7B6006 StatTools TrngMod

Part No.
7B6006
Cat No. 184 9017
Statistical Tools of Quality Control:

Training Module
Export authorized under general license GTDA (General Technical Data Available)
IMPORTANT
The information contained herein is based on the experience and knowledge relating to the subject
matter gained by Ortho-Clinical Diagnostics, Inc. prior to publication.
No patent license is granted by the information.
Ortho-Clinical Diagnostics, Inc. reserves the right to change this information without notice, and
makes no warranty, express or implied, with respect to the information. The company shall not be
liable for any loss or damage, including consequential or special damages resulting from the use
of this information, even if loss or damage is caused by its negligence or other fault.
VITROS is a trademark of Ortho-Clinical Diagnostics, Inc.
 Ortho-Clinical Diagnostics, Inc., 1995 - 2006. All rights reserved.

Revision History
Revision Date Description

2006-03-31 Updated to include Quality-Control Rule Section.
1/00 Publication formatting characters removed.
8/99 Changed company name to Ortho-Clinical Diagnostics, Inc. a
Johnson & Johnson Company.
5/97 Reprint; introduction of Johnson & Johnson Vitros Chemistry
Systems trademark nomenclature.
12/95 First release of training module.
Part No. 7B6006 Statistical Tools of Quality Control: Training Module iii
2006-03-31 VITROS Chemistry Systems
List of Revised Pages
Each page in your manual should be at the date listed below:
Revision Date Section Page

2006-03-31 All chapters All pages
1/00 All chapters All pages
iv Statistical Tools of Quality Control: Training Module Part No. 7B6006

VITROS Chemistry Systems 2006-03-31
About This Manual
Purpose of Manual
The purpose of this module is to provide laboratory personnel with working
knowledge of statistics needed to perform quality-control procedures. This
module can be used as a review or as an introduction to basic statistics.
Audience
This module is meant to be used by laboratory personnel working with
VITROS Chemistry Systems.
How This Manual Is Organized

The introduction identifies what you will be able to do as a result of this
training. The body of the module presents the content about statistical tools
of quality control. Appropriate practice exercises are provided to assess
mastery of information. The final section contains answer keys to check
work. The basic activity flow for using this material is Read - Do - Review.
Additional Training Modules

Performance Verifiers: Training Module
Part No. J04498
CAT No. 680 0325
Additional training modules are available on orthoclinical.com. For

ordering information, please contact your Ortho-Clinical Diagnostics
Laboratory Specialist.
Part No. 7B6006 Statistical Tools of Quality Control: Training Module v

vi Statistical Tools of Quality Control: Training Module Part No. 7B6006
Table of Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Quality Assurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Control of Preanalytical Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Control of Analytical Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Control of Postanalytical Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Quality-Control Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Statistical Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Accuracy and Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Central Tendency and Mean. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
Statistical Measures of Precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
Quality-Control Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Selection and Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
Reconstitution of Quality-Control Materials . . . . . . . . . . . . . . . . . . . . . . . . .18
Storage of Quality-Control Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18
Level I and Level II Quality-Control Materials . . . . . . . . . . . . . . . . . . . . . . .19
What is a Quality-Control Rule?. . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Individual Quality-Control Rules . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Other Rules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
The 12s Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
The 13s Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
The 22s Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25
Interpreting Quality-Control Results . . . . . . . . . . . . . . . . . . . . . . . 27

Levey-Jennings Chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27
Calculating ±2 Standard Deviation Ranges . . . . . . . . . . . . . . . . . . . . . . . . . .28
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Practice Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Practice Exercise 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Practice Exercise 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
Answer Keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Part No. 7B6006 Statistical Tools of Quality Control: Training Module vii
Self-Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Self-Evaluation Answer Key . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Part No. 7B6006 Statistical Tools of Quality Control: Training Module viii
List of Figures
Figure 1. Accuracy and Precision . . . . . . . . . . . . . . . . . . . . . . 5
Figure 2. Variation in Height Measurement . . . . . . . . . . . . . . . 6

Figure 3. Frequency Distribution of Height Measurements . 7
Figure 4. Glucose Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Figure 5. Standard Deviation . . . . . . . . . . . . . . . . . . . . . . . . . 10

Figure 6. Example Standard Deviation Calculations . . . . . . 11
Figure 7. Calculating Standard Deviation . . . . . . . . . . . . . . . 12
Figure 8. Frequency Distribution . . . . . . . . . . . . . . . . . . . . . 14

Figure 9. Standard Deviation and Probability . . . . . . . . . . . 14
Figure 10. Types of Quality-Control Materials . . . . . . . . . . . 17
Figure 11. Three Quality-Control Rule . . . . . . . . . . . . . . . . . . 21
Figure 12. Rule 22s Violation 1 . . . . . . . . . . . . . . . . . . . . . . . . 23
Figure 15. Control Rule Violation 22s . . . . . . . . . . . . . . . . . . . 25
Figure 16. Control Rule Violation 13s . . . . . . . . . . . . . . . . . . . 26
Figure 17. Levey-Jennings Chart . . . . . . . . . . . . . . . . . . . . . 27
Figure 18. Standard Deviation Range . . . . . . . . . . . . . . . . . . 29

Figure 19. Using Levey-Jennings Chart . . . . . . . . . . . . . . . . 30
Part No. 7B6006 Statistical Tools of Quality Control: Training Module ix

Part No. 7B6006 Statistical Tools of Quality Control: Training Module x
Introduction
Introduction
The purpose of this module is to provide a summary of the basic concepts
and practices of quality control used in the laboratory.
After completing this module, you will be able to calculate important
values associated with these programs and correctly interpret the
information obtained from those calculations.
Objectives
Upon completion of this module, you will be able to:
• Discuss the importance of quality assurance and quality-control
programs in the clinical laboratory.
• Name several parameters that should be monitored and evaluated by a
properly designed quality assurance program.
• Explain how a quality-control system is used in the clinical laboratory to
verify the accuracy of patient test results.
• Define the following terms as they relate to a set of repeated
measurements: accuracy, precision, frequency distribution, and central
tendency.
• Calculate the following statistical parameters: mean, standard deviation,
coefficient of variation, and ±1, 2, & 3 standard deviation ranges.
• Select the appropriate type of quality-control material for a given
laboratory test procedure.
This module will take approximately two hours to complete. You will need
the following items:
• Calculator*
• Pencil
This module is designed with practice exercises to help you understand
each concept. The answers to the practice exercises can be found in the
answer keys at the end of the module.
*Most calculators can perform the math taught here—make sure you
understand the concepts.
Part No. 7B6006 Statistical Tools of Quality Control: Training Module 1

Quality Assurance
Quality Assurance
Quality-assurance programs are designed to ensure that the results provided
by the laboratory are accurate. This is accomplished by establishing
protocols for quality performance in the laboratory.
Quality assurance is a program of written procedures. These procedures
provide a systematic method of monitoring factors that may affect the
quality of laboratory test results. A quality-assurance program also outlines
a system for correcting any deficiencies found during the monitoring
process, reevaluating the systems after corrections, and informing staff of
any procedural changes. Quality-assurance programs are designed to
monitor and evaluate the quality and appropriateness of patient care
services, as well as to resolve problems.
Each laboratory test consists of numerous steps that begin when the
clinician decides to order a test on a patient, and ends when the clinician
uses the results for diagnosis and/or treatment of the patient. Many of these
steps can be implemented in several ways. They are known as variables.
Some variables affect the final test result if they are not carried out in a
specified way. For example, when a specimen is collected, it might go to the
laboratory for processing immediately, or a delay might occur so that it
arrives in the laboratory several hours later. If a transport delay causes the
substance that is to be measured to deteriorate, then testing an old specimen
would give erroneous results. The variable in this example is time.
Criteria should be established to reduce delays in the transport process if
necessary, and monitors should be established to verify that no delays
occur. Establishing the appropriate time frames and monitoring them is a
way to control the time variable.
All variables that affect the testing process should be controlled. This
includes preanalytical, analytical, and postanalytical variables. On the
following pages you will find examples of these variables and a description
of how they should be controlled.
Control of Preanalytical Variables

