Search

Online somatic cell count (SCC) measurement is widely used in dairy herds milked with automatic milking systems (AMS) and gives the opportunity to closely monitor individual cow udder health. Using automated SCC data, we observed cows displaying a remarkably regularly fluctuating SCC (rfSCC) pattern, which is described in this study. We aimed to (1) estimate the prevalence of rfSCC in cows milked by AMS, (2) characterize the rfSCC pattern, and (3) identify factors potentially associated with the rfSCC pattern. We analyzed 30-d episodes of composite SCC recordings of 1,000 cows from 55 dairy herds from 6 countries using an AMS with automated SCC measurement, and we identified the rfSCC pattern in 4.7% (95% CI: 3.5-6.2%) of these episodes. The rfSCC episodes had a median SCC of 701 × 1,000 cells/mL (2.5-97.5% quantile: 539-1,162), a median amplitude of 552 × 1,000 cells/mL (2.5-97.5% quantile: 409-886), and a median cycle length of 4.1 d (2.5-97.5% quantile: 3.7-4.9). Bacteriological culture data from quarter-milk samples collected every 2 wk in 1 Dutch AMS herd were analyzed, yielding no clear association between pathogen species and the rfSCC pattern found in that herd. Altogether, we described an intriguing phenomenon, present in almost 5% of the cows during a 1-mo study period. Further work is needed to quantify its importance in terms of udder health, but also to elucidate the mechanism behind this remarkable SCC pattern.

Automatic milking systems (AMS) are installed on a growing number of dairy farms worldwide. Management to support good udder health might be different on farms with an AMS compared with farms milking with a conventional milking system, as risk factors for mastitis on farms using an AMS may differ. The aim of this study was to identify farm level factors associated with mastitis on Dutch dairy farms using an AMS. In 2008, risk factor data were collected using a questionnaire combined with on-farm recordings of cow, stall, and AMS hygiene on 135 farms. These risk factor data were linked to 4 udder-health-associated dependent variables: average herd somatic cell count (HeSCCav), variance of the average herd somatic cell count (SCC) on test days (HeSCCvar), the average proportion of new high SCC cases (NHiSCC), and the farmer-reported annual incidence rate of clinical mastitis (IRCM). We employed regression models using multiple imputation to deal with missing values. Due to the high dimensionality of the risk factor data, we also performed nonlinear principal component analysis (NLPCA) and regressed the dependent variables on the principal components (PC). Good hygiene of cows and of AMS were found to be related to a lower HeSCCav and less NHiSCC. Effective postmilking teat disinfection was associated with a lower NHiSCC. A higher bulk tank milk SCC threshold for farmers' action was related to more NHiSCC. Larger farm size was related to lower HeSCCvar but higher NHiSCC. Negative attitude of farmers to animal health, higher frequency of checking AMS, and more time spent on viewing computer data were all positively related to higher IRCM. An NLPCA with 3 PC explained 16.3% of the variance in the risk factor variables. Only the first 2 PC were associated with mastitis. The first PC reflected older and larger farms with poor cow hygiene and AMS hygiene, and was related to higher HeSCCav and NHiSCC, whereas the second PC reflected newly built smaller farms with poor cow hygiene and low milk production, and was associated with higher HeSCCvar and NHiSCC, but lower IRCM. Our study suggests that many of the risk factors on conventional milking system farms are applicable to AMS farms, specifically concerning hygiene of the cows and the milking machine, but on large AMS farms, udder health may need more attention than on smaller AMS farms. Multiple imputation is instrumental to deal with missing values and NLPCA is a useful technique to process high dimensional data in our study.

