. To maximize sensitivity (Se), alert thresholds were based on ROC optimal classification cutoff. Diagnostic accuracy was met when the alert had a moderate area under the ROC curve (≥0.70), high accuracy (Acc; ≥0.80), high Se (≥0.80), and very high precision (Pre; ≥0.85). For alert development, deviations in rolling dividend milk intake with drinking speed had the best performance (10 L/d: ROC area under the curve [AUC] = 0.79, threshold ≤0.70; 15 L/d: ROC AUC = 0.82, threshold ≤0.60). Our diagnostic criteria were only met in calves offered 15 L/d (10 L/d: Se 75%, Acc 72%, Pre 92%, specificity [Sp] 55% vs. 15 L/d: Se 91%, Acc 91%, Pre 89%, Sp 73%). For holdout validation, rolling dividend milk intake with drinking speed met diagnostic criteria for one facility (threshold ≤0.60, Se 86%, Acc 82%, Pre 94%, Sp 50%). However, no milk feeding behavior alerts met diagnostic criteria for the second facility due to poor Se (relative change milk intake -0.36 threshold, Se 71%, Acc 70%, and Pre 97%). We suggest that changes in milk feeding behavior may indicate diarrhea bouts in dairy calves. Future research should validate this alert in commercial settings; furthermore, software updates, support, and new analytics might be required for on-farm application to implement these types of alerts.

Evaluation of sensor-based health monitoring in dairy cows: Exploiting rumination times for health alerts around parturition

The use of sensor-based measures of rumination time as a parameter for early disease detection has received significant attention in scientific research. This study aimed to assess the accuracy of health alerts triggered by a sensor-based accelerometer system within 2 different management strategies on a commercial dairy farm. Multiparous Holstein cows were enrolled during the dry-off period and randomly allocated to conventional (CON) or sensor-based (SEN) management groups at calving. All cows were monitored for disorders for a minimum of 10 DIM following standardized operating procedures (SOPs). The CON group (n = 199) followed an established monitoring protocol on the farm. The health alerts of this group were not available during the study but were later included in the analysis. The SEN group (n = 197) was only investigated when the sensor system triggered a health alert, and a more intensive monitoring approach according to the SOPs was implemented. To analyze the efficiency of the health alerts in detecting disorders, the sensitivity (SE) and specificity (SP) of health alerts were determined for the CON group. In addition, all cows were divided into 3 subgroups based on the status of the health alerts and their health status, to retrospectively compare the course of rumination time. Most health alerts (87%, n = 217) occurred on DIM 1. For the confirmation of diagnoses, health alerts showed SE and SP levels of 71% and 47% for CON cows. In SEN cows, a SE of 71% and 75% and SP of 48% and 43% were found for the detection of ketosis and hypocalcemia, respectively. The rumination time of the subgroups was affected by DIM and the interaction between DIM and the status of health alert and health condition.

The Prediction of Clinical Mastitis in Dairy Cows Based on Milk Yield, Rumination Time, and Milk Electrical Conductivity Using Machine Learning Algorithms

In commercial dairy farms, mastitis is associated with increased antimicrobial use and associated resistance, which may affect milk production. This study aimed to develop sensor-based prediction models for naturally occurring clinical bovine mastitis using nine machine learning algorithms with data from 447 mastitic and 2146 healthy cows obtained from five commercial farms in Northeast China. The variables were related to daily activity, rumination time, and daily milk yield of cows, as well as milk electrical conductivity. Both Z-standardized and non-standardized datasets pertaining to four specific stages of lactation were used to train and test prediction models. For all four subgroups, the Z-standardized dataset yielded better results than those of the non-standardized one, with the multilayer artificial neural net algorithm showing the best performance. Variables of importance had a similar rank in this algorithm, indicating the consistency of these variables as predictors for bovine mastitis in commercial farms with similar automatic systems. Moreover, the peak milk yield (PMY) of mastitic cows was significantly higher than that of healthy cows (p < 0.005), indicating that high-yielding cattle are more prone to mastitis. Our results show that machine learning algorithms are effective tools for predicting mastitis in dairy cows for immediate intervention and management in commercial farms.

Associations among post-partum rumen fill and motility, subclinical ketosis and fertility in Holstein dairy cows

This prospective observational study aimed to investigate the association of rumen fill and motility in post-partum Holstein cows with their future reproductive performance and subclinical ketosis (SCK). The study population consisted of two independent data sets: the first (DS1) included 237 cows from 6 herds and the second one (DS2) 709 cows from 9 herds. Rumen Fill Score (RFS) was transformed into a 3 level-trait, representing very low, low and adequate dry matter intake, respectively. A binary Rumen Contraction Score (RCS) was defined as: 0: <2 contractions/2 min, impaired rumen motility and 1: ≥2 contractions/2 min, normal rumen motility. A combined binary trait based on RFS and RCS (RFCS) was also established, representing unsatisfactory and satisfactory rumen function. Three SCK traits were defined, based on 3 different thresholds, SCK_I: BHB≥1,000 mmol/L, SCK_II: BHB≥1,100 mmol/L and SCK_III: BHB≥1,200 mmol/L. Scores were assessed and blood samples collected on day 7 (DS1) or day 8 (DS2), postpartum. Kaplan-Meier survival analysis, multivariable Cox proportional hazards models and Generalized Linear Mixed Models were performed to evaluate the association of rumen and SCK traits with reproduction. Herd, parity, calving season and several postparturient diseases were also included as potential explanatory variables. Mean days from calving to pregnancy after the 1st artificial insemination (AI) and from calving to pregnancy (all AIs) were shorter for levels of rumen traits representing adequate DMI and normal rumen motility; in most cases these differences were statistically significant in both datasets. Cows with adequate DMI and normal rumen motility (only in DS2) had greater hazard (hazard ratio [HR] = 1.84 and 1.61, for RFS and RFCS, respectively) and odds (odds ratio [OR] = 2.49 and 1.98, for RFS and RFCS, respectively) for pregnancy at 1st AI. Assessment of the association of examined rumen traits with hazard and odds for pregnancy at all AIs yielded statistically significant results in both datasets. For RFS, RCS and RFCS, HRs ranged from 1.57 to 3.31 and ORs from 1.95 to 4.83. No statistically significant associations with hazard and odds for pregnancy at 1st or all AIs were detected, for any of the 3 SCK traits, in either dataset. Overall, the combined RFCS trait constantly identified more than twice the number of cows with future reproductive problems than a positive SCK blood test.

Effects of targeted clinical examination based on alerts from automated health monitoring systems on herd health and performance of lactating dairy cows

Our objectives were to compare the proportion of lactating dairy cows diagnosed with health disorders (HD) and herd performance when using a health monitoring program designed to rely primarily but not exclusively on alerts from automated health monitoring (AHM) systems or a health monitoring program based primarily on systematic clinical examinations, milk yield monitoring, and visual observation of cows. In a clinical trial, at ∼30 d before expected parturition, nulliparous and parous Holstein cows, stratified by parity and days in gestation, were randomly assigned to the high-intensity clinical monitoring (HIC-M; n = 625) or automated monitoring (AUT-M; n = 624) treatment. Cows were fitted with a neck-attached rumination and physical activity monitoring tag, and individual daily milk yield data were collected from parlor milk meters. For cows in HIC-M, clinical examination was conducted daily until 10 d in milk (DIM) and then in response to milk yield reduction alerts or visual observation of clinical signs of HD over the course of 21 DIM. For cows in AUT-M, clinical examination until 21 DIM was because of health index (HI) score alerts and reduced milk yield alerts. The HI score alerts used were generated based on the manufacturer's settings for the system for the last 2-h period before cows were selected for examination. Visual observation of clinical signs of HD was used for identifying cows potentially missed by automated alerts. Binomial and quantitative data were analyzed by logistic regression and ANOVA with repeated measures, respectively. The percentage of cows diagnosed with at least 1 HD during the experimental treatments risk period tended to be greater and the incidence rate ratio of HD diagnosed was greater in the HIC-M than in the AUT-M treatment. We found no difference between treatments for cows that exited the herd up to 60 or 150 DIM, but more cows tended to exit the herd from 61 to 150 DIM in the HIC-M than in the AUT-M treatment. No differences were detectable between treatments in daily or total milk yield to 21 DIM or in weekly mean milk yield and total milk yield to 150 DIM. More cows were inseminated in estrus for first service if in the HIC-M treatment and had no HD diagnosed than if in the HIC-M treatment but with HD diagnosed, or in the AUT-M treatment and had no HD diagnosed. Cows in the AUT-M treatment with HD diagnosed did not differ from other groups. No differences between treatments were observed in pregnancies per artificial insemination or pregnancy loss for first service. Despite a reduction in the risk of diagnosis of HD, no evidence indicated that a health monitoring program that relied on AHM system alerts to select cows for clinical examination reduced herd performance compared with a more intensive program that included systematic clinical examinations of all cows for the first 10 DIM, reduced milk yield alerts, and visual observation. However, to obtain the same herd performance as with the HIC-M treatment, the AUT-M treatment required use of visual observation. In conclusion, a health monitoring program designed to rely primarily on targeted clinical examination based on alerts from automated health monitoring systems might be a feasible alternative to programs that rely more on clinical examination, provided that visual observation is used to identify cows not detected by automated alerts.

Metabolic-digestive clinical disorders of lactating dairy cows were associated with alterations of rumination, physical activity, and lying behavior monitored by an ear-attached sensor

The objective of this observational cohort study was to characterize the pattern of rumination time (RT), physical activity (PA), and lying time (LT) monitored by an automated health monitoring system, based on an ear-attached sensor, immediately before, during, and after clinical diagnosis (CD) of metabolic-digestive disorders. Sensor data were collected from 820 lactating Holstein cows monitored daily from calving up to 21 DIM for detection of health disorders (HD). Cows were grouped retrospectively in the no-clinical health disorder group (NCHD; n = 616) if no HD were diagnosed, or the metabolic-digestive group (METB-DIG; n = 58) if diagnosed with clinical ketosis or indigestion only. Cows with another clinical health disorder within -7 to +7 d of CD of displaced abomasum, clinical ketosis, or indigestion were included in the metabolic-digestive plus one group (METB-DIG+1; n = 25). Daily RT, PA, and LT, and absolute and relative changes within -7 to +7 d of CD were analyzed with linear mixed models with or without repeated measures. Rumination time and PA were smaller, and LT was greater for the METB-DIG and METB-DIG+1 group than for cows in the NCHD group for most days from -7 to +7 d of CD of HD. In general, daily RT, PA, and LT differences were larger between the METB-DIG+1 and NCHD groups than between the METB-DIG and NCHD groups. In most cases, RT and PA decreased to a nadir and LT increased to a peak immediately before or after CD of HD, with a return to levels similar to the NCHD group within 7 d of CD. Absolute values and relative changes from 5 d before CD to the day of the nadir for RT and PA or peak for LT were different for cows in the METB-DIG and METB-DIG+1 group than for the NCHD group. For PA, the METB-DIG+1 group had greater changes than the METB-DIG group. For cows affected by metabolic-digestive disorders, RT, PA, and LT on the day of CD and resolution of clinical signs were different than for cows in the NCHD group, but an increase in RT and PA or a decrease in LT was observed from the day of CD to the day of resolution of clinical signs. We conclude that dairy cows diagnosed with metabolic-digestive disorders including displaced abomasum, clinical ketosis, and indigestion presented substantial alterations in the pattern of RT, PA, and LT captured by an ear-attached sensor. Thus, automated health monitoring systems based on ear-attached sensors might be used as an aid for identifying cows with metabolic-digestive disorders. Moreover, RT, PA, and LT changes after CD might be positive indicators of recovery from metabolic-digestive disorders.