Variables that should be controlled before testing are called preanalytical
variables. Examples are:
• Selection of appropriate tests.
• Identification and preparation of the patient.
• Collecting and labeling of the sample.
• Transport of the sample to the laboratory; proper handling of the sample
from the time of collection until the time of testing.
• Placement of the sample in the laboratory workload.
• Storage of reagents used to perform tests.
2 Statistical Tools of Quality Control: Training Module Part No. 7B6006

Quality Assurance
• Periodically verifying that procedure manuals are complete and up to

date.
Preanalytical variables should be monitored periodically to assure that they
are being carried out properly.
Control of Analytical Variables

Variables that affect the outcome of the result while the actual test
procedure is in progress are analytical variables. They include:
• Quality of the water supply.
• Stability of the electrical power.
• Preparation of the reagents.
• Upkeep of the instruments and equipment.
• Use of the proper calibration materials.
• Regulation of the temperature in instruments with temperature control
compartments.
• Observance of the steps outlined in the laboratory procedure manual for
each test performed.
• Recording of patient results.
Monitoring the analytical variables and correcting them as needed helps
ensure quality test results. Many other variables may be involved in the
analytic part of the testing process, but may be different for each testing
method.
Control of Postanalytical Variables

Postanalytical variables can impact the reported result once the test
procedure is completed. They include:
• Checking the accuracy of calculations used in reporting patient results.
• Reporting the patient’s results to the doctor.
• Establishing and using appropriate “reference” or “normal” range.
• Immediate reporting of critical results to the doctor.
• Filing and storing the patients’ reports.
It is very important that the laboratory staff communicate with the health
care team of doctors, nurses, phlebotomists, and messengers to achieve the
goals outlined in a quality-assurance program.
A good quality-assurance program systematically checks each of the items
that may cause errors in the system. Some items, such as the temperatures
of water baths, are monitored daily, so that the problems might be corrected
before testing patient samples. Other items, such as the electrical power and
water quality, may be evaluated periodically. A quality-assurance program
should be designed so that all factors that affect the test system are
evaluated over an established time frame.

Quality Control
Quality Control
The quality-control program is a sub-unit of the quality-assurance program.
It focuses on the actual testing process in the laboratory. This program
includes a system designed to monitor the accuracy and precision of the
testing process. These terms will be described in more detail later in this
module.
Quality-control procedures are used to monitor performance of analytical
methods. Therefore, quality control is most useful to identify if the
analytical system is stable and free of significant error. If problems are
identified, corrective action should then be taken to solve the problem,
thereby enabling you to produce accurate patient reports.
Quality-Control Samples
Control procedures use quality-control samples to determine if the method
is working properly. The quality-control sample is a sample with known
values of the analyte being tested. These samples of known value have been
analyzed many times. Based upon the results obtained from the repeated
analysis of quality-control materials, a range of expected values for each
analyte is developed. If the quality-control samples fall outside this
expected range of values, an error may have developed in the test system
and patient sample results may be incorrect.
Evaluation of quality-control values gives you confidence that the patient
results are accurate before reporting them to the physician. These ranges for
each analyte in the control sample are calculated using statistical formulas.
Statistical Concepts
Statistical concepts serve as the foundation for the interpretation of quality-
control results that provide important information about whether or not the
test system is operating correctly. Any series of measurements can be
evaluated by computing and interpreting the statistics from that set of
measurements.
Two important concepts used to assess quality control are accuracy and
precision.
Accuracy and Precision

Accuracy
If a calculated average and the true value are identical or very close, it can
be said that the measurements used to calculate the average are accurate.
Therefore, let us define an accurate series of measurements as one that
produces an average that is equal to or is very close to the true value.

Quality Control
Calculating the average of the quality-control measurements is the first step

in establishing expected ranges.
Precision
Repeated measurements do not usually produce identical results. It is
normal to have a certain degree of fluctuation in repeated measurements.
This degree of variability is known as precision.
To illustrate the relationship between accuracy and precision, consider the
example in Figure 1.
Accuracy and Precision
Accurate Inaccurate Inaccurate

Precise Precise Imprecise
Figure 1. Accuracy and Precision
Look at this as a target and consider each bullet hole as a measurement. The
center of the bull’s-eye represents the true value of the thing that is
measured. When the measurements fall within the bull’s-eye, they are
considered to be accurate. If the bullet holes are clustered close together, the
measurements are considered to be precise.
In the target on the left, all of the results are within the bull’s-eye and are
clustered close together. These results represent a measurement that is both
accurate and precise. That is, they hit the bull’s-eye (true value) with little
variability.
In the center target all of the results are clustered within a small area but
they are not within the bull’s-eye. These results represent precision without
accuracy.

Quality Control
The target on the right represents a group of measurements that are neither
precise nor accurate. They neither hit the bull’s-eye nor cluster together.
These results are said to be both inaccurate and imprecise.
Statistical Measure of Accuracy

To illustrate this point, suppose 22 people measured one person’s height to
the nearest tenth of an inch. They probably would get several different
measurements, even if they used the same yardstick.
For example, if the person’s true height is 65.4 inches, the measurements
might range from 65.1 inches to 65.7 inches, as illustrated in Figure 2:
Variation in Height Measurement

Measurement Number Height (inches)
1 65.3
2 65.1
3 65.4
4 65.3
5 65.6
6 65.4
7 65.3
8 65.5
9 65.4
10 65.6
11 65.4
12 65.4
13 65.5
14 65.2
15 65.4
16 65.3
17 65.7
18 65.5
19 65.5
20 65.2
21 65.4
22 65.4
Figure 2. Variation in Height Measurement
Actually, some variation would occur even if the same person measured the
height of an individual 22 times.
A certain degree of variability is allowed in the clinical laboratory, but you
must learn to distinguish between acceptable variation and test errors, and
establish parameters to distinguish between the two.
In order to do this, it is important to learn how to do some necessary
computations. Applying these computations will help you distinguish
between acceptable variation and error.
How may times does each measurement appear in Figure 2?

Quality Control
Height measurement # of Occurrences (frequency)

65.1 I
65.2 II
65.3 IIII
65.4 IIII III
65.5 IIII
65.6 II
65.7 I
The measurements shown in Figure 2 are represented on the graph in

Figure 3:
10
MEAN
8
Frequency
0
65 65.1 65.2 65.3 65.4 65.5 65.6 65.7 65.8
Height (inches)
Figure 3. Frequency Distribution of Height Measurements
The individual plotted points in Figure 3 represent the height measurement

(x-axis) and the frequency of the measurement (y-axis). If many
measurements are available to plot, a bell-shaped curve results, as shown by
the solid line in Figure 3. If only a few data points are available, a straight
line results, as shown by the dotted line.
As shown on the graph, the value of 65.4 inches occurred 8 times, while the
value of 65.2 occurred only 2 times.
Notice that the measurements are symmetrically distributed. This means
that if the graph were cut in half, both halves would be almost identical.
Also notice the bell shape of the graph. Most of the measurements tend to
cluster around a central value. This is called the mean. If the results of a set
of measurements have a specific distribution around the mean, as shown
here, then they are said to have a normal distribution. There are additional
statistical values that can describe a normal distribution that will be
reviewed in later pages.