The goal of dry-cow therapy (DCT) is to reduce the prevalence of intramammary infections (IMI) by eliminating existing IMI at drying off and preventing new IMI from occurring during the dry period. Due to public health concerns, however, preventive use of antimicrobials has become questionable. In this study, we evaluated the effects of 8 scenarios for selecting animals for DCT, taking into account variation in parity and cow-level somatic cell count (SCC) at drying off. The aim of this study was to evaluate udder health, antimicrobial usage, and economics at the herd level when using different scenarios for selecting cows for DCT. To enable calculation and comparison of the effects of different scenarios to select cows for DCT in an "average" herd, we created an example herd, with a virtual herd size of 100 dairy cows to be calving during a year. Udder health, antimicrobial usage, and economics were evaluated during the dry period and the first 100 d in lactation, the period during which the greatest effect of DCT is expected. This leads to an estimated 13,551 cow-days at risk during a year in a 100-cow dairy herd. In addition to a blanket DCT (BDCT) scenario, we developed 7 scenarios to select cows for DCT based on SCC. The scenarios covered a range of possible approaches to select low-SCC cows for DCT, all based on cow-level SCC thresholds on the last milk recording before drying off. The incidence rate of clinical mastitis in the example herd varied from 11.6 to 14.5 cases of clinical mastitis per 10,000 cow-days at risk in the different scenarios, and the prevalence of subclinical mastitis varied from 38.8% in scenario 1 (BDCT) to 48.3% in scenario 8. Total antimicrobial usage for DCT and clinical mastitis treatment varied over the scenarios from 1.27 (scenario 8) to 3.15 animal daily dosages (BDCT), leading to a maximum reduction in antimicrobial usage of 60% for scenario 8 compared with BDCT. The total costs for each of the scenarios showed little variation, varying from €4,893 for scenario 5 to €5,383 for scenario 8. The effect of selective DCT compared with BDCT on udder health, antimicrobial usage, and economics is influenced by the SCC criteria used to select cows for DCT. Scenario 2 resulted in the lowest increases in clinical and subclinical mastitis compared with BDCT. The greatest reduction in antimicrobial usage was achieved under scenario 8. From an economic perspective, lowest costs were achieved with scenario 5. Drying off dairy cows with antimicrobials has an effect on udder health, antimicrobial usage, and economics.

The objective of this study was to develop a mathematical model to identify a scenario with the lowest costs for mastitis associated with the dry period while restricting the percentage of cows to be dried off with dry cow antimicrobials. Costs of clinical and subclinical mastitis as well as antimicrobial use were quantified. Based on data from a large field trial, a linear programming model was built with the goal to minimize the costs associated with antimicrobial use at drying off. To enable calculations on minimizing costs of dry cow treatment on herd-level by drying-off decisions in an “average” herd, we created an example herd. Cows were projected on 3 different types of herds, based on bulk tank somatic cell count, and were categorized in groups based on parity and somatic cell count from the last test recording before drying-off. Economically optimal use of antimicrobials was determined while restricting the maximum percentage of cows dried off with antimicrobials from 100 to 0%. This restriction reveals the relationship between the maximum percentage of cows dried off with antibiotics and the economic consequences. A sensitivity analysis was performed to evaluate the effect of variation in the most important input variables, with the effect of dry cow antimicrobials resulting in a lower or higher percentage of clinical and subclinical mastitis depending on being dried off with or without dry cow antimicrobials, respectively, and the milk price. From an economic perspective, blanket dry cow treatment seems not to be the optimal approach of dry cow therapy, although differences between approaches were small. With lower bulk tank somatic cell counts, more dry cow antimicrobials can be omitted without economic consequences. The economic impact of reducing the percentage of clinical mastitis was found to be much larger than reducing the bulk tank somatic cell count. The optimal percentage of cows to be dried off with antimicrobials depends on the udder health situation, expressed as the bulk tank somatic cell count and the incidence of clinical mastitis. For all evaluated types of herds, selective dry cow treatment was economically more beneficial than blanket dry cow treatment. Economic profits of selective dry cow treatment are greater if bulk tank somatic cell count and clinical mastitis incidence are lower. Economics is not an argument against reduction of dry cow antimicrobials by applying selective dry cow treatment.

