Bio-Rad Blackboard - Youden Charts


Uploaded by BIORADQC on 30.08.2012

Transcript:
Welcome to Bio-Rad Blackboard Training. This is a series of short informal tutorials brought
to you by industry experts. Today we’ll cover Youden charts.
Youden charts were developed in the 1950’s by Dr. William Youden. The chart allows the
user to compare two levels of controls and can help differentiate between systematic
and random error.
The chart can compare; level 1 and 2, level 1 and 3 or level 2 and 3.
The chart can be built to compare results to the means and standard deviations of a
consensus group, the laboratory itself or another instrument.
Let’s start by looking at the components that make up a Youden chart.
The vertical axis represents the means and standard deviations of one of the control
levels of the comparison group.
The horizontal axis represents the mean and standard deviations from the other control
level of the comparison group.
The vertical line in the middle of the chart represents the mean for the first level. The
horizontal line in the middle of the chart represents the mean for the second level.
The diagonal line is used to help identify problems with linearity and distinguishing
between random and systematic error.
Data points are plotted on the graph and will fall into one of 3 regions.
The white square will display points that are within 1 standard deviation of the mean
for both levels.
The light gray square will display points that are between 1 and 2 standard deviations
from the mean for both levels.
The dark gray square will display points that are between 2 and 3 standard deviations from
the mean for both levels.
Your goal is to be as close to the middle of the chart as possible. Think of the center
as a bull’s eye of a target. The more tightly points are clustered around the bull’s eye;
the better the precision is for each control.
If a point does not fall near the middle, notice where it does fall.
Plots that have the majority of points scattered all over rather than clustered around the
middle may indicate that there is a problem with the precision of the testing system.
Corrective action may need to be taken. We encourage you to take a look at our video
that covers precision and accuracy if you need more information.
Results can be considered linear when the points are on or near the angled line. Linear
means that both levels are responding to testing in the same direction.
When the points on the chart are along the angled line but are too high or too low recalibration
of the instrument might be necessary.
When the majority of the points are within the upper left and/or lower right quadrant
of the chart then problems with linearity may exist.
Points that are near the angled line and within the one or two standard deviation squares
show acceptable results.
Points that lie near the angled line but are outside the 2 standard deviation square indicate
systematic error.
Points that lie far from the angled line indicate random error.
As you can see, Youden Charts are quite helpful. We encourage you to explore our other Blackboard
videos that cover charts; including Levey-Jennings and Yundt charts.
For all of your QC needs, visit www.qcnet.com.