Quality Control
All of the statistical methods that will follow assume that data being
analyzed will produce a normal distribution or bell shaped-curve.
It is not usually necessary to actually plot frequency distributions for the
purposes outlined in this module. However, you should understand the
concept of frequency distribution. In our example, the majority of
measurements are clustered around a central point, in this case, 65.4 inches.
Assuming that each person used an accurate yardstick and proper
technique, this central point should be close to the true height of the
individual.
Central Tendency and Mean

The term associated with a clustering of measurements is called central
tendency.
The most commonly used measure of central tendency is the arithmetic
average or mean. The mean is an estimate of the true value.
Let’s examine how the concept of central tendency relates to laboratory
values.
Calculation of the Mean
Glucose Results
Day Glucose (mg/dL)
1 83 (x1)
2 89 (x2)
3 85
4 85
5 84
6 87
7 85
8 86
9 85
10 84
11 88
12 81
13 86
14 88
15 82
16 87
17 82
18 83
19 80
20 90 (x20)
Figure 4. Glucose Results
Figure 4 shows glucose results from the repeated analysis of one quality-
control sample. This sample was analyzed once a day for 20 days. From

Quality Control
these measurements, we can calculate the mean. Remember that the mean,
a measure of central tendency, is another term for average.
To calculate the mean or average, divide the sum of the measurements by
the total number of measurements:
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 )
----------------------------------------------------------------------------- = X = mean
n
where: n = total number of measurements
X 1 refers to the glucose result obtained on Day 1.
X 2 refers to the glucose result obtained on Day 2, and so on,
for 20 days.
X is the symbol for mean.
Let’s calculate the mean for the data in Figure 4.
1. Add all of the results (measurements). Glucose results are commonly
reported in units of mg/dL.
In this example, the sum of the 20 glucose results is 1700 mg/dL.
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 ) = 1700 mg/dL
2. Divide this sum by the total number of measurements (n), in this

example there are 20.
The number, 1700 divided by 20 is equal to 85.
1700
------------ = 85
20
Therefore the mean of the 20 glucose measurements is 85 mg/dL.
or
X = 85 mg/dL
Practice Exercises 1 and 2 will allow you optional practice calculating

the mean for two sets of quality-control data. The answers are provided
in the answer key at the end of this module.
NOTE: VITROS Chemistry Systems can calculate means and standard
deviations for collected quality-control data in the quality-control
software package! Therefore the mean is calculated for you.
Statistical Measures of Precision

To establish limits of acceptability for single observations of control
measurements in the laboratory, two related measures are calculated for

Quality Control
each set of quality-control materials. The control is tested at least 20 times
over a span of 20 to 30 days. Then the parameters are calculated using the
20 values obtained.
Important note: You can get a good estimate of precision by using 20 to 30
points. However, you do not need 20 to 30 points to get a good estimate of
the mean. Two to five points may provide a good estimate of the mean.
The parameters that provide information about the reproducibility or
precision of the measurements are:
• Variance
• Standard Deviation
Let’s look at them individually to understand what they mean.
Standard Deviation
Standard Deviation
Line A
Frequency
Line B
X
Measurement Units
Line A = Small SD, more precise
Line B = Large SD, less precise
Figure 5. Standard Deviation
Variance is a measure of variability. It is used to calculate standard

deviation by taking the square root of it. The formula for variance is
explained on the following page.
The value of the standard deviation is related to the spread of the
measurement values. The greater the spread of values around the mean, the
larger the standard deviation. Large standard deviations indicate less
precision.
The shape of the curve is established by the distribution of the
measurements. In Figure 5, notice that Line B has a wide distribution

Quality Control
compared to Line A. Line B has a greater standard deviation than Line A.

Line A has a smaller standard deviation because the values are closer to the
mean.
Calculation of Standard Deviation(s)

Figure 6 shows the formulas for the calculation of the variance and standard
deviation:
Calculating Standard Deviation
Variance = s = ∑ (Xi - X)
2
2
n-1
√
2
∑ (Xi - X)
Standard Deviation = s =
n-1
Where ∑ = the sum of

Xi =
each measurement
X = the mean
s = standard deviation
n = total number of measurements
√ = square root
Figure 6. Example of Standard Deviation Calculation
Note that standard deviation is the square root of the variance.

Recall that the symbol X refers to the mean.
The symbol Σ stands for “the sum of.”
The total number of measurements is represented by n in this formula.
The notation for the variance is s2. Variance is not used to establish the
quality-control limits. The standard deviation is the square root of the
variance. Its notation is a lower case s.
To calculate the variance, the mean is subtracted from each result and the
value obtained is squared. All of the squared values are then added together
and divided by the number of measurements minus one.
Many scientific calculators have been programmed to perform standard
deviation calculations according to this formula once all the values have
been entered. If you are using a calculator, verify you are using the sample
SD calculation that uses “n–1” as the denominator. Remember, the
VITROS Chemistry Systems can collect the data and calculate the means

Quality Control
and standard deviations! However, the following pages will demonstrate

how this calculation can be done by hand.
Calculating Standard Deviation

(1) (2) (3)
2
Results (Xi ) mg/dL (Xi X) (Xi X)
83 -2 4
89 +4 16
85 0 0
85 0 0
84 -1 1
87 +2 4
85 0 0
86 +1 1
85 0 0
84 -1 1
88 +3 9
81 -4 16
86 +1 1
88 +3 9
82 -3 9
87 +2 4
82 -3 9
83 -2 4
80 -5 25
90 +5 25
X = 85 mg/dL ∑ (Xi X) 2 = 138
Figure 7. Calculating Standard Deviation
Figure 7 is an example of the steps to be followed to calculate the standard

deviation of the glucose measurements previously described in Figure 4.
Let’s follow the steps to calculate the standard deviation:
1. List the values obtained by each of the 20 measurements as shown in
column 1.
2. Calculate the mean of the 20 measurements.
3. Subtract the mean from each value as shown in column 2.
For example, the difference between the first value, 83, and the mean,
85 = –2.
Note that the values that are less than the mean result in negative
numbers in column 2 and the results that are greater than the mean are
positive numbers.
4. Square each of the numbers obtained in column 2. The results of this
step are shown in column 3.
Remember, to square a number, you multiply the number by itself.
Two negative numbers multiplied together equal a positive number.
Consequently, all of the numbers in column 3 are positive numbers.

Quality Control
5. Add all of the numbers in column 3. The total of the squared

differences in column 3 is 138 mg2/dL2.
Σ ( X i – X )2 = 138
6. Divide the total of the squared differences by one less than the total of
measurements.
In this example the total number of measurements (n) is 20.
Therefore, n – 1 = 19.
ΣX i – X2
Variance = S2 = -------------------- = 138
--------- = 7.3
n–1 19
7. Obtain the square root of the result, 7.3 mg2/dL2:
7.3 = 2.7 mg/dL
Here is the complete flow of this calculation.
Σ ( X i – X )2 138
s = ------------------------- = --------- = 7.3 = 2.7 mg/dL
n–1 19
2.7 mg/dL is the standard deviation for this set of measurements.
In calculating the standard deviation, you can see that the value is an
indication of the spread of values about the mean.
Refer to Figure 8. The standard deviation can be used to calculate limits that
can be expected to contain a certain percentage of the results. For example,
we can expect that the region from (X -2 SD) to (X+2 SD) would contain
approximately 95.4% of all results.

Quality Control
Frequency Distribution
-2 SD Mean +2 SD
Figure 8. Frequency Distribution
Practice Exercises 3 and 4 will allow you to practice calculating SDs.

Follow the sequence of calculations exactly as described to obtain the
correct values.
Standard Deviation and Probability
X
Frequency
68.2%
A.
95.4%
B.
C. 99.7%
MEAN
-3s -2s -1s +1s +2s +3s
X
Figure 9. Standard Deviation and Probability
Figure 9 shows a standard frequency plot of a theoretical set of data points.

Quality Control
Once the standard deviation is known, the likelihood or probability that a

particular measurement will fall within specified ranges can be predicted.
The three ranges shown on this graph labeled A, B, and C correspond to ±1
standard deviation, ±2 standard deviations, and ±3 standard deviations
about the mean.
A frequency plot that looks like a symmetrical, bell-shaped curve can be
described as a “normal” distribution. A normal distribution can be
described by the mean and the standard deviation. Statisticians tell us that
68% of all data points fall within ±1 standard deviation for data that can be
described by a symmetrical, bell-shaped “normal” curve. Additionally,
about 95% of all data points will fall within ±2 standard deviations, and
99.7% of all data points will fall within ±3 standard deviations.
Most of the measurements are expected to be within this ±2 standard
deviation range. If 100 glucose measurements of the same control material
were taken, approximately 95 of them would fall within the ±2 standard
deviation range, while approximately 5 measurements would fall out of this
range.
If the ±2 standard deviation criteria is used to evaluate the testing system,
you can have 95% confidence that the measurement is correct when the
quality-control measurements fall within this range.
Which SDs Should the Lab Use?