Activity behaviors and relative changes in activity patterns recorded by precision technology were associated with diarrhea status in individually housed calves

The objective of this case-control study was to quantify any association of daily activity behaviors and relative changes in activity patterns (lying time, lying bouts, step count, activity index) with diarrhea status in preweaning dairy calves. Individually housed calves sourced from auction were health-scored daily for signs of diarrhea (fecal consistency loose or watery for 2 consecutive days) for the 28 d after arrival. Calves with diarrhea were pair-matched with healthy controls (n = 13, matched by arrival date, arrival weight, and diagnosis days to diarrheic calves). Mixed linear regression models were used to evaluate the association of diarrhea status, and the diarrhea status by day interaction with activity behaviors (d -3 to d 4) and relative changes in activity patterns (d -3 to d 4) relative to diagnosis of a diarrhea bout. The serum Brix percentage at arrival and daily temperature-humidity index from the calf barn were explored as quantitative covariates, with day as a repeated measure. The baseline for relative changes in activity patterns was set at 100% on d 0. Diarrheic calves were less active; they averaged fewer steps (119.1 ± 18.81 steps/d) than healthy calves (227.4 ± 18.81 steps/d, LSM ± SEM). Diarrheic calves also averaged lower activity indices (827.34 ± 93.092 daily index) than healthy calves (1,396.32 ± 93.092 daily index). We also found also a diarrhea status by day interaction for lying time on d -3, with diarrheic calves spending more time lying (20.80 ± 0.300 h/d) than healthy calves (19.25 ± 0.300 h/d). For relative changes in activity patterns, a diarrhea status by day interaction was detectable on d -2, where diarrheic calves had greater relative changes in step counts (diarrhea 634.85 ± 87.581% vs. healthy 216.51 ± 87.581%) and activity index (diarrhea 316.83 ± 35.692% vs. healthy 150.68 ± 35.692%). Lying bouts were not associated with diarrhea status. These results show that diarrheic calves were more lethargic, and they had relative changes in activity patterns 2 d before clinical signs of diarrhea. Future research should explore the potential of an activity alert that positively indicates an individually housed calf at risk for a diarrhea bout using deviations from relative changes in individual calf activity patterns.

Metritis and clinical mastitis events in lactating dairy cows were associated with altered patterns of rumination, physical activity, and lying behavior monitored by an ear-attached sensor

Understanding changes in parameters recorded by automated health monitoring systems based on ear-attached sensors on the days immediately before and after diagnosis of metritis and clinical mastitis can help develop dairy cow health monitoring strategies. The objective of this observational cohort study was to characterize rumination time, physical activity, and lying time monitored by an ear-attached sensor before, during, and after clinical diagnosis (CD) of metritis and clinical mastitis. Lactating Holsteins monitored daily for 21 d in milk for detection of health disorders were retrospectively included in the no clinical health disorder group (NCHD; n = 616) if no disorders were diagnosed. Cows were included in the metritis (MET; n = 69) or clinical mastitis (MAST; n = 36) group if diagnosed only with nonsevere metritis (watery, reddish, and fetid uterine discharge with or without pyrexia) or nonsevere clinical mastitis (visibly abnormal milk secretion with or without signs of udder inflammation, with no pyrexia and no systemic signs of disease), respectively. Cows diagnosed with severe metritis (signs of metritis plus systemic signs of disease) or severe clinical mastitis (signs of mastitis plus pyrexia and systemic signs of disease), and cows diagnosed with nonsevere metritis or clinical mastitis plus another disorder within -7 to +7 d of CD of metritis or clinical mastitis diagnosis, were included in the metritis plus (MET+; n = 25) or the clinical mastitis plus (MAST+; n = 15) group, respectively. Cows were fitted with an ear-attached accelerometer to measure rumination time, physical activity, and lying time. Mean daily values, mean value absolute change, and relative change for the mean daily value from 3 or 5 d before CD to the nadir for cows with metritis and clinical mastitis, respectively, were analyzed with linear mixed models with or without repeated measures. Rumination time and physical activity were lesser, and lying time was greater for the MET and MET+ groups than for the NCHD group for most days from -4 to +7 d of CD of metritis. Generally, daily rumination time, physical activity, and lying time differences were greater and more prolonged between the MET+ and NCHD than between the MET and NCHD groups. Similarly, cows in the MAST and MAST+ groups had lesser rumination time and physical activity than cows in the NCHD group for several days before diagnosis. Lying time was greater for the MAST+ than the NCHD group on d -1 and 0 relative to CD. Absolute values and relative changes from 3 d before CD to the day of the nadir for rumination time and physical activity, or peak for lying time, were different for cows in the MET and MET+ groups than for the NCHD group. Similar results were observed for the MAST and MAST+ groups compared with the NCHD group. For cows with metritis, either an increase in rumination time and physical activity or a decrease in lying time was observed from the day of CD to resolution of clinical signs, but no changes were observed for the NCHD. Cows with clinical mastitis and the NCHD group had different rumination times, physical activity, and lying times on the day of CD and resolution of clinical signs, but cows with clinical mastitis had no significant changes from the day of CD to resolution of clinical signs. We conclude that cows affected by metritis and clinical mastitis presented substantial alterations of the patterns of rumination time, physical activity, and lying time captured by an ear-attached sensor. Thus, automated health monitoring systems based on ear-attached sensors might be used as an aid for identifying cows with metritis and clinical mastitis. Moreover, behavioral parameter changes after CD might be good indicators of resolution of clinical signs of metritis but not mastitis.

Behavioral patterns as indicators of resilience after parturition in dairy cows

During the transition phase, dairy cows are susceptible to develop postpartum diseases. Cows that stay healthy or recover rapidly can be considered to be more resilient in comparison to those that develop postpartum diseases. An indication of loss of resilience will allow for early intervention with preventive and supportive measures before the onset of disease. We investigated which quantitative behavioral characteristics during the dry period could be used as indicators of reduced resilience after calving, using noninvasive Smart Tag neck and Smart Tag leg sensors in dairy cows (Nedap N.V.). We followed 180 cows during 2 wk before until 6 wk after parturition at 4 farms in the Netherlands. Serving as proxy for loss of resilience, as defined by the duration and severity of disease, a clinical assessment was performed twice weekly and blood samples were taken in the first and fifth week after parturition. For each cow, clinical and serum value deviations were aggregated into a total deficit score (TDS total). We also calculated TDS values relating to inflammation, locomotion, or metabolic problems, which were further divided into macro-mineral and liver-related deviations. Smart Tag neck and leg sensors provided continuous behavioral activity signals of which we calculated the average, variance, and autocorrelation during the dry period. Diurnal patterns in the behavioral activity signals were derived by fast Fourier transformation and the calculation of the nonperiodicity. To select significant predictors of resilience, we first performed a univariate analysis with TDS as dependent variable and the behavioral characteristics that were measured during the dry period, as potential predictors with cow as experimental unit. We included parity group as fixed effect and farm as random effect. Next, we performed multivariable analysis with only significant predictors, followed by a variable selection procedure to obtain a final linear mixed model with an optimal subset of predictors with parity group as fixed effect and farm as random effect. The TDS total was best predicted by average inactive time, nonperiodicity ruminating, nonperiodicity of bouts standing up and fast Fourier transformation stand still. Average inactive time was negatively correlated with average eating time, and these 2 predictors could be exchanged with only little difference in model performance. Our best performing model predicted TDS total at a cutoff level of 60 points, with a sensitivity of 79.5% and a specificity of 73.2% with a positive predicted value of 0.69 and a negative predicted value of 0.83. The models to predict the other TDS categories showed a lower predictive performance as compared with the TDS total model, which could be related to the limited sample size and therefore, low occurrence of problems within a specific TDS category. Furthermore, more resilient dairy cows are characterized by high averages of eating time with high regularity in rumination and low averages of inactive time. They reveal high regularity in standing time and transitions from lying to standing, in the dry period. These behaviors can be used as indicators of resilience and allow for preventive intervention during the dry period in vulnerable dairy cattle. However, further examination is still required to find clues for adequate intervention strategies.

A randomized clinical trial evaluating the effects of administration of acidogenic boluses at dry-off on rumination and activity behavior in the 14 subsequent days

Elevated milk production at dry-off can lead to increased udder pressure and, in turn, increased stress due to pain and discomfort, affecting natural behaviors. Administering acidogenic boluses at dry-off acts by inducing temporary and mild decreases in blood pH. This decreases dry matter intake, reduces milk yield, and increases cow comfort by lessening udder pressure. The objective of this study was to assess the effect of oral administration of acidogenic boluses at dry-off on total daily activity (TDA) and total daily rumination (TDR) behaviors in the first 2 wk of the dry period. This randomized clinical trial was conducted on a single farm and cows were randomly assigned to either treatment (TRT; n = 30) or control (CON; n = 34). The TRT group received 2 acidogenic boluses at dry-off and the CON group received no intervention. All cows received dry-cow therapy (intramammary antibiotic and internal teat sealant). The TDA and TDR data from 7 d before to 14 d after dry-off were measured using ear-mounted activity monitors. Analyses were performed using linear mixed-effects models with repeated measures. We observed a similar TDA in both groups throughout the study follow-up period. Overall, cows in the TRT group spent 17 min/d less time active than cows in the CON group in the first 2 wk after dry-off with the greatest difference observed on the second day of the dry period (TRT = 395 min/d; 95% CI: 370 to 420 vs. CON = 428 min/d; 95% CI: 404 to 451). The TRT group had lower TDR in the first 24 h after bolus administration (TRT = 437 min/d; 95% CI: 414 to 461 vs. CON = 488 min/d; 95% CI: 466 to 510) when compared with the CON group, but no differences were observed when comparing both groups in the 13 subsequent days. Our results indicate that administering acidogenic boluses at dry-off slightly decreased TDA during the first 2 wk of the dry period and decreased TDR on the first day after administration.