A large part of the world's resources are used to produce animal products. Efficient use of these resources is important to improve social well-being. Endemic animal diseases decrease production efficiency, because they require a higher level of input to produce the same amount of output or result in a lower output with the same amount of input. The optimal level of production with and without disease differs from farm to farm and depends on varying economic circumstances. Given these difficulties, making an accurate theoretical estimation of the economic impact of endemic diseases is challenging. Current approaches towards the economic assessment of endemic diseases are, therefore, quite pragmatic. For on-farm decision-making, the total costs consist of failure costs and preventive costs. Failure costs are associated with production losses (i.e. decreases in milk production, mortality and culling), treatment costs (i.e. veterinary treatment, drugs, and discarded milk) and the use of other resources associated with the occurrence of disease (i.e. increased labour costs). Preventive costs are associated with preventive measures in terms of equipment, consumables (e.g. diagnostics and chemicals) and the use of other resources to prevent diseases (i.e. increased labour). There is a substitution relationship between failure costs and preventive costs. That means that, in order to maximise profit at the farm level, the amount of resources invested in prevention should be chosen in such a way that total costs are minimised. The most studied endemic disease in animal production is mastitis. Most publications on mastitis only assess failure costs, and studies on assessing the total costs and best methods to determine an optimal level of prevention are scarce. Future challenges lie in researching frameworks that can assist decision-makers to establish optimal prevention levels for endemic diseases

The aim of this study was to assess farmers’ preferences for the performance characteristics of mastitis detection systems. Additionally, we looked at whether certain groups of farmers could be distinguished with specific preferences. Farmers’ opinions concerning mastitis detection systems, as well as general information about the farm and the farmer, were investigated with a standard questionnaire. The second part of the questionnaire was specifically aimed at elucidating preferences. Definitions of time windows and performance parameters, such as sensitivity and specificity, were incorporated into characteristics of a detection system (attributes) that reflect farmers’ daily experience. Based on data from 139 farmers, we concluded that, on average, they prefer a clinical mastitis detection system that produces a low number of false alerts, while alerting in good time and with emphasis on the more severe cases. These 3 attributes were evaluated as more important than the 3 other attributes, representing the costs of the detection system, the number of missed cases, and how long before clinical signs alerts need to be given. Variation in importance per attribute, however, was high, denoting that farmers’ preferences differ considerably. Although some significant relationships were found between farm characteristics and attributes, no clear groups of farmers with specific preferences could be distinguished. Based on these results, we advise making detection systems adaptable for the farmers to satisfy their preferences and to match the circumstances on the farm. Furthermore, these results support that for evaluation of detection algorithms comparisons have to be made at high levels of specificity (e.g., 99%) and time windows have to be kept small (preferably no more than 24 h). Key words: adaptive conjoint analysis, automatic milking system, farmer preference, mastitis detection

Due to the increasing use of automated milking systems, automated detection of clinical mastitis is becoming more important. Various in- or on-line diagnostic tests are in use, but generally suffer from false mastitis alerts. In this study, we explored a new diagnostic approach based on measurement of protease activity using fluorogenic protease substrates, which can be performed on site, at high speed, and at low costs. Samples from cows with clinical mastitis submitted for bacteriological culture at the University Farm Animal Practice were collected during several months and kept at -20°C until protease activity measurement. A reference set of milk samples from clinically healthy cows were collected on 9 different farms and were tested for protease activity directly and after freezing at -20°C to allow for comparison with the samples from clinical cases. The protease activity in mastitic milk samples was significantly higher than in samples from healthy animals. Based on 71 clinical mastitis samples and 180 milk samples from clinically healthy quarters, the area under the receiver-operating characteristic curve was estimated to be between 0.88 and 0.90, and at a threshold of 38 fluorescence per minute the test had a specificity of 0.99 at a sensitivity of 0.58. Protease activity measured in fresh milk from clinically healthy cows was significantly associated with somatic cell count and parity, but not with electrical conductivity, whereas protease activity in milk that had been frozen was statistically significantly associated with all 3 parameters. This study indicates that protease activity measurement as a stand-alone test can be used for detecting mastitis samples, using milk samples that have been frozen. Because protease activity acts in part on a different biological mechanism than somatic cell count or electrical conductivity, this test may increase the accuracy of mastitis diagnosis in combination with currently available in- or on-line tests in automated milking systems.

