Control Chart Package
Current add-on version: 1.0.0 (build 58)
The Control Chart Package allows sophisticated statistical process control (for example, by plotting Shewhart I-Charts as recommended in the USP <1010>). It provides visualization options for intervals, subcharts and sidecharts, and allows you to define colors by a secondary characteristic. You can also define rules based on events and time frames.
The Control Chart Package supports the following USP <1010> methods:
- Parameter statistics
- Individual chart (I-Chart)
- Control limits
- Nelson & Western Electric Company (WECO) rules
- Calculation of the standard deviation
To illustrate how you can use the Control Chart Package in connection with these methods, we provide sample documents you can download when activating the add-on in PLA 3.0.
Additional information about each sample document is provided in the document's Comment section.
Parameter statistics
The Control Chart Package automatically calculates the following parameter statistics for each data series: mean and median, standard deviation (SD), coefficient of variation (CV), 1st and 3rd quartile, min. and max. values, number of values, and missing values. The statistics are displayed on the Dashboard and in the reports.
In addition, you can configure confidence intervals and sidecharts to deeply analyze your data.
Individual chart (I-Chart)
I-Charts monitor data such as measurements and regression parameters at regular intervals, with each data point within the chart representing a sample or an observation respectively.
All charts generated by the Control Chart Package are I-Charts. You can manually add the chart data, or aggregate it from other PLA 3.0 documents such as Quantitative response assays or Dose-response assays.
Control limits
Control limits define a range of acceptable values, determined by an upper and a lower limit. Any value outside of this range is considered to be a rule violation and needs to be marked as such.
Use the Control Chart Package to set up rule sets with independently defined upper and lower control limits. These rules will be drawn as horizontal lines on your control chart.
Nelson & Western Electric Company (WECO) rules
These decision rules let you detect out-of-control or non-random conditions on control charts, based on deviations from the mean. They are also capable of detecting patterns or trends. For example, you can apply Nelson rule 3 to detect when six (or more) points in a row are continually increasing (or decreasing).
The Control Chart Package provides a set of predefined control rules (Nelson rules 1 to 8 and WECO rules 1 to 4) and also lets you set up user-defined rules. You can also base control rules on parameter statistics of historical data. In contrast to control limits, where you can define the limits independently, a control rule automatically applies to both the upper and the lower limits.
Calculation of the standard deviation
USP <1010> recommends calculating the standard deviation using differences between consecutive data points instead of comparison to the overall mean.
The Control Chart Package supports both methods for estimating the standard deviation.