Uncovering Patterns in Dairy Cow Behaviour: A Deep Learning Approach with Tri-Axial Accelerometer Data

The accurate detection of behavioural changes represents a promising method of detecting the early onset of disease in dairy cows. This study assessed the performance of deep learning (DL) in classifying dairy cows' behaviour from accelerometry data acquired by single sensors on the cows' left flanks and compared the results with those obtained through classical machine learning (ML) from the same raw data. Twelve cows with a tri-axial accelerometer were observed for 136 ± 29 min each to detect five main behaviours: standing still, moving, feeding, ruminating and resting. For each 8 s time interval, 15 metrics were calculated, obtaining a dataset of 211,720 observation units and 15 columns. The entire dataset was randomly split into training (80%) and testing (20%) datasets. The DL accuracy, precision and sensitivity/recall were calculated and compared with the performance of classical ML models. The best predictive model was an 8-layer convolutional neural network (CNN) with an overall accuracy and F1 score equal to 0.96. The precision, sensitivity/recall and F1 score of single behaviours had the following ranges: 0.93-0.99. The CNN outperformed all the classical ML algorithms. The CNN used to monitor the cows' conditions showed an overall high performance in successfully predicting multiple behaviours using a single accelerometer.

Eucalcemia during lipopolysaccharide challenge in postpartum dairy cows: I. Clinical, inflammatory, and metabolic response

Hypocalcemia induced by immune activation is a conserved response across mammalian species; however, administration of Ca is discouraged in other species as it is associated with increased morbidity and mortality. Early postpartum cows experience a decrease in circulating Ca concentration following acute inflammation. Corrective Ca therapy during the transition period, particularly in dairy cows experiencing acute disease, is common practice. However, the effect of Ca administration on the inflammatory response during acute immune activation is unknown. Our objective was to compare the clinical, inflammatory, and metabolic response to an intravenous (IV) lipopolysaccharide (LPS) challenge between postpartum cows infused, or not, with IV Ca to maintain eucalcemia. Cows (n = 14, 8 ± 1 d in milk) were enrolled in a matched-pair randomized controlled design to receive IV Ca (IVCa) or sterile 0.9% NaCl (CTRL) during an IV LPS challenge (0.040 or 0.045 µg of LPS/kg of body weight over 1 h). Ionized Ca (iCa) was monitored cow-side, and IV Ca infusion was adjusted in a eucalcemic clamp for 12 h following the start of LPS infusion. Cows were monitored during the 24 h following challenge and serial blood samples were collected to quantify concentrations of glucose, β-hydroxybutyrate, nonesterified fatty acids, urea nitrogen, cytokines, acute-phase proteins, and cortisol. Blood iCa concentration decreased to 0.87 ± 0.03 mM in CTRL during challenge, and by design, iCa concentration was maintained within 3% of baseline in IVCa. Body temperature, heart rate, and respiratory rate were monitored for 24 h following the start of challenge and did not differ between groups. A treatment × time interaction was identified such that serum cortisol concentrations increased in both groups at 2 h but decreased to a greater extent at 6 h in IVCa compared with CTRL. Rumination time (min/h) over the first 12 h following challenge was greater in IVCa, but total rumination time in the 24 h following challenge did not differ from CTRL. Serum glucose and nonesterified fatty acid concentrations decreased, and β-hydroxybutyrate and urea nitrogen concentrations increased over time, but did not differ between groups. Acute leukopenia occurred in both groups at 4 h before leukocytosis was observed at 24 h with total white blood cell counts returning to baseline within 72 h. Plasma concentrations of tumor necrosis factor (TNF) and interleukin-10 (IL-10) increased within 1 h following the start of challenge and did not differ between groups. Serum haptoglobin and serum amyloid A concentrations increased within the 24 h following challenge and were elevated through 72 h but did not differ between groups. Eucalcemia during the acute systemic inflammatory response did not alter the TNF or IL-10 cytokine response, or the acute-phase protein SAA and haptoglobin response in this LPS challenge model; however, eucalcemia was associated with a more rapid decline in cortisol response and greater rumination time in the first 12 h following challenge. We did not find evidence that eucalcemia exacerbated the inflammatory response in early postpartum cows, but Ca administration may alter the clinical response to acute systemic inflammation.

Association between Rumination Times Detected by an Ear Tag-Based Accelerometer System and Rumen Physiology in Dairy Cows

Monitoring rumination activity is considered a useful indicator for the early detection of diseases and metabolic disorders. Accelerometer-based sensor systems provide health alerts based on individual thresholds of rumination times in dairy cows. Detailed knowledge of the relationship between sensor-based rumination times and rumen physiology would help detect conspicuous animals and evaluate the treatment's success. This study aimed to investigate the association between sensor-based health alerts and rumen fluid characteristics in Holstein-Friesian cows at different stages of lactation. Rumen fluid was collected via a stomach tube from 63 pairs of cows with and without health alerts (ALRT vs NALRT). Pairs were matched based on the day of lactation, the number of lactations, and health criteria. Rumen fluid was collected during and after health alerts. The parameters of color, odor, consistency, pH, redox potential, sedimentation flotation time, and the number of protozoa were examined. Results showed differences between both groups in odor, rumen pH, sedimentation flotation time, and protozoan count at the first rumen fluid collection. Within the groups, greater variations in rumen fluid parameters were found for ALRT cows compared to NALRT cows. The interaction between health alert and stage of lactation did not affect the rumen fluid parameters.

Evaluation of an indwelling bolus equipped with a triaxial accelerometer for the characterisation of the diurnal pattern of bovine reticuloruminal contractions

This observational study aimed to describe the diurnal pattern of reticuloruminal contraction rate (RRCR) and the proportion of time spent ruminating by cattle, using two commercial devices equipped with triaxial accelerometers: an indwelling bolus (placed in the reticulum) and a neck collar. The three objectives of this study were firstly to determine whether the indwelling bolus provided observations consistent with RRCR as determined by clinical examination using auscultation and ultrasound, secondly to compare estimates of time spent ruminating using the indwelling bolus and a collar-based accelerometer, and finally to describe the diurnal pattern of RRCR using the indwelling bolus data. Six rumen-fistulated, non-lactating Jersey cows were fitted with an indwelling bolus (SmaXtec Animal Care GmbH, Graz, Austria) and a neck collar (Silent Herdsman, Afimilk Ltd. Kibbutz Afikim, Israel), and data were collected over two weeks. Cattle were housed together in a single straw-bedded pen and fed ad libitum hay. To assess the agreement between the indwelling bolus and traditional methods of assessing reticuloruminal contractility in the first week, the RRCR was determined over 10 min, twice a day, by ultrasound and auscultation. Mean inter-contraction intervals (ICI) derived from bolus and ultrasound, and from auscultation were 40.4 ± 4.7, 40.1 ± 4.0 and 38.4 ± 3.3 s. Bland-Altmann plots showed similar performance of the methods with small biases. The Pearson correlation coefficient for the time spent ruminating derived from neck collars and indwelling boluses was 0.72 (highly significant, P < 0.001). The indwelling boluses generated a consistent diurnal pattern for all the cows. In conclusion, a robust relationship was observed between clinical observation and the indwelling boluses for estimation of ICI and, similarly, between the indwelling bolus and neck collar for estimating rumination time. The indwelling boluses showed a clear diurnal pattern for RRCR and time spent ruminating, indicating that they should be useful for assessing reticuloruminal motility.

Emergency vaccination of cattle against lumpy skin disease: Evaluation of safety, efficacy, and potency of MEVAC® LSD vaccine containing Neethling strain

Lumpy skin disease (LSD) is an emerging disease of cattle causing significantly high economic losses. Control of LSD depends on the use of homologous attenuated LSD virus strains isolated originally from South Africa (the Neethling strain). The virus belongs to the genus Capripoxvirus, which includes sheep pox virus and goat pox virus. The present study was conducted to evaluate the safety and efficacy of a new live attenuated LSD vaccine produced by Middle East for Vaccines (MEVAC^®) based on the Neethling strain. Tests were performed both in Egypt and Vietnam. Safety was evaluated by inoculation of five cattle with 10 times the recommended dose and observation of the animals for 14 days. Immunogenicity was tested at different periods post-vaccination (PV) in animals receiving the recommended doses of the vaccine using ELISA and virus neutralization test. Five cows were used to determine the protection index (PI) and non-vaccinated control cattle were included. Three calves were challenged by intradermal inoculation of the wild virus (5 × 10⁵ TCID₅₀) 28 days PV. Field or mass vaccination experiments were conducted in Vietnam during national campaigns in the summer of 2021 with 4301 vaccinated animals closely monitored after vaccination. In the field, around 2% (80/4301) of the animals showed hyper-reactivity, and 0.6% (24/4301) showed small skin swellings that disappeared within few hours PV. Abortion was recorded in three animals (0.3% 3/867). Challenged animals were resistant to clinical disease and PI value was 3.5 log₁₀. Meanwhile, antibody levels determined by the ELISA were inconsistent among animals and laboratories during the study period. Overall, the findings point to a new safe and effective LSD vaccine.

Behavioral Fingerprinting: Acceleration Sensors for Identifying Changes in Livestock Health

During disease or toxin challenges, the behavioral activities of grazing animals alter in response to adverse situations, potentially providing an indicator of their welfare status. Behavioral changes such as feeding behavior, rumination and physical behavior as well as expressive behavior, can serve as indicators of animal health and welfare. Sometimes behavioral changes are subtle and occur gradually, often missed by infrequent visual monitoring until the condition becomes acute. There is growing popularity in the use of sensors for monitoring animal health. Acceleration sensors have been designed to attach to ears, jaws, noses, collars and legs to detect the behavioral changes of cattle and sheep. So far, some automated acceleration sensors with high accuracies have been found to have the capacity to remotely monitor the behavioral patterns of cattle and sheep. These acceleration sensors have the potential to identify behavioral patterns of farm animals for monitoring changes in behavior which can indicate a deterioration in health. Here, we review the current automated accelerometer systems and the evidence they can detect behavioral patterns of animals for the application of potential directions and future solutions for automatically monitoring and the early detection of health concerns in grazing animals.

Sensitivity and Specificity for the Detection of Clinical Mastitis by Automatic Milking Systems in Bavarian Dairy Herds

In automatic milking systems (AMSs), the detection of clinical mastitis (CM) and the subsequent separation of abnormal milk should be reliably performed by commercial AMSs. Therefore, the objectives of this cross-sectional study were (1) to determine the sensitivity (SN) and specificity (SP) of CM detection of AMS by the four most common manufacturers in Bavarian dairy farms, and (2) to identify routinely collected cow data (AMS and monthly test day data of the regional Dairy Herd Improvement Association (DHIA)) that could improve the SN and SP of clinical mastitis detection. Bavarian dairy farms with AMS from the manufacturers DeLaval, GEA Farm Technologies, Lely, and Lemmer-Fullwood were recruited with the aim of sampling at least 40 cows with clinical mastitis per AMS manufacturer in addition to clinically healthy ones. During a single farm visit, cow-level milking information was first electronically extracted from each AMS and then all lactating cows examined for their udder health status in the barn. Clinical mastitis was defined as at least the presence of visibly abnormal milk. In addition, available DHIA test results from the previous six months were collected. None of the manufacturers provided a definition for clinical mastitis (i.e., visually abnormal milk), therefore, the SN and SP of AMS warning lists for udder health were assessed for each manufacturer individually, based on the clinical evaluation results. Generalized linear mixed models (GLMMs) with herd as random effect were used to determine the potential influence of routinely recorded parameters on SN and SP. A total of 7411 cows on 114 farms were assessed; of these, 7096 cows could be matched to AMS data and were included in the analysis. The prevalence of clinical mastitis was 3.4% (239 cows). When considering the 95% confidence interval (95% CI), all but one manufacturer achieved the minimum SN limit of >80%: DeLaval (SN: 61.4% (95% CI: 49.0%−72.8%)), GEA (75.9% (62.4%−86.5%)), Lely (78.2% (67.4%−86.8%)), and Lemmer-Fullwood (67.6% (50.2%−82.0%)). However, none of the evaluated AMSs achieved the minimum SP limit of 99%: DeLaval (SP: 89.3% (95% CI: 87.7%−90.7%)), GEA (79.2% (77.1%−81.2%)), Lely (86.2% (84.6%−87.7%)), and Lemmer-Fullwood (92.2% (90.8%−93.5%)). All AMS manufacturers’ robots showed an association of SP with cow classification based on somatic cell count (SCC) measurement from the last two DHIA test results: cows that were above the threshold of 100,000 cells/mL for subclinical mastitis on both test days had lower chances of being classified as healthy by the AMS compared to cows that were below the threshold. In conclusion, the detection of clinical mastitis cases was satisfactory across AMS manufacturers. However, the low SP will lead to unnecessarily discarded milk and increased workload to assess potentially false-positive mastitis cases. Based on the results of our study, farmers must evaluate all available data (test day data, AMS data, and daily assessment of their cows in the barn) to make decisions about individual cows and to ultimately ensure animal welfare, food quality, and the economic viability of their farm.