Information related to mastitis risk factors is useful for the design and implementation of clinical mastitis (CM) control programs. The first objective of our study was to model the risk of CM under Brazilian conditions, using cow-specific risk factors. Our second objective was to explore which risk factors were associated with the occurrence of the most common pathogens involved in Brazilian CM infections. The analyses were based on 65 months of data from 9,789 dairy cows and 12,464 CM cases. Cow-specific risk factors that could easily be measured in standard Brazilian dairy farms were used in the statistical analyses, which included logistic regression and multinomial logistic regression. The first month of lactation, high somatic cell count, rainy season and history of clinical mastitis cases were factors associated with CM for both primiparous and multiparous cows. In addition, parity and breed were also associated risk factors for multiparous cows. Of all CM cases, 54% showed positive bacteriological culturing results from which 57% were classified as environmental pathogens, with a large percentage of coliforms (35%). Coagulase-negative Staphylococcus (16%), Streptococcus uberis (9%), Streptococcus agalactiae (7%) and other Streptococci (9%) were also common pathogens. Among the pathogens analyzed, the association of cow-specific risk factors, such as Zebu breed (OR=5.84, 95%CI 3.77–10.77) and accumulated history of SCC (1.76, 95%CI 1.37–2.27), was different for CM caused by Coagulase-negative Staphylococcus and S. agalactiae in comparison to CM caused by coliforms. Our results suggest that CM control programs in Brazil should specially consider the recent history of clinical mastitis cases and the beginning of the lactations, mainly during the rainy season as important risk factor for mastitis.

To improve cow management in large dairy herds, sensors have been developed that can measure physiological, behavioral, and production indicators on individual cows. Recently, the number of dairy farms using sensor systems has increased. It is not known, however, to what extent sensor systems are used on dairy farms, and the reasons why farmers invest or not in sensor systems are unclear. The first objective of this study was to give an overview of the sensor systems currently used in the Netherlands. The second objective was to investigate the reasons for investing or not investing in sensor systems. The third objective was to characterize farms with and without sensor systems. A survey was developed to investigate first, the reasons for investing or not in sensor systems and, then, how the sensor systems are used in daily cow management. The survey was sent to 1,672 Dutch dairy farmers. The final data set consisted of 512 dairy farms (response rate of 30.6%); 202 farms indicated that they had sensor systems and 310 farms indicated that they did not have sensor systems. A wide variety of sensor systems was used on Dutch dairy farms; those for mastitis detection and estrus detection were the most-used sensor systems. The use of sensor systems was different for farms using an automatic milking system (AMS) and a conventional milking system (CMS). Reasons for investing were different for different sensor systems. For sensor systems attached to the AMS, the farmers made no conscious decision to invest: they answered that the sensors were standard in the AMS or were bought for reduced cost with the AMS. The main reasons for investing in estrus detection sensor systems were improving detection rates, gaining insights into the fertility level of the herd, improving profitability of the farm, and reducing labor. Main reasons for not investing in sensor systems were economically related. It was very difficult to characterize farms with and without sensor systems. Farms with CMS and sensor systems had more cows than CMS farms without sensor systems. Furthermore, farms with sensor systems had fewer labor hours per cow compared with farms without sensor systems. Other farm characteristics (age of the farmer, availability of a successor, growth in herd size, milk production per cow, number of cows per hectare, and milk production per hectare) did not differ for farms with and without sensor systems.

Op veel bedrijven is mastitis een onderschatte kostenpost waarbij een bedrijfsspecifieke aanpak op basis van beschikbare data loont. Dit bleek tijdens de masterclass uiergezondheid voor dierenartsen, georganiseerd door Gezondheidsdienst voor Dieren en DeLaval in Steenwijk.

