A study of methods for evaluating the success of the transition period in early-lactation dairy cows.

Three transition monitors were developed in this study that serve on 2 levels: the individual cow level and the herd level. On the first level they screen all cows for potential onset of postparturient health disorders and could be used to trigger implementation of more specific diagnostic initiatives. On the second level they can be used within herd to monitor the implementation of transition protocols and evaluate the transition management on the farm, signaling potential problems before clinical disease onset. The performance of 3 transition monitors based on daily milk yield (MY) within the first 7d in milk was evaluated in 3 herds with differing transition management intensity. The 3 monitors considered were increase in MY (LINE), average MY (MY7), and the difference between MY7 and expected MY (transition success measure, TSM). Transition monitors were evaluated not only as within-herd predictors of individual cow transition problems but also as indicators of herd transition management failures by relating their value with probability of early-lactation health disorders, culling, and treatment cost. Analysis of logistic models, correlations, and sensitivity and specificity estimates identified TSM as the most reliable measure of transition failure on both the individual cow level as well as the farm level across all study herds, with best performance achieved in herds with the most intensive postpartum cow management. As evaluated by logistic regression models, TSM was able to successfully predict the probability of a cow remaining healthy for the first 21d of lactation (c-statistic between 0.68 and 0.78), and probability of culling by 100d in milk (c-statistic between 0.73 and 0.86). Total cost of treatment by 21d in milk also showed the strongest correlation with TSM, with correlation coefficients ranging between 0.2 and 0.4. Statistical-process control cumulative sum charts for TSM designed to monitor postpartum management process in the herd identified transition failure events with at least 90% sensitivity at specificity above 92% within a 14-d window of 7d before and 7d after the event.

A novel method of analyzing daily milk production and electrical conductivity to predict disease onset.

This study evaluates the changes in milk production (yield; MY) and milk electrical conductivity (MEC) before and after disease diagnosis and proposes a cow health monitoring scheme based on observing individual daily MY and MEC. All reproductive and health events were recorded on occurrence, and MY and MEC were collected at each milking from January 2004 through November 2006 for 587 cows. The first 24 mo (January 2004 until December 2005) were used to investigate the effects of disease on MY and MEC, model MY and MEC of healthy animals, and develop a health monitoring scheme to detect disease based on changes in a cow's MY or MEC. The remaining 11 mo of data (January to November 2006) were used to compare the performance of the health monitoring schemes developed in this study to the disease detection system currently used on the farm. Mixed model was used to examine the effect of diseases on MY and MEC. Days in milk (DIM), DIM x DIM, and ambient temperature were entered as quantitative variables and number of calves, parity, calving difficulty, day relative to breeding, day of somatotropin treatment, and 25 health event categories were entered as categorical variables. Significant changes in MY and MEC were observed as early as 10 and 9 d before diagnosis. Greatest cumulative effect on MY over the 59-d evaluation period was estimated for miscellaneous digestive disorders (mainly diarrhea) and udder scald, at -304.42 and -304.17 kg, respectively. The greatest average daily effect was estimated for milk fever with a 10.36-kg decrease in MY and 8.3% increase in MEC. Milk yield and MEC was modeled by an autoregressive model using a subset of healthy cow records. Six different self-starting cumulative sum and Shewhart charting schemes were designed using 3 different specificities (98, 99, and 99.5%) and based on MY alone or MY and MEC. Monitoring schemes developed in this study issue alerts earlier relative to the day of diagnosis of udder, reproductive, or metabolic problems, are more sensitive, and give fewer false-positive alerts than the disease detection system currently used on the farm.

Water intake and dry matter intake changes as a feeding management tool and indicator of health and estrus status in dairy cows.

This study investigates whether dry matter (DM) or water intake is affected by the presence of disease or estrus in dairy cows and whether water intake can serve as an accurate substitute for monitoring changes in DM intake (DMI). A combined cumulative sum (CUSUM) and Shewhart monitoring scheme is proposed to detect DMI changes and emerging disease or estrus. Daily readings from 35 inline water meters for 35 water cups in a tie-stall barn at the University of Minnesota were collected from September 2005 until June 2006. Two cows were assigned to each water cup. Individual DMI were recorded for each of the 70 cows on the study. All drug or hoof treatments administered to the cows along with breeding and calving events were also recorded and classified as 1 of the following 6 event categories: estrus, calving, mastitis, fever, hoof treatment, and other. Analysis of covariance was used to identify factors significantly changing intake. Only the first 150 d in milk (DIM) were considered in the analysis. Six event categories plus DIM, ambient temperature, relative humidity, and parity were entered as independents into the model. Calving, primiparity, and health events categorized as "other" were associated with decreased DM and water intake. Mastitis decreased DMI and fever negatively affected water intake. Both intakes increased with DIM, and water intake decreased with increase in humidity. Covariance analysis was used to investigate the relationship between DMI and water intake. In model 1, analysis was done for a pair of cows, whereas model 2 modeled DMI of the whole group of 70 cows. Water intake, ambient temperature, humidity, and DIM were entered as independents in both models and parity was entered in model 1. Polynomial models and 2-way interactions were also considered. Water intake, ambient temperature, DIM, and DIM(2) were kept in final models 1 and 2, and parity was kept in model 1. Final models for cow pairs and a group of 70 cows resulted in R(2) of 0.50 and 0.82, respectively. The proposed CUSUM-Shewhart DMI monitoring scheme successfully detected emerging disease even in the first week of lactation. Monitoring water intake can serve as an alternative to measurements of DMI for groups of cows and has the potential of predicting change in individual cow health and estrus status.