High Precision Classification of Resting and Eating Behaviors of Cattle by Using a Collar-Fitted Triaxial Accelerometer Sensor

Cattle are less active than humans. Hence, it was hypothesized in this study that transmitting acceleration signals at a 1 min sampling interval to reduce storage load has the potential to improve the performance of motion sensors without affecting the precision of behavior classification. The behavior classification performance in terms of precision, sensitivity, and the F1-score of the 1 min serial datasets segmented in 3, 4, and 5 min window sizes based on nine algorithms were determined. The collar-fitted triaxial accelerometer sensor was attached on the right side of the neck of the two fattening Korean steers (age: 20 months) and the steers were observed for 6 h on day one, 10 h on day two, and 7 h on day three. The acceleration signals and visual observations were time synchronized and analyzed based on the objectives. The resting behavior was most correctly classified using the combination of a 4 min window size and the long short-term memory (LSTM) algorithm which resulted in 89% high precision, 81% high sensitivity, and 85% high F1-score. High classification performance (79% precision, 88% sensitivity, and 83% F1-score) was also obtained in classifying the eating behavior using the same classification method (4 min window size and an LSTM algorithm). The most poorly classified behavior was the active behavior. This study showed that the collar-fitted triaxial sensor measuring 1 min serial signals could be used as a tool for detecting the resting and eating behaviors of cattle in high precision by segmenting the acceleration signals in a 4 min window size and by using the LSTM classification algorithm.

Opportunities to monitor animal welfare using the five freedoms with precision livestock management on rangelands

Advances in technology have led to precision livestock management, a developing research field. Precision livestock management has potential to improve sustainable meat production through continuous, real-time tracking which can help livestock managers remotely monitor and enhance animal welfare in extensive rangeland systems. The combination of global positioning systems (GPS) and accessible data transmission gives livestock managers the ability to locate animals in arduous weather, track animal patterns throughout the grazing season, and improve handling practices. Accelerometers fitted to ear tags or collars have the potential to identify behavioral changes through variation in the intensity of movement that can occur during grazing, the onset of disease, parturition or responses to other environmental and management stressors. The ability to remotely detect disease, parturition, or effects of stress, combined with appropriate algorithms and data analysis, can be used to notify livestock managers and expedite response times to bolster animal welfare and productivity. The “Five Freedoms” were developed to help guide the evaluation and impact of management practices on animal welfare. These freedoms and welfare concerns differ between intensive (i.e., feed lot) and extensive (i.e., rangeland) systems. The provisions of the Five Freedoms can be used as a conceptual framework to demonstrate how precision livestock management can be used to improve the welfare of livestock grazing on extensive rangeland systems.

The Early Prediction of Common Disorders in Dairy Cows Monitored by Automatic Systems with Machine Learning Algorithms

We use multidimensional data from automated monitoring systems and milking systems to predict disorders of dairy cows by employing eight machine learning algorithms. The data included the season, days in milking, parity, age at the time of disorders, milk yield (kg/day), activity (unitless), six variables related to rumination time, and two variables related to the electrical conductivity of milk. We analyze 131 sick cows and 149 healthy cows with identical lactation days and parity; all data are collected on the same day, which corresponds to the diagnosis day for disordered cows. For disordered cows, each variable, except the ratio of rumination time from daytime to nighttime, displays a decreasing/increasing trend from d-7 or d-3 to d0 and/or d-1, with the d0, d-1, or d-2 values reaching the minimum or maximum. The test data sensitivity for three algorithms exceeded 80%, and the accuracies of the eight algorithms ranged from 65.08% to 84.21%. The area under the curve (AUC) of the three algorithms was >80%. Overall, Rpart best predicts the disorders with an accuracy, precision, and AUC of 81.58%, 92.86%, and 0.908, respectively. The machine learning algorithms may be an appropriate and powerful decision support and monitoring tool to detect herds with common health disorders.

Assessment of feeding, ruminating and locomotion behaviors in dairy cows around calving – a retrospective clinical study to early detect spontaneous disease appearance

The study aims to verify the usefulness of new intervals-based algorithms for clinical interpretation of animal behavior in dairy cows around calving period. Thirteen activities associated with feeding-ruminating-locomotion-behaviors of 42 adult Holstein-Friesian cows were continuously monitored for the week (wk) -2, wk -1 and wk +1 relative to calving (overall 30’340 min/animal). Soon after, animals were retrospectively assigned to group-S (at least one spontaneous diseases; n = 24) and group-H (healthy; n = 18). The average activities performed by the groups, recorded by RumiWatch® halter and pedometer, were compared at the different weekly intervals. The average activities on the day of clinical diagnosis (dd0), as well as one (dd-1) and two days before (dd-2) were also assessed. Differences of dd0 vs. dd-1 (ΔD1), dd0 vs. wk -1 (ΔD2), and wk +1 vs. wk -1 (Δweeks) were calculated. Variables showing significant differences between the groups were used for a univariate logistic regression, a receiver operating characteristic analysis, and a multivariate logistic regression model. At wk +1 and dd0, eating- and ruminating-time, eating- and ruminate-chews and ruminating boluses were significantly lower in group-S as compared to group-H, while other activity time was higher. For ΔD2 and Δweeks, the differences of eating- and ruminating-time, as well as of eating-and ruminate-chews were significantly lower in group-S as compared to group-H. Concerning the locomotion behaviors, the lying time was significantly higher in group-S vs. group-H at wk +1 and dd-2. The number of strides was significantly lower in group-S compared to group-H at wk +1. The model including eating-chews, ruminate-chews and other activity time reached the highest accuracy in detecting sick cows in wk +1 (area under the curve: 81%; sensitivity: 73.7%; specificity: 82.4%). Some of the new algorithms for the clinical interpretation of cow behaviour as described in this study may contribute to monitoring animals’ health around calving.

Sensor based time budgets in commercial Dutch dairy herds vary over lactation cycles and within 24 hours

Cows from 8 commercial Dutch dairy farms were equipped with 2 sensors to study their complete time budgets of eating, rumination, lying, standing and walking times as derived from a neck and a leg sensor. Daily sensor data of 1074 cows with 3201 lactations was used from 1 month prepartum until 10 months postpartum. Farms provided data over a 5 year period. The final models (lactational time budget and 24h time budget) showed significant effects of parity, farm and calving season. When primiparous cows were introduced in the lactational herd, they showed a decrease in lying time of 215 min (95% CI: 187–242) and an increase in standing time of 159 min (95% CI: 138–179), walking time of 23 min (95% CI: 20–26) and rumination time of 69 min (95% CI: 57–82). Eating time in primiparous cows increased from 1 month prepartum until 9 months in lactation with 88 min (95% CI: 76–101) and then remained stable until the end of lactation. Parity 2 and parity 3+ cows decreased in eating time by 30 min (95% CI: 20–40) and 26 min (95% CI: 18–33), respectively, from 1 month before to 1 month after calving. Until month 6, eating time increased 11 min (95% CI: 1–22) for parity 2, and 24 min (95% CI: 16–32) for parity 3+. From 1 month before calving to 1 month after calving, they showed an increase in ruminating of 17 min (95% CI: 6–28) and 28 min (95% CI: 21–35), an increase in standing time of 117 min (95% CI: 100–135) and 133 min (95% CI: 121–146), while lying time decreased with 113 min (95% CI: 91–136) and 130 min (95% CI: 114–146), for parity 2 and 3+, respectively. After month 1 in milk to the end of lactation, lying time increased 67 min (95% CI: 49–85) for parity 2, and 77 min (95% CI: 53–100) for parity 3+. Lactational time budget patterns are comparable between all 8 farms, but cows on conventional milking system (CMS) farms with pasture access appear to show higher standing and walking time, and spent less time lying compared to cows on automatic milking system (AMS) farms without pasture access. Every behavioral parameter presented a 24h pattern. Cows eat, stand and walk during the day and lie down and ruminate during the night. Daily patterns in time budgets on all farms are comparable except for walking time. During the day, cows on CMS farms with pasture access spent more time walking than cows on AMS farms without pasture access. The average 24h pattern between parities is comparable, but primiparous cows spent more time walking during daytime compared to older cows. These results indicate a specific behavioral pattern per parameter from the last month prepartum until 10 months postpartum with different patterns between parities but comparable patterns across farms. Furthermore, cows appear to have a circadian rhythm with varying time budgets in the transition period and during lactation.

Time-Consuming, but Necessary: A Wide Range of Measures Should Be Included in Welfare Assessments for Dairy Herds

Animal welfare assessments that measure welfare outcomes, including behavior and health, can be highly valid. However, the time and skill required are major barriers to their use. We explored whether feasibility of welfare outcome assessment for dairy herds may be improved by rationalizing the number of measures included. We compared two approaches: analyzing whether strong pairwise associations between measures existed, enabling the subsequent exclusion of associated measures; and identifying possible summary measures—“iceberg indicators”—of dairy herd welfare that could predict herd welfare status. A cross-sectional study of dairy herd welfare was undertaken by a single assessor on 51 English farms, in which 96 welfare outcome measures were assessed. All measures showed at least one pairwise association; percentage of lame cows showed the most (33 correlations). However, most correlations were weak–moderate, suggesting limited scope for excluding measures from protocols based on pairwise relationships. A composite measure of the largest portion of herd welfare status was then identified via Principal Component Analysis (Principal Component 1, accounting for 16.9% of variance), and linear regression revealed that 22 measures correlated with this. Of these 22, agreement statistics indicated that percentage of lame cows and qualitative descriptors of “calmness” and “happiness” best predicted Principal Component 1. However, even these correctly classified only ~50% of farms according to which quartile of the Principal Component 1 they occupied. Further research is recommended, but results suggest that welfare assessments incorporating many diverse measures remain necessary to provide sufficient detail about dairy herd welfare.