In the past few decades, farms have increased in size and the focus of management has changed from curative to preventive. To help farmers cope with these changes, veterinarians offer veterinary herd health management (VHHM) programs, whose major objective is to support the farmer in reaching his farm performance goals. The association between farm performance and participation in VHHM, however, remains unknown. The aim of this paper was to compare farm performance parameters between participants and nonparticipants in VHHM and to differentiate within participation to evaluate the possible added value of VHHM on the farm. Five thousand farmers received a questionnaire about the level of VHHM on their farm. Farm performance parameters of these 5,000 farms were provided. For all respondents (n = 1,013), farm performance was compared between participants and nonparticipants and within level of participation, using linear mixed and linear regression models. Farmers who participated in VHHM produced 336 kg of milk/cow per year more and their average milk somatic cell count (SCC) was 8,340 cells/mL lower than farmers who did not participate in VHHM. Participating herds, however, had an older age at first calving (+12 d), a lower 56-d nonreturn rate percentage (-3.34%), and a higher number of inseminations per cow (+0.09 inseminations). They also had more cows culled per year (+1.05%), and a lower age at culling (-70 d). Participants in the most-extended form of VHHM (level 3) had a lower SCC (-19,800 cells/mL), fewer cows with high SCC (-1.70%), fewer cows with new high SCC (-0.47%), a shorter calving interval (-6.01 d), and fewer inseminations per heifer (-0.07 inseminations) than participants in the least-extended form of VHHM (level 1). Level 3 participants, however, also had more cows culled per year (+1.74%) and a lower age at culling (-103 d). Discussing specific topics with the veterinarian (milk production, fertility, and udder health) had only marginal effects on improving the farm performance parameters related to those topics. Given the relevance of fertility on the farm and the focus on longevity by society, it is important to determine underlying reasons for the negative associations of these topics with participation in VHHM. A longitudinal study could provide answers to this. For now, veterinarians should be aware of the associations. The increased milk production and milk quality could help the marketing of VHHM to farmers.

Veterinary herd health management (VHHM) programmes are meant to support herd health and farmers’ income (Brand and Guard 1996). They were introduced in the Netherlands in the 1970s (Sol and Renkema 1984) and at present many veterinarians provide them to farmers. VHHM comprises a basic structure of goal setting, planning, execution and evaluation. Farms are visited every four to six weeks, where the veterinarian inspects the animals, evaluates gathered data and provides advice (Brand and Guard 1996). Ideally, VHHM combines animal health, food safety, animal welfare and public health with farm management and economics (Noordhuizen and Wentink 2001, LeBlanc and others 2006). VHHM programmes are used not only in the Netherlands, but on a wider scale, for instance in the UK (Wassell and Esslemont 1992) and Denmark (Kristensen and Enevoldsen 2008). The farmers in Europe have to produce under strict, often expensive and laborious, regulations while competing with commercial farmers outside the EU who are not subjected to the same rules (Cannas de Silva and others 2006). As dairy farmers strive for further efficiency in production, driven by market economics, the risks and consequences of poor health and suboptimal production increase (Sibley 2006). VHHM programmes are meant to help farmers to produce products of high quality for a low cost price (Brand and Guard 1996). Its primary objectives include the optimisation of herd health, productivity, quality of products and profitability of the dairy enterprise (Blood and others 1978). In practice, some farms or veterinarians embrace the concepts of VHHM by active participation, while others do not (Derks and others 2012). Also, veterinarians are not always able to meet farmers’ requirements for VHHM (Hall and Wapenaar 2012, Derks 2013). Its efficiency is hard to determine. There is, for instance, no recent information on the cost implications of this programme with regard to perceived profitability in farms. A limited number of controlled studies were carried out in the early 1970s and 1980s to evaluate the effects of VHHM on farm performance (Williamson 1980, Sol and others 1984). It was shown that a VHHM programme produced considerable benefits to participating farmers. A follow-up study showed approximately 8 per cent increase in margin per cow compared with the initial margin, using 1974–1975 as a base year (Hogeveen and others 1992). Since that time, no economic studies were carried out on the effects of these programmes. This study evaluates the economic relationship between participation in a VHHM programme and farm performance on dairy farms and estimates the costs incurred by the participation in this programme as well as the net returns (NR), including factors that influence the NR on a farm.