Predicting somatic cell count standard violations based on herd's bulk tank somatic cell count. Part I: Analyzing variation.

The present study examines the relationship between the bulk tank somatic cell count (SCC) mean and sigma (an estimate of variation) and the probability of exceeding a SCC standard. Daily or every other day, bulk tank SCC data were collected for 24 mo from 1,501 herds. Mean and sigma were estimated for each herd monthly and were compared between months and herd production categories using Kruskal-Wallis nonparametric ANOVA. The effect of month on bulk tank SCC mean and sigma was significant, with estimates for all summer months and most of the spring and fall months being significantly greater than estimates of mean and sigma in December 2004. Logistic regression models were developed to examine the relationship between month and herd production and the odds of a herd exceeding a SCC standard. The odds of exceeding a bulk tank SCC standard were significantly greater in the summer months and for smaller herds. A grid was constructed determining the probability of exceeding any of 5 SCC standards (200,000 to 600,000 cells/mL, step 100,000 cells/mL) in the following month, based on the mean and sigma of the past month. The violation probability grid can be used to assess the prospect of meeting quality premium goals and to proactively encourage more consistent performance in all the processes affecting bulk tank SCC.

Predicting somatic cell count standard violations based on herd's bulk tank somatic cell count. Part II: Consistency index.

The present study examines the capability of 1,501 herds in the Upper Midwest and the performance of statistical process control charts and indices as a way of monitoring and controlling milk quality on the farm. For 24 mo, daily or every other day bulk tank somatic cell count (SCC) data were collected. Consistency indices for 5 different SCC standards were developed. The indices calculate the maximum variation allowed to meet a desired SCC level at a given mean bulk tank SCC and were used to identify herds not capable of meeting a specific SCC standard. Consistency index method was compared with a test identifying future bulk tank SCC standard violators based on herds' past violations. The performance of the consistency index test and the past violation method was evaluated by logistic regression. The comparison focused on detection probability and certainty associated with a result. For the 5 SCC levels, detection probability and certainty associated with a result ranged from 51 to 98%. Detection probability of all violators and certainty associated with a negative result was greater for the consistency index across all 5 SCC levels (by 0.7 to 7.4% and 2.1 to 5.1%, respectively). Control charts were plotted and monthly consistency indices calculated for individual farms. Charts in combination with the consistency indices would warn from 66 to 80% of the herds about an upcoming violation within 30 d before it occurred. They offer a proactive approach to maintaining consistently high milk quality. By assessing process capability and distinguishing between significant changes and random variation in bulk tank SCC, tools presented in this article encourage fact-based decisions in dairy farm milk quality management.

Bulk tank somatic cell counts analyzed by statistical process control tools to identify and monitor subclinical mastitis incidence.

The objective of this study was to examine the relationship between monthly Dairy Herd Improvement (DHI) subclinical mastitis and new infection rate estimates and daily bulk tank somatic cell count (SCC) summarized by statistical process control tools. Dairy Herd Improvement Association test-day subclinical mastitis and new infection rate estimates along with daily or every other day bulk tank SCC data were collected for 12 mo of 2003 from 275 Upper Midwest dairy herds. Herds were divided into 5 herd production categories. A linear score [LNS = ln(BTSCC/100,000)/0.693147 + 3] was calculated for each individual bulk tank SCC. For both the raw SCC and the transformed data, the mean and sigma were calculated using the statistical quality control individual measurement and moving range chart procedure of Statistical Analysis System. One hundred eighty-three herds of the 275 herds from the study data set were then randomly selected and the raw (method 1) and transformed (method 2) bulk tank SCC mean and sigma were used to develop models for predicting subclinical mastitis and new infection rate estimates. Herd production category was also included in all models as 5 dummy variables. Models were validated by calculating estimates of subclinical mastitis and new infection rates for the remaining 92 herds and plotting them against observed values of each of the dependents. Only herd production category and bulk tank SCC mean were significant and remained in the final models. High R2 values (0.83 and 0.81 for methods 1 and 2, respectively) indicated a strong correlation between the bulk tank SCC and herd's subclinical mastitis prevalence. The standard errors of the estimate were 4.02 and 4.28% for methods 1 and 2, respectively, and decreased with increasing herd production. As a case study, Shewhart Individual Measurement Charts were plotted from the bulk tank SCC to identify shifts in mastitis incidence. Four of 5 charts examined signaled a change in bulk tank SCC before the DHI test day identified the change in subclinical mastitis prevalence. It can be concluded that applying statistical process control tools to daily bulk tank SCC can be used to estimate subclinical mastitis prevalence in the herd and observe for change in the subclinical mastitis status. Single DHI test day estimates of new infection rate were insufficient to accurately describe its dynamics.