Machine Learning to Detect Posture and Behavior in Dairy Cows: Information from an Accelerometer on the Animal's Left Flank

Simple Summary

This study analyzed the possibility of automatically detecting dairy cow behavior by combining the use of a single triaxial accelerometer applied to the animal’s left flank with a machine learning technique. This combination enabled the detection of posture and the main types of behavior that are extremely useful in evaluating the animal’s welfare and health such as resting, feeding, and rumination with a high degree of accuracy. The novelty of the study was the success in reaching a high accuracy in detecting five different behaviors and the animal posture by using a single sensor and allowing farmers to save money. To the best of our knowledge, this is the first study that has successfully explored the feasibility of locating a sensor on the animal’s left flank, showing the opportunity of automatically measuring some physiological parameters, such as those ones related to respiration and rumen health, in a non-invasive way.

Abstract

The aim of the present study was to develop a model to identify posture and behavior from data collected by a triaxial accelerometer located on the left flank of dairy cows and evaluate its accuracy and precision. Twelve Italian Red-and-White lactating cows were equipped with an accelerometer and observed on average for 136 ± 29 min per cow by two trained operators as a reference. The acceleration data were grouped in time windows of 8 s overlapping by 33.0%, for a total of 35,133 rows. For each row, 32 different features were extracted and used by machine learning algorithms for the classification of posture and behavior. To build up a predictive model, the dataset was split in training and testing datasets, characterized by 75.0 and 25.0% of the observations, respectively. Four algorithms were tested: Random Forest, K Nearest Neighbors, Extreme Boosting Algorithm (XGB), and Support Vector Machine. The XGB model showed the best accuracy (0.99) and Cohen’s kappa (0.99) in predicting posture, whereas the Random Forest model had the highest overall accuracy in predicting behaviors (0.76), showing a balanced accuracy from 0.96 for resting to 0.77 for moving. Overall, very accurate detection of the posture and resting behavior were achieved.

Quantitatively determining the somatic cell count of raw milk using dielectric spectra and support vector regression

To investigate the potential of dielectric spectroscopy in quantitatively determining the somatic cell count (SCC) of raw milk, the dielectric spectra of 301 raw milk samples at different SCC were collected using coaxial probe technology in the frequency range of 20 to 4,500 MHz. Standard normal variate, Mahalanobis distance, and joint x-y distances sample division were used to pretreat spectra, detect outliers, and divide samples, respectively. Principal component analysis and variable importance in projection (VIP) methods were used to reduce data dimension and select characteristic variables (CVR), respectively. The full spectra, 16 principal components obtained by principal component analysis, and 86 CVR selected by VIP were used as inputs, respectively, to establish different support vector regression models. The results showed that the nonlinear support vector regression models based on the full spectra and selected CVR using VIP had the best prediction performance, with the standard error of prediction and residual predictive deviation of 0.19 log SCC/mL and 2.37, respectively. The study provided a novel method for online or in situ detection of the SCC of raw milk in production, processing, and consumption.

Training and Validating a Machine Learning Model for the Sensor-Based Monitoring of Lying Behavior in Dairy Cows on Pasture and in the Barn

Simple Summary

There are various systems available for health monitoring and heat detection in dairy cows. By continuously monitoring different behavioral patterns (e.g., lying, ruminating, and feeding), these systems detect behavioral changes linked to health disorders and estrous. Most of the systems were developed for cows kept indoors, and only a few systems are available for pasture-based farms. The systems developed for the barn failed to detect the targeted behavior and thereby its changes on the pasture and vice versa. Therefore, our goal was to train and validate a machine learning model for the automated prediction of lying behavior in dairy cows kept on pastures, as well as indoors. Data collection was conducted on three dairy farms where cows were equipped with the collar-based prototype of the monitoring system and recorded with cameras in parallel. The derived dataset was used to develop the machine learning model. The model performed well in predicting lying behavior in dairy cows both on the pasture and in the barn. Therefore, the building of the model presents a successful first step towards the development of a monitoring system for dairy cows kept on pasture and in the barn.

Abstract

Monitoring systems assist farmers in monitoring the health of dairy cows by predicting behavioral patterns (e.g., lying) and their changes with machine learning models. However, the available systems were developed either for indoors or for pasture and fail to predict the behavior in other locations. Therefore, the goal of our study was to train and evaluate a model for the prediction of lying on a pasture and in the barn. On three farms, 7–11 dairy cows each were equipped with the prototype of the monitoring system containing an accelerometer, a magnetometer and a gyroscope. Video observations on the pasture and in the barn provided ground truth data. We used 34.5 h of datasets from pasture for training and 480.5 h from both locations for evaluating. In comparison, random forest, an orientation-independent feature set with 5 s windows without overlap, achieved the highest accuracy. Sensitivity, specificity and accuracy were 95.6%, 80.5% and 87.4%, respectively. Accuracy on the pasture (93.2%) exceeded accuracy in the barn (81.4%). Ruminating while standing was the most confused with lying. Out of individual lying bouts, 95.6 and 93.4% were identified on the pasture and in the barn, respectively. Adding a model for standing up events and lying down events could improve the prediction of lying in the barn.

Behavioral changes in dairy cows with metritis in seasonal calving pasture-based dairy system

The aims of this study were to determine changes in lying and rumination behavior of transition dairy cows that were healthy or affected by metritis during the autumn and spring calving season in a temperate climate and determine the effect of some weather variables (precipitation and heat stress) on these behaviors. A total of 103 multiparous cows were monitored during the autumn (February to April) and spring calving season (July to October) from 10 d before to 10 d after calving. Cows were chosen retrospectively by diagnosis of metritis [autumn season, (n = 11); spring season, (n = 13)] or as healthy [autumn season, (n = 25); spring season, (n = 25)] based on vaginal discharge characteristics evaluated during the first 10 days in milk. In all animals, electronic data loggers recorded lying (Hobo Pendant G Acceleration, Onset Computer Corp.) and rumination behavior (Hi-Tag rumination monitoring system, SCR Engineers Ltd.) during the study period. We included precipitation level (>1 mm/d = with rain, and ≤1 mm/d = without rain) and heat stress [no heat stress = temperature-humidity index (THI) < 68 vs. heat stress = THI ≥ 68] as weather factors that may have affected lying and rumination behavior during the spring and autumn season in a temperate climate, respectively. Metritis during the spring calving season was associated only with longer lying times (≥1.3 h/d) after calving. During the autumn calving season cows with metritis lay down longer the day of calving (~2.1 h/d) and had fewer lying bouts of longer duration during the prepartum period compared with healthy cows. Rumination time did not differ by health status during the spring calving season, whereas cows with metritis during autumn ruminated 30, 21, and 24 min/d less than healthy cows during the prepartum, calving, and postpartum period, respectively. Precipitation and heat stress were associated with decreased daily lying and rumination time in sick cows. Our results indicate that differences in lying and rumination behavior depended on the metritis status, and support the idea that weather factors such as rainfall or heat stress requires to be considered in analyses of transition cow behavior in seasonal calving pasture-based dairy systems.

Motivations and attitudes of Brazilian dairy farmers regarding the use of automated behaviour recording and analysis systems

In this Research Communication we investigate the motivations of Brazilian dairy farmers to adopt automated behaviour recording and analysis systems (ABRS) and their attitudes towards the alerts that are issued. Thirty-eight farmers participated in the study distributed into two groups, ABRS users (USERS, n = 16) and non-users (NON-USERS, n = 22). In the USERS group 16 farmers accepted being interviewed, answering a semi-structured interview conducted by telephone, and the answers were transcribed and codified. In the NON-USERS group, 22 farmers answered an online questionnaire. Descriptive analysis was applied to coded answers. Most farmers were young individuals under 40 years of age, with undergraduate or graduate degrees and having recently started their productive activities, after a family succession process. Herd size varied with an overall average of approximately 100 cows. Oestrus detection and cow's health monitoring were the main reasons given to invest in this technology, and cost was the most important factor that prevented farmers from purchasing ABRS. All farmers in USERS affirmed that they observed the target cows after receiving a health or an oestrus alert. Farmers believed that they were able to intervene in the evolution of the animals' health status, as the alerts gave a window of three to four days before the onset of clinical signs of diseases, anticipating the start of the treatment.The alerts issued by the monitoring systems helped farmers to reduce the number of cows to be observed and to identify pre-clinically sick and oestrous animals more easily. Difficulties in illness detection and lack of definite protocols impaired the decision making process and early treatment, albeit farmers believed ABRS improved the farm's routine and reproductive rates.

Transition cow metabolites and physical traits influence days to first postpartum ovulation in dairy cows

Physical activities are associated with the health of transition dairy cows and pregnancy outcomes are positively related to early resumption of postpartum estrous cycles. The objective was to assess key metabolites and patterns of prepartum and postpartum physical activity as they relate to the onset of first postpartum ovulation in lactating dairy cows. Close-up dry Holstein cows (n = 82) and late gestation heifers (n = 78) were enrolled beginning 3 wk before expected calving date (Day 0). Cows were fit with Cow SensOor ear tags to assess transitional changes in eating, resting, rumination, high activity, and ear-surface temperatures. Rectal temperatures were assessed and blood samples were collected on Days 0, 3, 7, and 14 to measure concentrations of glucose, free fatty acids (FFA), β-hydroxybutyrate (BHB), calcium, and haptoglobin. Body condition scores (BCS) and body weights (BW) were measured weekly, and blood samples were collected weekly from Day 21 ± 3 through 70 ± 3 to quantify changes in progesterone to detect luteal function after ovulation. Cows first ovulating before median Day 33 were classified as early (n = 76), whereas those first ovulating after Day 33 were classified as late (n = 84). Early ovulating cows first ovulated earlier (P < 0.001) than the late ovulation cows (24.3 ± 1.2 d [range: 16-32 d] vs. 48.8 ± 1.2 d [range: 33-74 d]), respectively. Mean days to first ovulation excluded seven cows that failed to ovulate before insemination. Compared with late ovulating cows, early ovulating cows had lesser (P < 0.05) concentrations of FFA, BHB, and haptoglobin on Days 0, 3, 7, and 14 in addition to having lesser (P < 0.05) rectal temperatures and ear-surface temperatures. Ear-surface temperatures began to decrease 4 d before parturition and remained less (P < 0.05) after calving than cows that subsequently ovulated late. Early ovulating cows tended (P = 0.07) to spend more time eating, and less (P = 0.02) time resting and being active during the first 3 wk after calving, and lost less (P = 0.03) BW and BCS during the first 9 wk compared with late ovulating cows. Although no differences were detected in yields of milk or energy-corrected milk during the first 9 wk after calving, early compared with late ovulating cows produced more (P < 0.01) milk protein. We concluded that metabolic measures during the first 2 wk after calving, and physical and behavioral traits are associated with the onset of postpartum ovarian activity.