There are three sources for SDs that the lab can use.
• If you choose to use medically useful criteria, you will need to have the
medical director of the laboratory establish precision goals and the SD
guidelines and rules that are used in determining when a quality-control
failure is large enough to require action. Medically useful criteria are
often described by the term coefficient of variation (CV). CV is
defined as :
SD
------- × 100%
X
It is a convenient way of expressing precision in a manner applicable to
any concentration level.
• If you are using VITROS Performance Verifiers as a daily control fluid,
we recommend the within-lab SDs from the Performance Verifier Assay
Sheets. These values incorporate the expected sources of variation that
can occur when you are using one lot number of slides overtime.
• If you choose to use an SD that represents actual performance in your
laboratory (whether using VITROS Performance Verifiers or another
manufacturer’s fluid) you must be sure that this estimate includes all the
sources of variability that can occur within a slide lot.

Quality-Control Materials
Up to this point you have learned the calculations to estimate statistical
parameters used in most quality-control systems. The remainder of this module
will focus on the composition and use of quality-control materials and the
recording and interpretation of quality-control data.
Selection and Use

Recall that the goal of quality control is to monitor the accuracy and precision of
the test system. This is accomplished by including quality-control samples along
with patient samples in the test run or by following instructions from instrument
or reagent manufacturers. Quality-control samples for tests run on VITROS 250,
350, 950 an 5,1 FS Chemistry Systems should be analyzed every 24 hours.
However, your lab may choose to run quality-control samples more frequently.
Quality-control materials are designed to duplicate the chemical and physical
properties of the patient samples being tested. Therefore:
• The basic composition (called “matrix”) of a control should be similar to that
of the patient samples. Controls for urine are usually a different composition
than controls for serum. The controls are made to “look like” the fluid they are
meant to “test like” (that is, urine controls have less protein and are less
viscous than serum controls).
The value for a urine analyte will be different from a value for the same analyte
found in plasma or serum. For example, sodium levels may be determined in both
serum and urine, but each usually requires its own control material.
• Either the laboratory procedure manual or the manufacturer’s method sheet
should specify the appropriate type of quality-control material for each test.
• Because many quality-control samples are composed of human serum or
urine, all should be handled as if they contain biohazardous material.
Two major types of quality-control materials may be purchased:
Assayed controls are control materials that have assigned values for the analytes
contained in the control. VITROS Performance Verifiers have been assayed by
Ortho-Clinical Diagnostics and multiple laboratories. A within-lab SD and range
of means describes expected system performance. Other manufacturers may
provide controls assayed for VITROS Chemistry Systems. Please note that we
have no control over other manufacturer’s value assignment processes, the
timeliness of the processes, or how applicable or adequate those processes may
be for VITROS Chemistry Systems.
Unassayed controls have not had repeated measurements performed by the
manufacturer. They do not have assigned values.
If unassayed controls are used, the mean must be established by the laboratory
before use in a daily quality-control program. This requires multiple analysis of
the control material before it is put into use.

Both types of commercial quality-control materials may be prepared in

large batches by the manufacturer. Each batch is assigned a specific
number. These numbers are referred to as lot numbers.
Important points to remember about control lots used for daily quality
control:
• Many labs prefer to have one year’s supply of each control on hand.
They should be stored according to the manufacturer’s directions.
• Different lot numbers of control fluid must not be mixed together
because each has a different concentration of analytes.
• The mean must be calculated or evaluated before a new lot number of
quality-control material is used because each lot number has its own
concentration of the analyte.
The mean should also be reassessed when new lot numbers of reagents are
put into use. For VITROS Systems, you should re-evaluate means when
new lot numbers of slides, VITROS Reference Fluid, or Immuno-Wash
Fluid are put into use.
• Use only the lot number of quality-control material for which you have
established the mean and standard deviation.
• SDs should be derived from the laboratory’s established precision goals
for each analyte. Some manufacturers provide recommended SDs for
each analyte. For VITROS Chemistry Systems, this information is
provided on the Performance Verifier sheets.
Types of Quality-Control Materials
Quality-Control
Materials
Liquid Lyophilized
Form “Freeze Dried”
Figure 10. Types of Quality-Control Materials
Commercially available quality-control materials (Figure 10) can be liquid

or lyophilized. Liquid controls may contain stabilizers such as ethylene
glycol, a chemical preservative. Controls made with ethylene glycol are

incompatible with VITROS Slides and should not be used on VITROS

Chemistry Systems.
Lyophilized or freeze-dried controls are preserved by dehydration, which is
the removal of water. During lyophilization, the material is quickly frozen
and then placed under low pressure to remove the water.
Reconstitution of Quality-Control Materials

To prepare lyophilized quality-control materials for use, the water removed
in the manufacturing process must be replaced. The addition of purified
water, or other diluent, as specified by the manufacturer, is called
reconstitution.
Instructions in the manufacturer’s product insert must be followed exactly
when reconstituting quality-control material.
• The correct diluting fluid and proper pipetting techniques using
appropriate pipettes must be followed since the final concentrations of
analytes will be affected by the volume of the diluent. VITROS fluids, as
well as other brands, are optimized when made up with reusable
volumetric glass pipettes.
• The vial should be opened slowly and carefully to avoid loss of any
contents.
• After adding diluent to the lyophilized quality-control material, the
laboratorian’s initials and the date the control was reconstituted should
be written on the bottle.
• The lyophilate should be dissolved in the diluent for the time
recommended by the manufacturer and then gently mixed before use.
• Use and store the reconstituted fluid according to the manufacturer’s
directions provided on an insert sheet.
If the quality-control material is prepared incorrectly, results obtained may
be incorrect. The test method or equipment may not be properly evaluated
and delays and errors in reporting patient results are possible.
Storage of Quality-Control Materials

It is important to store reconstituted quality-control materials at the
manufacturer’s recommended temperature to prevent deterioration. Most
reconstituted lyophilized quality-control materials should not be left at
room temperature (18°–28°C; 64°–82°F) for long periods of time, as some
of the constituents may deteriorate.
It is a good practice to remove a small quantity, called an aliquot, of quality-
control material for testing and then replace the vial in a refrigerator as soon
as possible. For control fluids run on VITROS Chemistry Systems, the
control fluids should be transferred into a sample cup and capped. Restore
the remaining control fluid to the appropriate storage, usually a refrigerator
(2°–8°C; 36°–46°F).

Allow the quality-control fluid in the cup to come up to room temperature

(following the recommendation of the manufacturer), and then run the fluid
on the analyzer. Cold fluids are more viscous and will spread less evenly on
the slide. VITROS slides are optimized for use with room temperature
fluids.
Also, be sure to use reconstituted control materials within the
manufacturer’s specified time period for each constituent.
Level I and Level II Quality-Control Materials

IMPORTANT: Some VITROS Performance Verifiers have 3 levels. Mean
and standard deviation must be established for all levels in
use.
A mean and standard deviation must be established or available for each
concentration and lot number of control material. Traditionally, two
different quality-control materials with different concentrations of analytes
are analyzed every 24 hours as long as patient samples are being tested.
Controls commonly have values assigned so that one level has the
concentration of analytes at or near the reference interval for those analytes.
A second level has concentrations at or near concentrations associated with
the presence of disease or a medical decision point.
Using two levels allows you to evaluate the performance of the analyzer
system at different concentration ranges. That is, at least two levels are
required to evaluate the accuracy and precision of the test system.

What is a Quality-Control Rule?
What is a Quality-Control Rule?

A quality-control rule is used to help you decide when your analyzer system
is performing within expectations. Expectations can be determined by you
or by the manufacturer of the system. When the system is performing as
expected, you have confidence in the patient results reported by the system.
Quality-control rules describe statistical procedures used to interpret
quality-control data and identify unusual occurrences that characterize
unstable analytical system performance.
Most of the rules use the mean and standard deviation (SD) to describe a
situation. The mean and SD are calculated from quality-control data.
Manufacturers of quality-control (QC) fluids and equipment sometimes
determine mean and SD values that they recommend for use in the lab.
Control rules are written in the following format: “X y.” Control rules
combine either a statistical procedure or the number of observations (X)
and a control limit (y).
For example, “13s” is a control rule where 1 observation of a QC fluid result
(X=1) is reviewed to see if that one result differs from the mean by ±3 SD
(y=3s).
This is the same as using a Levey-Jennings chart with an established mean
and an acceptable range set at ±3 SD around the mean.
Quality-control rules may be used individually or in combination with each
other to determine if the analyzer system is functioning properly. The next
section tells you which quality-control rules are appropriate for use on
VITROS Systems. And after that, we will describe how to use those rules
together to determine whether your VITROS System is functioning
properly.

