|
|
| Each dairy operation can be viewed as a system of processes directly or indirectly impacting one another and ultimately creating the quality of the end product, raw milk. |
|
|
 |
 |
|
| Are we improving, staying the same, or getting worse? |
by Mireille Chahine, University of Idaho
Joanna M. Lukas and Jeffrey K. Reneau, University of Minnesota
Profitable production of safe, high quality raw milk from healthy cows is the primary goal of every dairy. All who are involved in managing dairies know that this goal can only be accomplished by a lot of hard work and a firm commitment to quality management practices. Excellent employee training, good communications between all staff, and standardization of herd management procedures are keys to success especially as herd size expands.
Each dairy operation can be viewed as a system of processes directly or indirectly impacting one another and ultimately creating the quality of the end product, raw milk. This makes the quality of the milk produced on a farm a reflection of the performance of all the processes that constitute a dairy operation. Therefore monitoring that quality is critical for the herd manager and all employees to know how the herd is doing.
We receive this kind of feedback from our creamery when at every pick up the milk in the bulk tank is sampled and tested for fat, protein and SCC. However rarely do we take full advantage of this information. Test results in the form of rows and columns of numbers are hard to read or interpret. Summaries in the form of monthly or yearly averages enable us to compare one month or year to another. They do not however help us distinguish between natural month to month variation and/or notice a trend that we should be concerned about. In addition, these summaries usually come too late and are therefore taken out of context making it hard for a herd manager to tie any shift in fat, protein or SCC to a specific day, event or change on the farm.
Daily monitoring of milk quality represented in a more "comprehensible" form than tables of numbers would help provide information about herd’s past, present and future performance in time for managers and other dairy employees to make appropriate decision.
Statistical Process Control
Statistical process control (SPC) is a tool that has been used for a long time in other branches of the industry to manage quality. SPC measures process behavior with control charts. It looks at the operation as a system of processes and allows us to know when REAL change has occurred. By examining a control chart we are able to answer one very important question: Are we improving, staying the same, or getting worse? The sole purpose of SPC techniques is to allow production managers to distinguish with statistical CERTAINTY the difference between “normal” (common cause) and “abnormal” (special cause) variation. In recent years, some work has been done to adapt these techniques to the dairy industry. Daily and weekly herd management data can be graphed into easily understood control charts that provide powerful day-to-day decisions aids. |
| |
 |
| |
| Steps for Developing a Control Chart |
- Start with a times series chart.
- Add a centerline for central reference
- Add control limits computed from the data
- Apply tests to distinguish REAL change from random variation
Setting Control Limits
Plotting data in a form of a time series plot (Figure 1) already helps identify the more obvious trends or changes in our process. However, only after adding control limits will we be able to discover the more subtle but significant changes (Figure 2). Control limits are calculated based on the data collected in a time frame called the "control period". Experience has shown that 20 data points are needed to calculate credible control limits. When initiating a control chart, it is appropriate to use the first 20 data points collected as the control period. Once control charting is established, the control period can be set depending on the question being asked. If the intent is to monitor a process for the purpose of maintaining a stable process then using data from an apparently stable period makes sense. If the intent is to use the control chart to evaluate the introduction of a new product or a change in process procedure, etc., then the control period should be calculated from data collected just prior to the introduction of the product or change in procedure.
How to Interpret a Control Chart
When any of the following conditions are met, the process being measured is considered "out of control".
- A single data point outside the upper or lower control limits.
- Two of three consecutive data points on one side of the center line and more than two standard deviations away from the center line.
- Four of five consecutive data points on one side of the center line and more than one standard deviation away from the center line.
- Eight or more consecutive points on one side of the center line.
Adaptation of Statistical Process Control for Milk Quality Monitoring
If we now look at a dairy operation as a system of processes we will find that there are numerous applications for SPC in dairy herd management; from monitoring feed intake, health or reproductive performance to the most obvious and already possible: milk quality monitoring. On most dairies, daily milk quality data are readily available. As mentioned previously it is very hard to draw conclusion by looking at rows and columns of data or at a single number. Looking at a chart, however, allows us to identify where the process is heading. Did we have a significant increase in BTSCC in the last month? Did the new teat dip that we purchased improve milk quality on the dairy? Has a change in our bedding routine affected BTSCC?
MilkLab Control Charting System is now Available on the Web (http://www.dairyperformance.com).
MilkLab is an Internet-based monitoring system. Since the system is automated, no on-farm data entry is required for daily SCC, fat%, protein% and milk urea nitrogen tests. Data can be automatically uploaded to the MilkLab’s Internet website directly from any lab in the world. There is also the capability to upload DMI and daily milk production data to the website from the farm for control charting purposes. Any farm that subscribes to the service ($12.99 per month) can give website access of their data to their farm consultants. In addition, there is a built in alert system (Figure 3) that will automatically email or fax the producer, or anyone given data access, when any variable has significantly changed to exceed a minimum or maximum setting or is deemed “out of control” by SPC rules. In effect, this provides a 24/7 “real time” watch system for the producer and any of the consultants that have been given access to the herd data. The turnaround time from milk sample pickup at the farm to control chart is 2 to 4 days. Therefore, the time between potential management cause and effect is minimized.
Conclusion
Monitoring not only cow performance, but also monitoring and organizing people performance will be the hallmarks of dairy management success in the future. Where possible, the use of daily monitoring and SPC techniques will provide a powerful means of making accurate day-to-day herd management decisions. |
|
|
|
Email to a friend
Resource Summary