Changes in social and feeding behaviors, activity, and salivary serum amyloid A in cows with subclinical mastitis

The aim of this study was to identify detailed changes in behavior, and in salivary serum amyloid A (SAA), associated with subclinical mastitis. This included standard sickness behaviors, such as decreased activity, feeding and drinking (here labeled "core maintenance" behaviors), and less well-studied social, grooming, and exploratory behaviors (here labeled "luxury" behaviors). Luxury behaviors are biologically predicted to change at lower levels of mastitis infection and are, therefore, particularly relevant to detecting subclinical mastitis. Salivary serum amyloid A is a physiological marker of systemic inflammation, with levels in milk and serum already known to increase during subclinical mastitis. We investigated whether the same was true for SAA in cow saliva. Data were collected for 17 matched pairs of commercial barn-housed Holstein-Friesian cows. Each pair comprised a cow with subclinical mastitis (SCM) and a healthy control (CTRL), identified using somatic cell count (SCC; SCM: SCC >200 × 1,000 cells/mL; CTRL: SCC <100 × 1,000 cells/mL). SCM cows were selected for study ad hoc, at which point they were paired with a CTRL cow, based upon parity and calving date; consequently, the full data set was accrued over several months. Data were collected for each pair over 3 d: SCC (d 1), behavior (d 2), salivary SAA (d 3). All behaviors performed by the focal cows over a single 24-h period were coded retrospectively from video footage, and differences between the SCM and CTRL groups were investigated using the main data set and a subset of data corresponding to the hour immediately following morning food delivery. Saliva was collected using cotton swabs and analyzed for SAA using commercially available ELISA kits. We report, for the first time, that an increase in salivary SAA occurs during subclinical mastitis; SAA was higher in SCM cows and demonstrated a positive (weak) correlation with SCC. The behavioral comparisons revealed that SCM cows displayed reductions in activity (behavioral transitions and distance moved), social exploration, social reactivity (here: likelihood to be displaced following receipt of agonism), performance of social grooming and head butts, and the receipt of agonistic noncontact challenges. In addition, SCM cows received more head swipes, and spent a greater proportion of time lying with their head on their flank than CTRL cows. The SCM cows also displayed an altered feeding pattern; they spent a greater proportion of feeding time in direct contact with 2 conspecifics, and a lower proportion of feeding time at self-locking feed barriers, than CTRL cows. Behavioral measures were found to correlate, albeit loosely, with serum SAA in a direction consistent with predictions for sickness behavior. These included positive correlations with lying duration and the receipt of all agonistic behavior, and negative correlations with feeding, drinking, the performance of all social and all agonistic behavior, and social reactivity. We conclude that changes in salivary SAA, social behavior, and activity offer potential in the detection of subclinical mastitis and recommend further investigation to substantiate and refine our findings.

Assessment of the Relationship between Postpartum Health and Mid-Lactation Performance, Behavior, and Feed Efficiency in Holstein Dairy Cows

The objective of this study was to investigate the relationships between postpartum health disorders and mid-lactation performance, feed efficiency, and sensor-derived behavioral traits. Multiparous cows (n = 179) were monitored for health disorders for 21 days postpartum and enrolled in a 45-day trial between 50 to 200 days in milk, wherein feed intake, milk yield and components, body weight, body condition score, and activity, lying, and feeding behaviors were recorded. Feed efficiency was measured as residual feed intake and the ratio of fat- or energy-corrected milk to dry matter intake. Cows were classified as either having hyperketonemia (HYK; n = 72) or not (n = 107) and grouped by frequency of postpartum health disorders: none (HLT; n = 94), one (DIS; n = 63), or ≥2 (DIS+; n = 22). Cows that were diagnosed with HYK had higher mid-lactation yields of fat- and energy-corrected milk. No differences in feed efficiency were detected between HYK or health status groups. Highly active mid-lactation time was higher in healthy animals, and rumination time was lower in ≥4th lactation cows compared with HYK or DIS and DIS+ cows. Differences in mid-lactation behaviors between HYK and health status groups may reflect the long-term impacts of health disorders. The lack of a relationship between postpartum health and mid-lactation feed efficiency indicates that health disorders do not have long-lasting impacts on feed efficiency.

Do we automatically detect health- or general welfare-related issues? A framework

The early detection of health disorders is a central goal in livestock production. Thus, a great demand for technologies enabling the automated detection of such issues exists. However, despite decades of research, precision livestock farming (PLF) technologies with sufficient accuracy and ready for implementation on commercial farms are rare. A central factor impeding technological development is likely the use of non-specific indicators for various issues. On commercial farms, where animals are exposed to changing environmental conditions, where they undergo different internal states and, most importantly, where they can be challenged by more than one issue at a time, such an approach leads inevitably to errors. To improve the accuracy of PLF technologies, the presented framework proposes a categorization of the aim of detection of issues related to general welfare, disease and distress and defined disease. Each decision level provides a different degree of information and therefore requires indicators varying in specificity. Based on these considerations, it becomes apparent that while most technologies aim to detect a defined health issue, they facilitate only the identification of issues related to general welfare. To achieve detection of specific issues, new indicators such as rhythmicity patterns of behaviour or physiological processes should be examined.

Variable selection for monitoring sickness behavior in lactating dairy cattle with the application of control charts

The present observational study investigated the application of multivariate cumulative sum (MCUSUM) control charts by including variables selected by principal component analysis and partial least squares (PLS) regression to identify sickness behavior in dairy cattle. Therefore, sensor information (24 variables) was collected from 480 milking cows on a German dairy farm between September 2018 and December 2019. These variables were gathered in potentially different scenarios on farm. In total, data from 749 animals were available for evaluation. Variables were chosen based on the information of 499 cows (62 healthy; 437 sick) with 93,598 observations. The available diagnoses were collected together to form 1,025 sickness events. Hence, the different numbers of selected variables were included into the MCUSUM control charts. The performance of the MCUSUM control charts was evaluated by a 10-fold cross validation; hence, 90% of the original data set (749 cows) represented the training data, and the remaining 10% was used to test the training results. On average, the 10 training data sets included 124,871 observations with 1,392 sickness events, and the 10 testing data sets included, on average, 13,704 observations with 153 sickness events. The MCUSUM generated from the variables selected by principal component analysis showed comparable results in training and testing in all scenarios; therefore, 70.0 to 97.4% of the sickness events were detected. The false-positive rates ranged from 8.5 to 29.6%, and thus they created at least 2.6 false-positive alerts per day in testing. The variables selected by the PLS regression approach showed comparable sickness detection rates (70.0-99.9%) as well as false-positive rates (8.2-62.8%) in most scenarios. The best performing scenario produced 2.5 false-positive alerts in testing. Summarizing, both approaches showed potential for practical implementation; however, the PLS variable selection approach showed fewer false positives. Therefore, the PLS regression approach could generate a more reliable sickness detection algorithm, if combined with MCUSUM control charts, and considered for practical implementation.

Rumination time around dry-off relative to the development of diseases in early-lactation cows

Monitoring rumination time (RT) around the time of calving is an effective way of identifying cows at risk of disease in early lactation. However, this only allows for the identification of cows a few days before the onset of clinical signs; thus, effective preventive measures cannot be implemented. Recent research has suggested that biomarkers of immune and metabolic function measured at dry-off (DO) can predict higher disease risk in early lactation. Nevertheless, the extent to which RT around DO is associated with early-lactation disease risk remains unexplored. Thus, the objective of this study was to compare RT in the weeks before and after DO between cows that did and did not experience health disorders in early lactation. For this, we conducted an observational retrospective cohort study utilizing the records available from a large commercial dairy herd in which RT is recorded daily using an automated system. Daily RT from -7 to +14 d relative to DO from 2,258 DO cycles and their respective health records in the first 60 d in milk were used. Differences in RT between animals with and without a disease history were tested with the Student t-test with Bonferroni adjustment. Mixed linear regression analyses were performed to assess differences in RT around DO and the association of RT with the occurrence of mastitis, metritis, retained placenta, hyperketonemia, lameness, hypocalcemia, pneumonia, and displaced abomasum. Rumination time decreased abruptly at DO and remained lower for 3 to 4 d compared with the days before DO. On average, cows affected by hyperketonemia and lameness ruminated 9.83 ± 6.40 and 15.00 ± 6.08 min/d less than unaffected cows, respectively. Cows that developed lameness in the first 60 d in milk showed reduced RT from 1 to 3 d following DO compared with cows that were not diagnosed with lameness in early lactation. However, RT around DO was not associated with the occurrence of the other health disorders studied here. Our results demonstrate that DO is a stressful event for dairy cows resulting in decreased RT for several days. Furthermore, the association between RT around DO and some early-lactation diseases suggests that RT could be a useful tool to identify at-risk cows early enough to allow for preventive interventions. Further studies should investigate the diagnostic utility of incorporating RT data early in the dry period in the disease prediction algorithms of rumination sensors.

Retrospective cohort study of management procedures associated with dairy herd-level eradication of Streptococcus agalactiae in the Danish surveillance program

The aim of this observational retrospective cohort study was to identify management procedures that are associated with herd-level eradication of Streptococcus agalactiae in dairy herds. The objective was to compare herds that recovered from Strep. agalactiae with herds that remained infected with Strep. agalactiae on the basis of specific management procedures. Data from the Danish surveillance program for Strep. agalactiae, where all milk delivering dairy herds are tested yearly, were used to identify study herds. One hundred ninety-six herds that were classified in the program as infected with Strep. agalactiae, in both January 2013 and January 2014, were identified as study herds. These were followed until January 2017. One hundred forty-four herds remained infected every year until January 2017. Forty-six herds recovered from Strep. agalactiae after January 2014 (were tested negative continuously after January 2015, January 2016, or January 2017 and remained noninfected in the program from recovery until January 2017). Herd characteristics and management procedures were obtained through the Danish Cattle Database. Herd characteristics included herd size, yield, milking system, and bulk milk somatic cell count (SCC). Management procedures included the proportion of cows culled within 100 d after calving due to mastitis, the extent of diagnoses relative to the extent of mastitis treatments, the proportion of cows treated for mastitis during lactation, the proportion of cows treated for mastitis early in lactation, the proportion of cows treated at dry-off, and the median length of the dry period for cows receiving dry cow treatment. All variables were calculated on herd level. Multivariable logistic regression was used to analyze the association between herd infection status and management procedures. A higher proportion of culling due to mastitis within 100 d from calving was associated with a higher probability of herd-level recovery from Strep. agalactiae in herds with conventional milking system. For example, herds with conventional milking, a bulk milk SCC of 260,000 cells/mL, and 10% early culling due to mastitis had a recovery probability of 0.13, whereas similar herds with 20% early culling due to mastitis had a recovery probability of 0.15. A higher proportion of mastitis treatments within 250 d postcalving was associated with a higher probability of herd-level recovery for herds with a relatively high bulk milk SCC. For example, herds with conventional milking, a bulk milk SCC of 260,000 cells/mL, and 10% lactational mastitis treatments had a recovery probability of 0.12, whereas similar herds with 20% lactational mastitis treatments had a recovery probability of 0.15. Herds with a low bulk milk SCC (<220,000 cells/mL) combined with a low proportion of lactational treatments (<0.2) had a relatively high probability of herd-level recovery (>0.2). Additional variables, including the proportion of dry cow treatments, were not associated with herd-level recovery from Strep. agalactiae.