Individual Quality-Control Rules

The approach presented in this document uses two levels of quality-control
fluid. However, the number and concentration of quality-control fluids
should be selected based on your testing needs.
Quality-control rules are applied to the data of the 2 levels of quality-
control fluid individually and together.
There are three quality-control rules that are appropriate to use on the
VITROS System. They are the 12s rule, the 22s rule, and the 13s rule. Table
1 lists these quality-control rules:
Rule Quality-Control Rule Explanation

12s The 12s rule describes a situation where one control value for any of the two
levels of control is outside the mean ± 2 SD.
13s The 13s rule describes a situation where one control value for any of the two
levels of control is outside the mean ± 3 SD.
22s The 22s rule describes a situation where both control fluids or two consecutive
control values of one fluid are outside the mean by either +2 SD or –2 SD.
Table 1. Quality-Control Rules
The three quality-control rules just described are used together when the
mean value for the quality-control fluids has been established. The diagram
below should help you better understand the three rules.
Three Quality-Control Rules

Control
Data
In Control:
12s No Continue &
Violated?
Accept Results
Yes No
13s No 2 2s
Violated? Violated?
Yes Yes
Out of Control - Investigate
Figure 11. Three Quality-Control Rules

Other Rules
Additional rules, like the Westgard rules, are also available, such as the 41s
rule and the 101s rule. These rules are used only after a 12s violation is
noted. The 41s rule violation occurs when four consecutive control
observations fall on the same side of the mean and all four are outside of
±1 SD. The 101s rule violation occurs when 10 consecutive control
observations fall on the same side of the mean outside of ±1 SD.
These rules are useful when a system is subject to frequent changes, like
frequent calibration. They are very sensitive to small systematic errors.
Because VITROS Chemistry Systems are so infrequently calibrated, using
these rules can cause you to react when nothing is wrong and is not
recommended. Therefore, not all rules are appropriate for VITROS
Systems.
Let’s take a detailed look at the three rules that are appropriate for VITROS
Systems. We will use Levey-Jennings charts to help “visualize” the control
rules, but you do not need these charts to use the rules.
The 12s Rule

The 12s rule is a warning rule. This rule is violated when one of the two
levels of control results falls outside of the mean ±2 standard deviation
range. One control level result is within ±2 SD and the other control level
is outside ±2 SD.
Violation of the 12s rule constitutes a warning and requires evaluation to
determine whether the 13s rule and the 22s rule have also been violated. It
cautions you to watch closely for present or developing sources of error.
The 12s warning rule by itself is not cause for rejecting a run. When this rule
has been violated, the next rule in the sequence must be evaluated.
For example, suppose that the mean of a Level I control material for
cholesterol is 180 mg/dL and the ±2 standard deviation range is from 160
to 200 mg/dL. A result either above 200 mg/dL or below 160 mg/dL would
violate the 12s warning rule.
The 13s Rule

If the 12s warning rule is violated, the 13s rule is evaluated next. The 13s rule
is violated when either result of the two levels of control fall outside of the
mean ±3 standard deviation range for either control level.
When this rule is violated, your system behavior may have changed and
investigation is needed before proceeding. No other rules need to be
evaluated. Remember, 99.7% of the time, results are expected to fall within
the mean ±3 SD range.

If the 13s rule is violated, the method must be evaluated to determine the
source of the error. The problem must then be corrected and the controls
repeated before running patient samples.
The 22s Rule

If the 13s rule is not violated, the 22s rule is evaluated next. The 22s rule can
be violated three different ways (Figures 1 to 3).
1. One QC point of Level I control and one QC point of Level II when
assayed at the same time are greater than their respective 2 SD limits
with same sign (+ or -). In this example, both values are outside the
+2 SD limit.
Rule 2 2s Violation 1
Level I Level II
+2SD +2SD
X X
-2SD -2SD
1 2 1 2
Days Days
Figure 12. Rule 22s Violation 1

2. Two sequential QC points from the same fluid are greater than the
same 2 SD limit. In this example, the 2 recent values for Level I are
outside the +2 SD limit.
Level I Level II
+2SD +2SD
X X
-2SD -2SD
1 2 1 2
Days Days
3. Two sequential QC points, one from Level I and the other from Level
II, are outside of their respective 2 SD limits with the same sign
(+ or -). In this example, the value for Level I on day one is outside
+2 SD limit and the value for Level II on day two is also outside
+2 SD limit.
Level I Level II
+2SD +2SD
X X
-2SD -2SD
1 2 1 2
Days Days

If the 22s rule is violated, your system’s behavior has changed and
investigation is needed. If a problem is identified, it must be corrected and
the controls repeated. The patient results can be reported only when the
controls used in the repeat assay do not violate any of the rules.
Once you have corrected the problem that caused a rule to be violated, the
control values that indicated a problem should not be used to evaluate the
22s status next time controls are assayed (but do not delete from QC
records).
Examples
The following are examples of the evaluation process using these three
control rules.
Control Rule Violations 22s
96 Glucose (Level I)
(+3s)
92
(+2s) X= 84 mg/dL
88
(+1s) s= 4 mg/dL
Glucose 84
mg/dL (X)
80
(-1s)
76
(-2s)
72
(-3s)
0 5 10 15 20 25 30 31
Day of the Month
214 Glucose (Level II)

(+3s)
206
(+2s) X= 180 mg/dL
188 s= 8 mg/dL
(+1s)
Glucose
180
mg/dL (X)
172
(-1s)
164
(-2s)
156
(-3s)
0 5 10 15 20 25 30 31
Day of the Month
Figure 15. Control Rule Violation 22s
In this example (Figure 15), on day 10 of these Levey-Jennings charts, each

quality-control level violated the 12s warning rule. The 13s rule, however,
was not violated. Further inspection shows that the 22s rule is also violated
since both controls are outside of the two standard deviation range.
Remember, one way that a 22s rule violation occurs is when the quality-
control results of both levels either exceed the +2 standard deviation limit
or exceed the -2 standard deviation limit. In this example, the results of both
quality-control levels exceed the +2 standard deviation limit. This method

is considered to be out of control on day 10. Identify and correct the cause
of the rule violation.
Control Rule Violations 13s

13s Violation
96 Glucose (Level I)
(+3s)
92
(+2s) X= 84 mg/dL
88
(+1s) s= 4 mg/dL
Glucose 84
mg/dL (X)
80
(-1s)
76
(-2s)
72
(-3s)
0 5 10 15 20 25 30 31
Day of the Month
214 Glucose (Level II)

(+3s)
206
(+2s) X= 180 mg/dL
188 s= 8 mg/dL
(+1s)
Glucose
180
mg/dL (X)
172
(-1s)
164
(-2s)
156
(-3s)
0 5 10 15 20 25 30 31
Day of the Month
Figure 16. Control Rule Violation 13s
Figure 16 shows Levey-Jennings charts for the results of Level I and

Level II quality-control materials for a glucose method. Days 1 through 3
show both controls within the ±2 standard deviation range, indicating that
the method is in control. However, on day 4 the result for Level I is greater
than the +3 standard deviation limit. Therefore, this violates the 13s rule and
this method is considered to be out of control. Identify and correct the cause
of the rule violation.
To become familiar with using these rules, do Practice Exercises 1
and 2.