Because of increasing bulk milk somatic cell counts and continuous clinical mastitis problems in a substantial number of herds, a national mastitis control program was started in 2005 to improve udder health in the Netherlands. The program started with founding the Dutch Udder Health Centre (UGCN), which had the task to coordinate the program. The program consisted of 2 parts: a research part and a knowledge-transfer part, which were integrated as much as possible. The knowledge-transfer part comprised 2 communication strategies: a central and a peripheral approach. The central approach was based on educating farmers using comprehensive science-based and rational argumentation about mastitis prevention and included on-farm study group meetings. Comprehensive education materials were developed for farmers that were internally motivated to improve udder health. In the peripheral approach it was tried to motivate farmers to implement certain management measures using nontechnical arguments. Mass media campaigns were used that focused on one single aspect of mastitis prevention. These communication strategies, as well as an integrated approach between various stakeholders and different scientific disciplines were used to reach as many farmers as possible. It should be noted that, because this intervention took place at a national level, no control group was available, as it would be impossible to isolate farmers from all forms of communication for 5 years. Based on several studies executed during and after the program, however, the results suggest that udder health seemed to have improved on a national level during the course of the program from 2005 to 2010. Within a cohort of dairy herds monitored during the program, the prevalence of subclinical mastitis did not change significantly (23.0 in 2004 vs. 22.2 in 2009). The incidence rate of clinical mastitis, however, decreased significantly, from 33.5 to 28.1 quarter cases per 100 cow years at risk. The most important elements of the farmers' mindset toward mastitis control also changed favorably. The simulated costs of mastitis per farm were reduced compared with a situation in which the mastitis would not have changed, with € 400 per year. When this amount is extrapolated to all Dutch farms, the sector as a whole reduced the total costs of mastitis by € 8 million per year. It is difficult to assign the improved udder health completely to the efforts of the program due to the lack of a control group. Nevertheless, investing € 8 million by the Dutch dairy industry in a 5-yr national mastitis control program likely improved udder health and seemed to pay for itself financially.

Poor hygiene is an important risk factor for reduced udder health. Because the teat cleaning process is done automatically on farms with an automatic milking system (AMS), hygiene management might differ. The aim of this study was to determine the relationship between hygiene and udder health on farms with an AMS at the farm level as well as at the cow level. Information on hygiene and udder health was collected on 151 Dutch dairy farms with an AMS. Teams of 2 veterinary students collected data with the use of a partially open-ended questionnaire and scoring protocols for hygiene of the cows, cleanliness of the AMS, and functioning of the AMS. Milk production records from the Dutch dairy herd information association were also collected. Stepwise general linear models were used to analyze the relation between hygiene and udder health at farm level. Dependent variables were average herd somatic cell count (SCC), the average percentage of new cows with a high SCC, and the incidence rate of clinical mastitis, all in the year preceding the farm visit. The annual average herd SCC was positively related to the proportion of cows with dirty teats before milking and the proportion of cows with dirty thighs. The annual average percentage of new cows with a high SCC was positively related to the proportion of cows with dirty teats before milking and the proportion of milkings where teats were not covered with teat disinfecting spray by the AMS. The annual incidence rate of clinical mastitis was positively related to the frequency of replacing the milking filters. At the cow level, hygiene scores of the udder, thighs, and legs (range 1 to 4, where 1 is clean and 4 is very dirty) were related with cow SCC from the milk production test day closest to the farm visit using a general linear mixed model. The relationship between cow SCC and the hygiene score of the udder was positive. Key words: udder health; hygiene; automatic milking system

Mastitis kost geld. Maar hoeveel? Beduidend meer dan melkveehouders denken, zo blijkt uit recent onderzoek van het Uiergezondheidscentrum Nederland en de faculteit Diergeneeskunde onder 64 veehouders. Vooral het productieverlies bij subklinische mastitis kost veel euro’s