Prepartum behavior changes in dry Holstein cows at risk of postpartum diseases

The objective of this study was to identify changes in prepartum behavior associated with the incidence of postpartum diseases in dairy cows. Multiparous Holstein cows (n = 489) were monitored with accelerometers for 3 wk prepartum. Accelerometers measured steps, time at the feed bunk, frequency of meals, lying bouts, and lying time. Postpartum health was monitored from 0 to 30 d in milk and cases of metritis, mastitis, retained placenta, displaced abomasum (DA), ketosis, and hypocalcemia were recorded. A multivariate linear mixed model was used to assess differences in behavior between diseased and not diagnosed diseased cows. A multivariate logistic regression was used to predict the occurrence of diseases. Predictors were selected using a manual backward stepwise selection process of variables until all remaining predictors had a P < 0.10. Models were submitted to a leave-one-out cross-validation process, and sensitivity, specificity, false discovery rate, and false omission rate were calculated. On average, over the 3-wk prepartum period, cows not diagnosed diseased (n = 345) took 1,613 ± 38 steps, spent 181 ± 7.1 min at the feed bunk, had 8.3 ± 0.17 meals, had 9.8 ± 0.32 lying bouts, and spent 742 ± 11.3 min lying per day. Behavior of diseased cows (n = 144) did not differ from those not diagnosed diseased. However, differences for specific diseases were observed, being significant in the week prepartum. When considering changes in behavior for only the week before calving, cows with metritis had more lying bouts (+21%), cows with DA had fewer meals (-24%) and tended to take fewer steps (-18%), and cows with ketosis had fewer meals (-22%) and spent less time at the feed bunk (-40%). Prediction models with the best outcomes were found for DA and ketosis using data of the prepartum week only. The model for DA included time at the feed bunk. Cross-validation resulted in a 80% sensitivity, 58.1% specificity, 59.2% accuracy, 91.2% false discovery rate, and 1.7% false omission rate. The model for ketosis included time at the feed bunk and number of meals. Cross-validation resulted in 64.3% sensitivity, 59.3% specificity, 59.5% accuracy, 93.0% false discovery rate, and 2.8% false omission rate. Prepartum behavior of cows affected with metritis, DA, and ketosis was different from that of cows not diagnosed with diseases. Prediction equations were able to classify cows at high or low risk of ketosis and DA and can be used in taking management decisions, but the high false discovery rates requires further refinement.

Effects of adding an automated monitoring device to the health screening of postpartum Holstein cows on survival and productive and reproductive performances

Automated monitoring devices (AMD) have become more affordable, and consequently more popular among dairy producers. We hypothesized that the addition of AMD-generated health alerts to a health-screening program improves survival, milk production, and reproductive success. In addition, we hypothesized that cows diagnosed with clinical disease that have AMD alerts are at greater risk of culling, lower milk production, and decreased risk of pregnancy than cows without AMD alerts. Holstein cows (nulliparous = 282, parous = 328) were enrolled at -60 ± 3 d (d 0 = calving), when they were fitted with an AMD and assigned randomly to 1 of 2 health-screening strategies: (1) control: AMD alerts not provided to farm personnel; and (2) automated device: AMD alerts provided to farm personnel. Twice daily, study personnel determined which cows had AMD alerts (health index ≤79, rumination <200 min/d, or difference between current rumination and the average of the 3 preceding days <0) and provided the information to farm personnel. Farm personnel examined cows at 3, 5, and 9 d in milk (DIM) and when daily milk yield decreased ≥25% on consecutive days. We detected no differences between health-screening strategies regarding morbidity (control = 49.7 ± 3.3%, automated device = 52.8 ± 3.2%), but the interaction between health-screening strategy and parity tended to be associated with the number of clinical diseases per cow (primiparous: control = 0.46 ± 0.06, automated device = 0.65 ± 0.07 cases/cow; multiparous: 0.88 ± 0.08, automated device = 0.86 ± 0.08 cases/cow). Cows enrolled in the automated device strategy were more likely to be treated with supportive therapy (64.4 ± 3.1 vs. 55.0 ± 3.2%), whereas primiparous cows in the automated device strategy were more likely to be treated with anti-inflammatory drugs than those in the control strategy (41.6 ± 4.7 vs. 23.8 ± 4.0%). Health-screening strategy did not affect survival or total milk yield up to 22 wk postpartum, but cows in the automated device strategy had reduced risk of pregnancy after the first 2 services (54.5 ± 3.0 vs. 46.2 ± 3.2%). Cows diagnosed with a clinical disease without AMD alerts had reduced risk of removal from the herd by 150 DIM (5.7 ± 2.0 vs. 19.0 ± 3.3%), greater risk of pregnancy after the first 2 services (49.6 ± 4.5 vs. 33.6 ± 3.9%), and greater milk by 22 wk postpartum (6.7 ± 0.2 vs. 5.3 ± 0.2 × 10³ kg) than cows diagnosed with a clinical disease that had an AMD alert. Adding AMD-generated health alerts to the health screening of postpartum cows in a herd with an existing screening program did not improve survival, milk yield, or reproductive success. In addition, AMD alerts in cows diagnosed with a clinical disease may be indicative of the future success of such cows.

Accelerometer systems as tools for health and welfare assessment in cattle and pigs - A review

Welfare assessment has traditionally been performed by direct observation by humans, providing information at only selected points in time. Recently, this assessment method has been questioned, as 'Precision Livestock Farming' technologies may be able to deliver more valid, reliable and feasible real-time data at the individual level and serve as early monitoring systems for animal welfare. The aim of this paper is to describe how accelerometers can be used for welfare assessment based on the principles of the Welfare Quality assessment protocol. Algorithm development is based mainly on the detection of behavioural traits. So far, high accuracies have been found for movement and resting behaviours in cows and pigs, while algorithm development for feeding and drinking behaviours in pigs lag behind progress in cows where valid algorithms are already available. Welfare studies have used accelerometer technology to address the effects on behaviour of diet, daily cycle, enrichment, housing, social mixing, oestrus, lameness and disease. Additional aspects to consider before a decision is made upon its use in research and in practical applications include battery life and sensor location. While accelerometer systems for cows are already being used by farmers, application in pigs has mainly remained at the research level.

A survey on sensor systems used in Italian dairy farms and comparison between performances of similar herds equipped or not equipped with sensors

Sensor systems (SS) were developed over the last few decades to help dairy farmers manage their herds. Such systems can provide both data and alerts to several productive, behavioral, and physiological indicators on individual cows. Currently, there is still a lack of knowledge on both the proportion of dairy farms that invested in SS and type of SS installed. Additionally, it is still unclear whether the performances of herds equipped with SS differ from those of similar herds managed without any technological aid. Therefore, the aims of this study were (1) to provide an insight into SS spread among Italian dairy farms and (2) to analyze the performances of similar herds equipped or not equipped with SS. To reach the former goal, a large survey was carried out on 964 dairy farms in the northeast of Italy. Farmers were interviewed by the technicians of the regional breeders association to collect information on the type of SS installed on farms and the main parameters recorded. Overall, 42% of the surveyed farms had at least 1 SS, and most of them (72%) reared more than 50 cows. Sensors for measuring individual cow milk yield were the most prevalent type installed (39% of the surveyed farms), whereas only 15% of farms had SS for estrus detection. More sophisticated parameters, such as rumination, were automatically monitored in less than 5% of the farms. To reach the latter goal of the study, a subset of 100 Holstein dairy farms with similar characteristics was selected: half of them were equipped with SS for monitoring at least individual milk yield and estrus, and the other half were managed without any SS. Average herd productive and reproductive data from official test days over 3 yr were analyzed. The outcomes of the comparison showed that farms with SS had higher mature-equivalent milk production. Further clustering analysis of the same 100 farms partitioned them into 3 clusters based on herd productive and reproductive data. Results of the Chi-squared test showed that the proportion of farms equipped with SS was greater in the cluster with the best performance (e.g., higher milk yield and shorter calving interval). However, the presence of a few farms equipped with SS in the least productive cluster for the same parameters pointed out that although the installation of SS may support farmers in time- and labor-saving or in data recording, it is not a guarantee of better herd performance.

Transition dairy cow health is associated with first postpartum ovulation risk, metabolic status, milk production, rumination, and physical activity

Our objective was to determine the association of health status during the first 60 d in milk (DIM) and first postpartum ovulation risk, physical activities recorded by an activity monitor, and metabolic and milk measures in Holstein cows. Late-gestation heifers and close-up dry cows in 1 herd fitted with CowManager SensOors (Agis, Harmelen, the Netherlands) were enrolled in the study 3 wk before expected parturition to assess ear skin temperature and daily rumination, eating, inactivity, and activity times. Blood samples were collected at calving (d 0), and on d 3, 7, and 14 to assess concentrations of free fatty acids, β-hydroxybutyrate (BHB), calcium, glucose, and haptoglobin. In addition, weekly measures were conducted for body condition, body weight, and progesterone through 63 ± 3 DIM when ovulation was synchronized (GnRH-1 - 7 d - PGF_2α - 24 h - PGF_2α - 32 h - GnRH-2 - 16 h - artificial insemination). Disease diagnosed in 68 of 160 cows (42.5%) was distributed equally between primiparous (48.5%) and multiparous (51.5%) cows. Cows were classified as diseased when any case of metritis, digestive disorders, ketosis, hypocalcemia, calving problems, mastitis, or lameness occurred during the first 60 DIM. Odds of early ovulation by median postpartum d 33 was 1.92 times greater in healthy than diseased cows. Incidence of individual diseases included metritis (18.8%), digestive disorders (17.5%), ketosis (BHB >10 mg/dL; 11.9%), hypocalcemia (Ca <2.2 mmol/L; 10.6%), calving problems (6.3%), mastitis (3.1%), and lameness (3.1%). Odds of early ovulation were 2.48, 2.65, and 5.72 times greater in healthy cows compared with cows diagnosed with metritis, digestive disorders, or ketosis, respectively. Diseased compared with healthy cows had greater concentrations of free fatty acids, BHB, haptoglobin, greater rectal temperature, and lesser concentration of serum calcium on d 0, 3, 7, and 14 than healthy cows. Plasma glucose was not affected by health status, but was lesser in concentration on d 3, 7, and 14 compared with day of calving. Weekly (calving through 9 wk) body condition scores tended to be and weekly body weights were greater in healthy compared with diseased cows. Activity measures differed by health status during prepartum (d -14 through -1) and postpartum (d 0 through 20) periods except for eating time. Healthy cows spent less time being inactive during both periods compared with diseased cows and had greater postpartum rumination times than diseased cows. Mean daily milk yield during the first 14 wk in milk was greater in healthy than diseased cows by 2.1 ± 0.8 kg. We conclude that disease negatively affects early postpartum ovulation risk and is associated with measurable changes in periparturient physical activity and postpartum metabolic profiles.