Interpreting Quality-Control Results

To interpret your quality-control data, the results on the controls are
compared to a range of acceptable values. The most common criteria used
to define the range of acceptable values are the mean ±2 SD and the mean
± 3 SD. These ranges can be thought of as having an upper and lower limit.
For a control result to be acceptable it must fall within these limits.
Levey-Jennings Chart
Once quality-control ranges are established, each quality-control result
should be recorded and evaluated before patient-test results are reported.
One of the most common manual ways to do this is to plot the control
results on a Levey-Jennings chart. The mean and acceptable ranges of the
quality-control material are plotted against the days of the month as shown
in Figure 17.
Glucose
mg/dL Glucose (Level I)
96
(+3s) X= 84 mg/dL
92 s= 4 mg/dL
(+2s)
88
(+1s)
84
(X)
80
(–1s)
76
(–2s)
72
(–3s)
1 5 10 15 20 25 30 31
Day of the Month
Figure 17. Levey-Jennings Chart
The Levey-Jennings quality-control chart provides a convenient way to

record and visualize quality-control results over a period of time.
Please note that the VITROS Chemistry System can record quality-control
data when the sample is run. You can request printouts of Levey-Jennings
charts from the VITROS quality-control program. See “Calculating
±2 Standard Deviation Ranges” in this section for calculation.
To obtain the values included in the control limit range two mathematical
calculations must be performed. First calculate the mean and select the

standard deviation used. Then subtract multiples of the standard deviation
value from the mean to determine the lower range limit. Then add multiples
of the standard deviation value to the mean to determine the upper range
limit. These two calculated limits identify the standard deviation range.
Actual examples of how to calculate ranges of ±2 and ±3 standard
deviations will be shown in the next section.
When using the Levey-Jennings system, you must develop a Levey-
Jennings quality-control chart for each level of quality-control material for
each method used.
Calculating ±2 Standard Deviation Ranges

To calculate the range of values included in 2 standard deviations for a level
of quality-control material, you use the mean and the standard deviation
obtained from repeated measurements of that control level.
1. Multiply the standard deviation by 2.
2. Subtract that value from the mean to give the lower limit (-2 standard
deviations) and add it to the mean to get the upper limit (+2 standard
deviations.) The values between the lower limit and the upper limit
define the acceptable range of operations. The numbers for the actual
limits are included in this range.
To calculate the range for 3 standard deviations, follow the same procedure
except that the standard deviation value is multiplied by three before being
added to or subtracted from the mean.
Let’s look at the following example of calculating a range of ±2 SD:
1. The mean for a Level I glucose control is 84 mg/dL, and the standard
deviation is 4 mg/dL.
Mean = 84 mg/dL
Standard Deviation = 4.0 mg/dL
2. Multiplying the SD of 4 by 2 = 8 mg/dL.
3. The range that includes ±2 SD is then calculated. This is done by
adding 8 to 84 to determine the upper limit and then subtracting 8
from 84 to determine the lower limit.
Mean +2 standard deviations = upper limit value
84 + (4.0 x 2) =
84 + 8.0 = 92.0 mg/dL
Mean -2 standard deviations = lower limit value
84 - (4.0 x 2) =
84 - 8.0 = 76.0 mg/dL

Therefore, the range that includes 2 standard deviations would include

all the values from 76 mg/dL to 92 mg/dL.
Mean ±2 Standard Deviation = range of values
76.0 to 92.0 mg/dL
Remember that 4.5% of the time, or approximately 1 in 20 times, a quality-
control value will fall outside of this range due to chance (even when there
is nothing wrong).
Figure 18 represents an analyte with a mean of 84 and 1 SD of 4.0 mg/dL.
The dotted line on either side of the mean represents the +2 and –2 standard
deviation limits. In the example, the standard deviation = 4.0 and 2 x 4.0 =
8. Thus, the line above the mean is 84 + 8, or 92, and the line below the
mean is 84 – 8, or 76.
Standard Deviation Range
+2 SD (92.0 mg/dL)
Test value
mg/dL
Mean (84.0 mg/dL)
-2 SD (76.0 mg/dL)
Days
Figure 18. Standard Deviation Range
To become familiar with calculating standard deviation ranges, do

Practice Exercises 5 and 6.

Figure 19 shows the relationship between a normal distribution of repeated
measurements and a Levey-Jennings control chart.
Glucose
mg/dL
Glucose (Level I)
-1s MEAN +1s +2s +3s

96
(+3s) X= 84 mg/dL
92 s= 4 mg/dL
(+2s)
88
95.4%
99.7%
68.2%
(+1s)
(X)
84
(X)
80
-2s
(-1s)
76
-3s
(-2s)
72
(-3s)
0 5 10 15 20 25 30 31
Day of the Month
Figure 19. Using Levey-Jennings Chart
The normal distribution is rotated 90° so that the relationships between the
mean and standard deviations on the two charts can be seen more readily.
The distribution obtained by repeatedly testing the Level I control for
glucose is shown on the left side of Figure 19. The probability distribution
is displayed in this way to illustrate that the mean and standard deviation
ranges shown are also represented on the y-axis of a Levey-Jennings chart.
The right portion of the figure shows a Levey-Jennings chart using the same
glucose data.

Summary
Summary
This module presented the basic math concepts used for quality control in
the clinical laboratory. You now know how to calculate a mean and a
standard deviation. You know how to create a Levey-Jennings chart, and
you have learned basic information about quality-control fluids.
In the next module you will learn more about interpreting Levey-Jennings
plots and determining the proper mean and SD to use in setting up these
plots.
This concludes “Statistical Tools of Quality Control.” You may now
complete the Self-Evaluation.

Practice Exercises
Practice Exercises
Practice Exercise 1
Using the formula below, calculate the mean for the following data:
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 )
----------------------------------------------------------------------------- = X = mean
n
Day Calcium (mg/dL)

1 9.7
2 9.3
3 10.0
4 9.8
5 9.7
6 9.9
7 9.6
8 10.1
9 9.8
10 9.6
11 9.9
12 9.7
13 9.8
14 9.8
15 9.5
16 9.4
17 10.0
18 9.9
19 9.7
20 9.7
Fill in the blanks:

1. The sum of all measurements is
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 ) = ____________________
2. The total number of measurements is n = ________________
The calculated mean is X = ___________________.

Check your work with the answer key.

Practice Exercises
Practice Exercise 2
Calculate the mean for the following new data:
Day Creatinine (mg/dL)
1 3.6
2 3.8
3 3.4
4 3.9
5 4.0
6 4.1
7 4.1
8 4.2
9 3.5
10 4.1
11 4.2
12 4.1
13 3.9
14 3.8
15 4.1
16 3.6
17 3.9
18 4.2
19 4.3
20 4.0
1. The mean or X = _________________

Practice Exercises
Practice Exercise 3
Using this formula, calculate the mean and standard deviation for the
following data:
Σ ( X i – X )2
Standard Deviation = s = -------------------------
n–1
Where:
X i = each measurement
X = the mean
Σ = “the sum of”
= square root
Day Glucose (mg/dL)

1 205
2 192
3 190
4 213
5 201
6 194
7 199
8 215
9 216
10 192
11 194
12 198
13 202
14 205
15 210
16 201
17 194
18 197
19 204
20 198
1. The mean or X = _________________

Practice Exercises
Calculate the standard deviation. Fill in the blanks and complete the table
below.
Column 1 Column 2 Column 3
( Xi – X ) ( X i – X )2
1 205
2 192
3 190
4 213
5 201
6 194
7 199
8 215
9 216
10 192
11 194
12 198
13 202
14 205
15 210
16 201
17 194
18 197
19 204
20 198
2. Subtract the mean from each value to obtain the ( X i – X ) values and
record each value in column 2.
3. Square each of the numbers obtained in column 2 to obtain the

( X i – X )2 values and record these in column 3.
4. Add all of the numbers in column 3 to obtain the sum of the squared
differences.
Σ ( X i – X )2 = _______mg2/dL2.
5. Divide the sum by one less than the total of measurements to obtain
your variance or s2
The total number of measurements (n) is __________.
Therefore, n – 1 = __________________.
Σ ( X i – X )2
----------------------- = ________= s2
n–1

Practice Exercises
6. Take the square root of the result in step 5 to obtain the standard
deviation:
s 2 = ______mg/dL = s

Practice Exercises
Practice Exercise 4
Calculate the mean and standard deviation for the following data.
( X i – X )2
Urea Nitrogen
Day
(mg/dL) ( Xi – X )
1 53
2 50
3 55
4 62
5 51
6 51
7 55
8 50
9 62
10 53
11 54
12 55
13 57
14 48
15 50
16 56
17 52
18 57
19 58
20 51
1. The mean or X = _________________
2. The standard deviation = ____________________________.