Treatment of subclinical mastitis is traditionally no common practice. However, some veterinarians regard treatment of some types of subclinical mastitis to be effective. The goal of this research was to develop a stochastic Monte Carlo simulation model to support decisions around treatment of chronic subclinical mastitis caused by Streptococcus uberis. Factors in the model include, amongst others, the probability of spontaneous cure, probability of the cow becoming clinically diseased, transmission of infection to other cows, cure rate under treatment and physiological effects of the infection. The average economic damage (with basic input parameters for the Dutch circumstances) of an untreated chronic subclinical mastitis case caused by S. uberis is € 88. With a short (3 day) treatment, the average damage was higher. For the average cow, treatment is not economical efficient. Sensitivity analysis showed that this might depend on some specific cow and farm factors. Moreover, the spread of economic damage indicates that the risk of a high damage is much higher when a cow with chronic subclinical mastitis is not treated

Veel melkveehouders hebben een abonnement op het celgetalprogramma van het NRS. Dit betekent dat bij een proefmelking niet alleen de standaardmetingen gedaan worden, maar dat ook het aantal cellen in de melk gemeten wordt. Koeien worden op basis van een te hoog celgetal geattendeerd. Ook is het mogelijk met het celgetal-B.O. programma bacteriologisch onderzoek (B.O.) uit te voeren bij koeien met attenties. De resultaten van dit programma zijn beoordeeld op twee bedrijven: een bedrijf met een laag tankmelkcelgetal en een bedrijf met een hoog tankmelkcelgetal. Het bleek dat dat celgetalprogramma op beide bedrijven de meeste koeien met chronische celgetalproblemen attendeerde. De resultaten van het celgetal-B.O. programma vonden op beide bedrijven de belangrijkste bacteriën. Op een bedrijf met een laag celgetal waren echter meer melkmonsters negatief.

Beschrijving van de representatie van de benodigde en aanwezige kennis voor de pathogeendiagnose bij mastitis op koeniveau met behulp van een Bayesiaans netwerk

Dit artikel beschrijft in het kort de gebruikte methode, het uiteindelijke kennissysteem en de resultaten van de validatie

In dit overzichtsartikel zijn enkele hoofdzaken in de epidemiologische ontwikkelingen op het gebied van uiergezondheid in Nederland samengevat

Highlights? Study the usefulness of sensor measurements to predict clinical mastitis pathogens. ? Applying decision-tree induction to sensor data collected during automatic milking. ? Currently available sensors cannot be used to predict clinical mastitis pathogens. Automatic milking systems produce mastitis alert lists that report cows likely to have clinical mastitis (CM). A farmer has to check these listed cows to confirm a CM case and to start an antimicrobial treatment if necessary. In order to make a more informed decision, it would be beneficial to have information about the CM causal pathogen at the same time a cow is listed on the mastitis alert list. Therefore, this study explored whether decision-tree induction was able to predict the Gram-status of CM causal pathogens using in-line sensor measurements from automatic milking systems. Data were collected at nine Dutch dairy farms milking with automatic milking systems and included 140 bacteriological cultured CM cases with sensor measurements of electrical conductivity, colors red, green, and blue and milk yield for analyses. In total, 110 CM cases were classified as Gram-positive CM cases and 30 as Gram-negative. Stratified randomization was used to divide the data in a training set (n=96) for model development, and a test set (n=44) for validation. The decision tree used three variables to predict the Gram-status of the CM causal pathogen; two variables were based on electrical conductivity measurements, and one on measurements of the color blue. This decision tree had an accuracy of 90.6% and a kappa value of 0.76 based on data in the training set. When only those CM cases were considered with extreme high probability estimates for their Gram-status (either positive or negative), 74% of all records in the training set could be classified with a stratified accuracy of 97.1%. When validated, the decision tree performed poorly; accuracy dropped to 54.5% and the kappa value to -0.20. The stratified accuracy calculated for 75% of all records in the test set was 66.7%. Predicting the CM causal pathogen showed a similar poor result; the decision tree had an accuracy of 27.9% and a kappa of 0.12, based on data in the test set. Based on these results, it is concluded that decision-tree induction in conjunction with sensor information from the electrical conductivity, color, and milk yield provide insufficient discriminative power to predict the Gram-status or the CM causal pathogen itself.