Using Sensor Data to Detect Lameness and Mastitis Treatment Events in Dairy Cows: A Comparison of Classification Models

The aim of this study was to develop classification models for mastitis and lameness treatments in Holstein dairy cows as the target variables based on continuous data from herd management software with modern machine learning methods. Data was collected over a period of 40 months from a total of 167 different cows with daily individual sensor information containing milking parameters, pedometer activity, feed and water intake, and body weight (in the form of differently aggregated data) as well as the entered treatment data. To identify the most important predictors for mastitis and lameness treatments, respectively, Random Forest feature importance, Pearson’s correlation and sequential forward feature selection were applied. With the selected predictors, various machine learning models such as Logistic Regression (LR), Support Vector Machine (SVM), K-nearest neighbors (KNN), Gaussian Naïve Bayes (GNB), Extra Trees Classifier (ET) and different ensemble methods such as Random Forest (RF) were trained. Their performance was compared using the receiver operator characteristic (ROC) area-under-curve (AUC), as well as sensitivity, block sensitivity and specificity. In addition, sampling methods were compared: Over- and undersampling as compensation for the expected unbalanced training data had a high impact on the ratio of sensitivity and specificity in the classification of the test data, but with regard to AUC, random oversampling and SMOTE (Synthetic Minority Over-sampling) even showed significantly lower values than with non-sampled data. The best model, ET, obtained a mean AUC of 0.79 for mastitis and 0.71 for lameness, respectively, based on testing data from practical conditions and is recommended by us for this type of data, but GNB, LR and RF were only marginally worse, and random oversampling and SMOTE even showed significantly lower values than without sampling. We recommend the use of these models as a benchmark for similar self-learning classification tasks. The classification models presented here retain their interpretability with the ability to present feature importances to the farmer in contrast to the “black box” models of Deep Learning methods.

Effects of postpartum acetylsalicylic acid on metabolic status, health, and production in lactating dairy cattle

The transition period is one of the most challenging times for dairy cattle. Previous research suggests that treatment of postpartum cows with anti-inflammatory drugs may decrease pain and inflammation, enhancing cow welfare and performance during this challenging period. However, these strategies involve numerous time-consuming interventions, which require extra labor and do not fit modern farm logistics. The objective of this experiment was to assess the effects of acetylsalicylic acid (ASA) every 24 h for 2 d after calving on (1) daily milk yield, daily milk conductivity, and daily rumination during the first 60 days in milk (DIM), and 305-d mature-equivalent milk, milk fat, and milk protein yields, (2) body condition score, β-hydroxybutyrate (BHB), and haptoglobin, and (3) incidence of clinical diseases during the first 60 DIM. Dairy cows (n = 246) from a dairy farm located in Pennsylvania were enrolled in this experiment. Cows were blocked by parity and assigned randomly to 1 of 2 treatments: (1) ASA (n = 121), in which cows received 2 treatments with ASA (200 mg/kg; 4 boluses), the first within 12 h after parturition and the second 24 h later; or (2) untreated (UNT; n = 125), in which cows remained untreated. Blood samples were collected at 30 ± 6 h, 7 ± 3 d, and 14 ± 3 d after calving to measure BHB and haptoglobin concentrations. Body condition score was assessed at enrollment, 7 ± 3 DIM, 14 ± 3 DIM, and 50 ± 10 DIM. Furthermore, incidences of diseases, daily rumination, daily milk yield, and daily milk conductivity during the first 60 DIM and 305-d mature-equivalent milk, milk fat, and milk protein yields were collected from on-farm computer records. The data were analyzed using mixed multiple linear and logistic regression models as a randomized complete block design. Multiparous cows treated with ASA produced 1.64 kg/d more milk compared with multiparous cows that remained untreated (ASA = 41.66 ± 0.88 kg/d; UNT = 40.02 ± 0.81 kg/d) during the first 60 DIM. There was no difference in daily milk conductivity and rumination between treatments. Cows treated with ASA had lower concentration of BHB (ASA = 1.16 ± 0.64 mmol/L; UNT = 1.23 ± 0.80 mmol/L) during the first 14 ± 3 DIM and had higher body condition score within the first 50 ± 10 DIM compared with cows that remained UNT. There were no differences in circulating concentrations of haptoglobin between treatments. These results support previous findings showing that the use of anti-inflammatory drugs after calving may increase milk production and affect the metabolic status of dairy cows.

Identifying gram-negative and gram-positive clinical mastitis using daily milk component and behavioral sensor data

Opportunities exist for automated animal health monitoring and early detection of diseases such as mastitis with greater on-farm adoption of precision technologies. Our objective was to evaluate time series changes in individual milk component or behavioral variables for all clinical mastitis (CM) cases (ACM), for CM caused by gram-negative (GN) or gram-positive (GP) pathogens, or CM cases in which no pathogen was isolated (NPI). We developed algorithms using a combination of milk and activity parameters for predicting each of these infection types. Milk and activity data were collated for the 14 d preceding a CM event (n = 170) and for controls (n = 166) matched for breed, parity, and days in milk. Explanatory variables in the univariate and multiple regression models were the slope change in milk (milk yield, conductivity, somatic cell count, lactose percentage, protein percentage, and fat percentage) and activity parameters (steps, lying time, lying bout duration, and number of lying bouts) over 7 d. Slopes were estimated using linear regression between d -7 and -5, d -7 and -4, d -7 and -3, d -7 and -2, and d -7 and -1 relative to CM detection for all parameters. Univariate analyses determined significant slope ranges for explanatory variables against the 4 responses: ACM, GN, GP, and NPI. Next, all slope ranges were offered into the multivariate models for the same 4 responses using 3 baselines: d -10, -7, and -3 relative to CM detection. In the univariate analysis, no explanatory variables were significant indicators of ACM, whereas at least 1 parameter was significant for each of GN, GP, and NPI models. Superior sensitivity (Se) and specificity (Sp) estimates were observed for the best GP (Se = 82%, Sp = 87%) and NPI (Se = 80%, Sp = 94%) multiple regression models compared with the best ACM (Se = 73%, Sp = 75%) and GN (Se = 71%, Sp = 74%) models. Sensitivity for the GN model was greater at the baseline closest to the day of CM detection (d -3), whereas the opposite was observed for the GP and NPI model as Se was maximized at the d -10 baseline. Based on this screening of relationships, milk and activity sensor data could be used in CM detection systems.

Associations between motion activity, ketosis risk and estrus behavior in dairy cattle

Ketosis (acetonaemia) is a metabolic disorder that occurs in cattle when energy demand exceeds energy intake and results in a negative energy balance. The course of the disease often starts with a subclinical phase, so early detection is crucial for decisive strategies. The aim of this study was to determine whether daily motion activity could be used as an indicator of subclinical ketosis in early lactation and to evaluatethe effect of subclinical ketosis on activity at estrus. The study was carried out on a 75-cow dairy farm over 6 months. Data were collected from 48 cows between day 0 and day 70 post-partum. Beta-Hydroxybutyrate (BHB) concentrations were evaluated in milk samples using rapid on-site ketosis tests. A test was considered positive at a concentration of >100 μmol/l. The animals were divided into two groups: group 'Healthy' (H) and group 'Ketosis' (K). Once the on-site test was positive, the cows were assigned to group K. Progesterone concentrations were evaluated in milk by photometric detection of the colour reaction of a competitive, heterologous enzyme immunoassay (EIA). Each drop from ≥0.3 ng/ml to <0.3 ng/ml with a subsequent increase to ≥0.3 ng/ml was considered estrus. Daily milk yield, concentrate intake and motion activity were recorded from a computerized dairy management system with the associated software (DairyPlan C21). Animals in group K had lower average daily activity levels than animals in group H. In this study, statistically significant reduced motion activity in animals in group K was observed on days 6-12 post-partum (P < 0.001, χ² test) compared with the herd mean daily motion activity. Furthermore, a significant association could be found between motion activity and group affiliation (logistic regression models). The sensitivity of the detection of cows at risk for ketosis was 81.8 %, and the specificity was 65.4 %, retrospectively determined by their activity behaviour. The mean motion activity on the day of estrus was significantly (P < 0.05) lower in animals in group K than in those in group H. This method may help to establish a future early warning system for the risk of ketosis in dairy cows. Thus, cows at risk may be identified for further targeted diagnostics and for selective treatment procedures. This study confirms the already reported lasting effect of subclinical ketosis on reproductive efficiency.

Alterations in sick dairy cows' daily behavioural patterns

The recent development of dairy production is characterised by increasing herd sizes and therefore increasingly complicated visual observation of cow behaviour, which is traditionally the basis for diagnoses of production diseases. The limitation of the direct visual behavioural observation due to the increasing number of individual cows implies a growing need for an automated detection of changes within behavioural patterns to identify cows that show sickness behaviour. Sensor systems can be used to measure behavioural patterns such as activity, resting, feeding and rumination. Behavioural patterns change with the occurrence of sickness but also interact with external factors. Changes such as prolonged lying duration or shortened feeding duration caused by metabolic disorders or infections, respectively, then serve as a detection tool for sick individuals. The aim of the present review is to outline the impact of production diseases on the daily behavioural patterns of dairy cows by referring to sickness behaviour.

On-farm use of disease alerts generated by precision dairy technology

Wearable precision dairy monitoring (PDM) technologies currently used to detect estrus may provide insight into disease detection. However, the incorporation of PDM into farm management and its perceived usefulness for dairy producers have not been explored. As the targeted end users of these products, information is needed on how producers use generated disease alerts as well as barriers to adoption and usefulness. The objective of this research was to assess the perceived usefulness producers attributed to alerts from a daily generated alert list designed to identify sick or injured cows or cows that experienced a major management change. Data from 1,171 cows on 4 commercial farms in Kentucky were collected from October 2015 to October 2016. Each cow was equipped with 2 PDM technologies: a leg tag (measuring activity in steps/d and lying time in h/d) and a neck collar (measuring eating time in h/d). Alerts were generated based on an individual cow's decrease of ≥30% in activity, lying, and eating time compared with each cow's 10-d moving mean. Producers sorted alerts into 3 overall categories: (1) the cow alert was perceived to be true and the cow was visually checked, (2) the cow alert was perceived to be true, but the cow was not visually checked, and (3) the cow alert behavior change was doubted by the producer and the cow was not visually checked. Further subdivisions were also provided to explain the choice for an overall category. Over the year, 24,012 cow alerts were generated (eating time, n = 9,543; lying time, n = 9,777; activity, n = 1,590; or a combination of behaviors, n = 3,102). Only 8% of the alerts were doubted by the producer. Although 55% of alerts were perceived to be true, producers visually assessed cows based on only 21% of the alerts with a large variation between farms (2 to 45% of the alerts visually assessed). Producers were more likely to completely ignore alerts over time. Producers were more likely to perceive cow alerts to be true and visually check cows when ≤20 alerts occurred per day, cows were fresh or in early lactation, alerts occurred during the work week, or when cow alerts were for eating time, activity, or a combination of multiple behaviors. Behavioral disease alerts must be improved and correspond to an actionable change for producers to use them. Incorporating herd management software, creating and managing alerts by lactation stage, and focusing on behaviors producers already find useful could improve future alert utilization.