Practice Exercises
Practice Exercise 5
Using the formula, calculate the mean, standard deviation, and the ±1, 2,
and 3 standard deviation ranges for Level I quality-control data shown
below.
Σ ( X i – X )2
n–1
Where:
X = the mean
= square root
( X i – X )2
Total Bilirubin
Day
(mg/dL) ( Xi – X )
1 0.7
2 0.4
3 0.5
4 0.6
5 0.6
6 0.7
7 0.5
8 0.6
9 0.7
10 0.8
11 0.6
12 0.7
13 0.5
14 0.7
15 0.7
16 0.8
17 0.9
18 0.7
19 0.6
20 0.7
1. The mean of X = _________________

Practice Exercises
To calculate the standard deviation and the ±1 standard deviation range,

follow steps 2 through 6 and fill in the blanks.
2. Calculate the sum of the ( X i – X )2 values.

Σ ( X i – X )2 = _______________.
3. Divide the sum of ( X i – X )2 by one less than the total of
measurements to obtain your variance or s2.
The total number of measurements (n) is ______.
Therefore, n – 1 = _________.
Σ ( X i – X )2
- = _______________ = s2 = variance.
----------------------
n–1
4. Take the square root of the variance to obtain the standard deviation:
s 2 = ______ = s
5. You now have the mean and the standard deviation. Calculate the ±1
standard deviation range by filling in the blanks:
Mean = _______.
Standard deviation = _________.
Multiply the standard deviation by 1:
standard deviation x 1 =__________ = 1 standard deviation
Calculate your ±1 standard deviation range by adding 1 standard
deviation to the mean to obtain the upper limit and then subtracting
1 standard deviation from the mean to obtain the lower limit:
mean +1 standard deviation =_________.
mean –1 standard deviation =_________.
Therefore the ±1 standard deviation range = _______ to _________.
6. Now calculate the ±2 and ±3 standard deviation ranges:

The ±2 standard deviation range = _______ to _________.
The ±3 standard deviation range = _______ to _________.

Practice Exercises
Practice Exercise 6
Calculate the mean, standard deviation, and ±1 SD range for this one level
of quality control.
Day Cholesterol (mg/dL)
1 372
2 360
3 348
4 381
5 365
6 368
7 385
8 373
9 370
10 359
11 355
12 389
13 364
14 371
15 364
16 352
17 378
18 344
19 375
20 371
1. The mean or X = _________________
The standard deviation = _______________.
The ±1 standard deviation range = ________ to _________.

Answer Keys
Answer Keys
Practice Exercise 1
Using the formula below, calculate the mean for the following data:
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 )
----------------------------------------------------------------------------- = X = mean
n
Day Calcium (mg/dL)

1 9.7
2 9.3
3 10.0
4 9.8
5 9.7
6 9.9
7 9.6
8 10.1
9 9.8
10 9.6
11 9.9
12 9.7
13 9.8
14 9.8
15 9.5
16 9.4
17 10.0
18 9.9
19 9.7
20 9.7
Fill in the blanks:

1. The sum of all measurements is
( X 1 ) + ( X 2 ) + ( X 3 ) + … + ( X 20 ) = 194.9
2. The total number of measurements is n = 20
The calculated mean is X = 9.7 mg/dL.

Answer Keys
Practice Exercise 2
Calculate the mean for the following new data:
Day Creatinine (mg/dL)
1 3.6
2 3.8
3 3.4
4 3.9
5 4.0
6 4.1
7 4.1
8 4.2
9 3.5
10 4.1
11 4.2
12 4.1
13 3.9
14 3.8
15 4.1
16 3.6
17 3.9
18 4.2
19 4.3
20 4.0
78.8
1. The mean or X = ---------- = 3.9
20

Answer Keys
Practice Exercise 3
Using this formula, calculate the mean and standard deviation for the
following data:
Σ ( X i – X )2
n–1
Where:
X = the mean
= square root
Day Glucose (mg/dL)

1 205
2 192
3 190
4 213
5 201
6 194
7 199
8 215
9 216
10 192
11 194
12 198
13 202
14 205
15 210
16 201
17 194
18 197
19 204
20 198
1. The mean or X = 201 mg/dL.

Answer Keys
Calculate the standard deviation. Fill in the blanks and complete the table
below.
Column 1 Column 2 Column 3
( Xi – X ) ( X i – X )2
1 205 4 16
2 192 –9 81
3 190 –11 121
4 213 12 144
5 201 0 0
6 194 –7 49
7 199 –2 4
8 215 14 196
9 216 15 225
10 192 –9 81
11 194 –7 49
12 198 –3 9
13 202 1 1
14 205 4 16
15 210 9 81
16 201 0 0
17 194 –7 49
18 197 –4 16
19 204 3 9
20 198 –3 9
2. Subtract the mean from each value to obtain the ( X i – X ) values and
record each value in column 2.
3. Square each of the numbers obtained in column 2 to obtain the
( X i – X )2 values and record these in column 3.
4. Add all of the numbers in column 3 to obtain the sum of the squared
differences.
Σ ( X i – X )2 = 1156 mg2/dL2.
5. Divide the sum by one less than the total of measurements to obtain
your variance or s2.
The total number of measurements (n) is 20.
Σ ( X i – X )2
----------------------- = 1156
------------ = 60 = s2
n–1 19

Answer Keys
6. Take the square root of the result in step 5 to obtain the standard
deviation:
s 2 = 7.8 mg/dL = s

Answer Keys
Practice Exercise 4
Calculate the mean and standard deviation for the following data.
( X i – X )2
Urea Nitrogen
Day
(mg/dL) ( Xi – X )
1 53 -1 1
2 50 -4 16
3 55 1 1
4 62 8 64
5 51 -3 9
6 51 -3 9
7 55 1 1
8 50 -4 16
9 62 8 64
10 53 -1 1
11 54 0 0
12 55 1 1
13 57 2 4
14 48 -6 36
15 50 -4 16
16 56 2 4
17 52 -2 4
18 57 3 9
19 58 4 16
20 51 -3 9
2. The standard deviation = 3.9 mg/dL.

Answer Keys
Practice Exercise 5
Using the formula, calculate the mean, standard deviation, and the ±1, 2,
and 3 standard deviation ranges for Level I quality-control data shown
below.
Σ ( X i – X )2
n–1
Where:
X = the mean
= square root
( X – X )2
Total Bilirubin
Day
(mg/dL)
X = 0.65
1 0.7 .05 .0025
2 0.4 –.25 .0625
3 0.5 –.15 .0225
4 0.6 –.05 .0025
5 0.6 –.05 .0025
6 0.7 .05 .0025
7 0.5 –.15 .0225
8 0.6 –.05 .0025
9 0.7 .05 .0025
10 0.8 .15 .0225
11 0.6 –.05 .0025
12 0.7 .05 .0025
13 0.5 –.15 .0225
14 0.7 .05 .0025
15 0.7 .05 .0025
16 0.8 .15 .0225
17 0.9 .25 .0625
18 0.7 .05 .0025
19 0.6 –.05 .0025
20 0.7 .05 .0025
1. The mean or X = .65 mg/dL.

Answer Keys
To calculate the standard deviation and the ±1 standard deviation range,
follow steps 2 through 6 and fill in the blanks.
2. Calculate the sum of the ( X i – X )2 values.

Σ ( X i – X )2 = 0.27
3. Divide the sum of ( X i – X )2 by one less than the total of
measurements to obtain your variance or s2.
The total number of measurements (n) is 20.
Σ ( X i – X )2
- = 0.0142 (use .014) = s2 = variance.
----------------------
n–1
4. Take the square root of the variance to obtain the standard deviation:
s 2 = 0.118 (round to 0.12) = s
5. You now have the mean and the standard deviation. Calculate the ±1
standard deviation range by filling in the blanks:
Mean = 0.65.
Standard deviation = 0.12.
Multiply the standard deviation by 1:
standard deviation x 1 = 0.12 = 1 standard deviation
Calculate your ±1 standard deviation range by adding 1 standard
deviation to the mean to obtain the upper limit and then subtracting
1 standard deviation from the mean to obtain the lower limit:
mean +1 standard deviation = 0.77.
mean –1 standard deviation = 0.53.
Therefore the ±1 standard deviation range = 0.53 to 0.77.
6. Now calculate the ±2 and ±3 standard deviation ranges:

The ±2 standard deviation range = 0.41 to 0.89.
The ±3 standard deviation range = 0.29 to 1.01.

