SELECTION OF QC PROCEDURE
LevyJennings Control Chart
Westgard Multirule chart
LevyJennings Control Chart
Control charts were first
introduced into the clinical chemistry laboratory by Levy and Jennings in 1950.
Here single control values are plotted directly. To use a LeveyJennings
control chart, follow these steps:
1. Analyze
control sample at least 20 different days. Calculate the mean and SD for those
results.
2. Construct a
control chart either manually on graph paper or using computer. In the yaxis put
control values along with mean ± 4SD which
covers about 95% or 99.7% of the measurements. Draw horizontal lines for the
mean and upper and lower control limits. Set the control limits as the mean ± 3SD when the number of control
observations, n is 2 or greater. When n is 1, control limits may be set as mean
± 2SD. Label
x axis in terms of time, using day, run number, control observations number.
3. Introduce
control specimen into each analytical run, record the values and plot each
value on the control chart.
4. When the control value fall within the control limits (±3s), interpret the run as being
in control and report the patient results. When a single control value exceeds
the control limits, stop the method; do not report patient results. Inspect
method to determine the cause for the errors. Resolve the problem, then repeat
the entire run specimen and control samples.
In practice 1_{2s} rule has
been used for rejection of test when n = 1. Whenever control value exceeds 2s
limit repetitive measurement of control and patient sample should be done. When
a second or repeated control value is outside 2s control limit, there must be
true rejection and problem solving procedure should be started.
Westgard Multirule chart
The Multirule procedure developed
by Westgard and associates uses multiple control rules for interpreting control
data. The procedures require a chart having lines for control limits drawn at
the mean ±
1s, 2s and 3s. This chart drawing is similar to leveyJennings chart.
The following control rules are used.
Rule

Description

1_{2s}

One control observation exceeding the mean ± 2s
Warning sign, control data should be tested
by other control rules, do not report the result

1_{3s}

One control observation exceeding the mean ± 3s
Rejection, due to random error

2_{2s}

Two consecutive control data exceeding same mean
plus 2s or mean minus 2s
Rejection, systematic error

R_{4s}

One observation exceeding mean plus 2s and
another exceeding mean minus 2s
Rejection, random error. Note: this rule applies
only within a run not between runs.

4_{1s}

Four consecutive observations exceeding mean plus
1s or mean minus 1s
Rejection, systematic error

10_{x}

Ten consecutive control data falling on one side
of mean (above or below)
Rejection, systematic error