Answer Keys
Practice Exercise 6
Calculate the mean, standard deviation, and ±1 SD range for this one level
of quality control.
Day Cholesterol (mg/dL)
1 372
2 360
3 348
4 381
5 365
6 368
7 385
8 373
9 370
10 359
11 355
12 389
13 364
14 371
15 364
16 352
17 378
18 344
19 375
20 371
The standard deviation = 11.9 mg/dL.

The ±1 standard deviation range = 355 to 379.

Self-Evaluation
Self-Evaluation
To complete this self-evaluation:
• Fill in the blanks to correctly complete the following statements.
• Calculate the mean, standard deviation, and the ±1, ±2, and ±3SD ranges
for 20 days of glucose data.
• Plot data on a Levey-Jennings chart and interpret the chart results.
1. A program designed to provide a systematic method to monitor and

evaluate factors which may affect the quality of laboratory test results
is_________________. A system designed to monitor the accuracy
and precision of the testing process is called___________________.
2. The three types of variables that may affect the testing process are:
3. The reproducibility of a value upon repeated measurement is referred

to as ______________ and the term used to define the quality or state
of being correct in clinical chemistry (e.g., when the calculated mean
and true value are very close) is _________________.
4. Current regulations require that a minimum of two levels of control be

run with patient samples at least every____________ of testing.
5. Given the 20 days of data for this one level of glucose (see data
below), calculate the mean, the standard deviation, and the ± 1, ± 2,
and ± 3 SD ranges.

Self-Evaluation
Glucose
Day
(mg/dL)
1 112
2 107
3 106
4 109
5 130
6 112
7 113
8 114
9 116
10 106
11 108
12 109
13 111
14 112
15 115
16 110
17 111
18 108
19 107
20 105
a. The mean or X = _________________
b. The standard deviation =__________________.
c. The ± 1 SD range is _________________ to ________________.

The ± 2 SD range is _________________ to ________________.
The ± 3 SD range is _________________ to ________________.

Self-Evaluation
d. Using the data on the previous page, label the Levey-Jennings chart
below and plot the first 8 days results.
+3SD
+2SD
+1SD
Mean
–1SD
–2SD
–3SD
Days 1 2 3 4 5 6 7 8

Self-Evaluation Answer Key

To complete this self-evaluation:
• Fill in the blanks to correctly complete the following statements.
• Calculate the mean, standard deviation, and the ±1, ±2, and ±3 SD
ranges for 20 days of glucose data.
• Plot data on a Levey-Jennings chart and interpret the chart results.
1. A program designed to provide a systematic method to monitor and

evaluate factors which may affect the quality of laboratory test results
is quality assurance. A system designed to monitor the accuracy and
precision of the testing process is called quality control.
2. The three types of variables that may affect the testing process are:
preanalytical variables
analytical variables
postanalytical variables
3. The reproducibility of a value upon repeated measurement is referred

to as precision and the term used to define the quality or state of being
correct in clinical chemistry (e.g., when the calculated mean and true
value are very close) is accuracy.
4. Current regulations require that a minimum of two levels of control be

run with patient samples at least every 24 hours of testing.
5. Given the 20 days of data for this one level of glucose (see data
following), calculate the mean, the standard deviation, and the ±1, ±2,
and ±3 SD ranges.

Glucose
Day
(mg/dL)
1 112
2 107
3 106
4 109
5 130
6 112
7 113
8 114
9 116
10 106
11 108
12 109
13 111
14 112
15 115
16 110
17 111
18 108
19 107
20 105
a. The mean or X = 111.1 mg/dL.
b. The standard deviation = 5.4.
c. The ±1 SD range is 105.7 to 116.5.

The ±2 SD range is 100.3 to 121.9.
The ±3 SD range is 94.9 to 127.3.

d. Using the data on the previous page, label the Levey-Jennings chart
below and plot the first 8 days results.
127.4 +3SD
121.9 +2SD
116.5 +1SD
◆
◆
111.1 Mean ◆ ◆
◆
◆
105.7 –1SD ◆
100.3 –2SD
94.9 –3SD
Days 1 2 3 4 5 6 7 8


Ortho-Clinical Diagnostics, Inc.
100 Indigo Creek Drive
Rochester, New York 14626-5101

7B6006 StatTools TrngMod

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

7B6006 StatTools TrngMod

Transféré par

Droits d'auteur :

Formats disponibles

Part No.

Statistical Tools of Quality Control:

VITROS is a trademark of Ortho-Clinical Diagnostics, Inc.

 Ortho-Clinical Diagnostics, Inc., 1995 - 2006. All rights reserved.

Revision Date Description

Revision Date Section Page

iv Statistical Tools of Quality Control: Training Module Part No. 7B6006

How This Manual Is Organized

Additional Training Modules

Additional training modules are available on orthoclinical.com. For

Part No. 7B6006 Statistical Tools of Quality Control: Training Module v

What is a Quality-Control Rule?. . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Interpreting Quality-Control Results . . . . . . . . . . . . . . . . . . . . . . . 27

Figure 1. Accuracy and Precision . . . . . . . . . . . . . . . . . . . . . . 5

Figure 2. Variation in Height Measurement . . . . . . . . . . . . . . . 6

Figure 4. Glucose Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Figure 5. Standard Deviation . . . . . . . . . . . . . . . . . . . . . . . . . 10

Figure 7. Calculating Standard Deviation . . . . . . . . . . . . . . . 12

Figure 8. Frequency Distribution . . . . . . . . . . . . . . . . . . . . . 14

Figure 10. Types of Quality-Control Materials . . . . . . . . . . . 17

Figure 11. Three Quality-Control Rule . . . . . . . . . . . . . . . . . . 21

Figure 12. Rule 22s Violation 1 . . . . . . . . . . . . . . . . . . . . . . . . 23

Figure 13. Rule 22s Violation 2 . . . . . . . . . . . . . . . . . . . . . . . . 24

Figure 14. Rule 22s Violation 3 . . . . . . . . . . . . . . . . . . . . . . . . 24

Figure 15. Control Rule Violation 22s . . . . . . . . . . . . . . . . . . . 25

Figure 16. Control Rule Violation 13s . . . . . . . . . . . . . . . . . . . 26

Figure 17. Levey-Jennings Chart . . . . . . . . . . . . . . . . . . . . . 27

Figure 18. Standard Deviation Range . . . . . . . . . . . . . . . . . . 29

Part No. 7B6006 Statistical Tools of Quality Control: Training Module ix

Part No. 7B6006 Statistical Tools of Quality Control: Training Module 1

Control of Preanalytical Variables

2 Statistical Tools of Quality Control: Training Module Part No. 7B6006

• Periodically verifying that procedure manuals are complete and up to

Control of Analytical Variables

Control of Postanalytical Variables

Part No. 7B6006 Statistical Tools of Quality Control: Training Module 3

Accuracy and Precision

4 Statistical Tools of Quality Control: Training Module Part No. 7B6006

Calculating the average of the quality-control measurements is the first step

Accuracy and Precision

Accurate Inaccurate Inaccurate

Figure 1. Accuracy and Precision

Part No. 7B6006 Statistical Tools of Quality Control: Training Module 5

Statistical Measure of Accuracy

Variation in Height Measurement

Figure 2. Variation in Height Measurement

6 Statistical Tools of Quality Control: Training Module Part No. 7B6006

Height measurement # of Occurrences (frequency)

The measurements shown in Figure 2 are represented on the graph in

Figure 3. Frequency Distribution of Height Measurements

The individual plotted points in Figure 3 represent the height measurement

Part No. 7B6006 Statistical Tools of Quality Control: Training Module 7

Central Tendency and Mean

Calculation of the Mean

Figure 4. Glucose Results

8 Statistical Tools of Quality Control: Training Module Part No. 7B6006

2. Divide this sum by the total number of measurements (n), in this

Practice Exercises 1 and 2 will allow you optional practice calculating

Statistical Measures of Precision

Part No. 7B6006 Statistical Tools of Quality Control: Training Module 9

Figure 5. Standard Deviation

Variance is a measure of variability. It is used to calculate standard

10 Statistical Tools of Quality Control: Training Module Part No. 7B6006