ESTABLISHMENT OF

CONTROL LIMITS

NURDIANA ZAINUDDIN
MED. LAB. TECHNOLOGY

1
The Quality System
Organization Personnel Equipment

Process
Purchasing & Control Information
(QC & EQA) &
Inventory Specimen
Management
Management

Documents Occurrence
Assessment
& Records Management

Process Customer Facilities &

Improvement Service Safety

2
The Quality Assurance Cycle

Patient/Client Prep
Sample Collection
Personnel Competency
Reporting Test Evaluations
•Data and Lab
Management
•Safety
•Customer
Service Sample Receipt
and Accessioning

Record Keeping

Quality Control Sample Transport

Testing

3
 Quality Control –
Quantitative Tests
• How to implement a laboratory
quality control program

4
 Implementing a QC Program –
Quantitative Tests
• Select high quality controls
• Collect at least 20 control values over a period of
20-30 days for each level of control
• Perform statistical analysis
• Develop Levey-Jennings chart
• Monitor control values using the Levey-Jennings
chart and/or Westgard rules
• Take immediate corrective action, if needed
Record actions taken

5
Quality Control - Quantitative

• Analysis of QC Data

6
 How to carry out this analysis?
• Need tools for data management and analysis
 Basic statistics skills
 Manual methods
 Graph paper
 Calculator
 Computer helpful
 Spreadsheet
• Important skills for laboratory personnel

7
 Analysis of Control Materials
• Need data set of at least 20 points, obtained
over a 30 day period
• Calculate mean, standard deviation,
coefficient of variation; determine target
ranges
• Develop Levey-Jennings charts, plot results

8
 Establishing Control Ranges
• Select appropriate controls
• Assay them repeatedly over time
 at least 20 data points
• Make sure any procedural variation is represented:
 different operators
 different times of day
• Determine the degree of variability in the data to establish
acceptable range

9
 Measurement of Variability
• A certain amount of variability will naturally occur
when a control is tested repeatedly.
• Variability is affected by operator technique,
environmental conditions, and the performance
characteristics of the assay method.
• The goal is to differentiate between variability due
to chance from that due to error.

10
 Measures of Central Tendency
• Data are frequently distributed about a
central value or a central location
• There are several terms to describe that
central location, or the ‘central tendency’
of a set of data

11
 Measures of Central Tendency
• Mean = the calculated average of the
values
• Mode = the value which occurs with the
greatest frequency
• Median = the value at the center (midpoint)
of the observations

12
Exercise 1: Population of months since
last examination
Patient Number of Months Since Last
Examination
1 12

2 13

3 14

4 15

5 16
13
Mean = the calculated average of the values

14
 Calculation of Mean

X = Mean
X1 = First result
X2 = Second result
Xn = Last result in series
n – Total number of results
15
• Mean number of months = (12+13+14+15+16)
5
= 70
5

= 14
16
MODE – the most frequent value in a set of
data
Exercise 2:
1, 5, 2, 3, 4, 5, 5, 7
=5

17
MEDIAN – middle of a set of value when
arranged in ascending order

Exercise 3:
1, 5, 2, 3, 4, 5, 5, 7, 8
arrange: 1,2,3,4,5,5,5,7,8

=5
18
Normal Distribution

19
 Normal Distribution
• All values are symmetrically distributed
around the mean – Gaussian Distribution
• Characteristic “bell-shaped” curve
• Assumed for all quality control statistics

20
 Normal Distribution

X
Frequency

-3 -2 -1 Mean 1 2 3

21
 Normal Distribution

16
14
Mean
12
# of Observations

10
8
6
4
2
0
192 194 196 198 200 202 204 206 208 210 212
Serum glucose (mg/dL)

22
 Accuracy and Precision
• Precision – refers to repeatability or reproducibility;
the ability to get the same result in subsequent tests
on the same sample.
• Accuracy – the closeness of a measured value to a
true value. The difference of the mean value to the
true value - inaccuracy

 Quality Control is used to monitor both the precision and

the accuracy of the assay in order to provide reliable
results.
23
 Precision and Accuracy
• Precise and inaccurate • Precise and accurate

24
Imprecise and inaccurate

25
 Measures of Dispersion
or Variability
• There are several terms that describe the
dispersion or variability of the data around the
mean:
Range
Variance
Standard Deviation
Coefficient of Variation

26
 Range
• Range refers to the difference or spread
between the highest and lowest observations.
• It is the simplest measure of dispersion.
• It makes no assumption about the shape of
the distribution or the central tendency of the
data.

27
 Calculation of Variance
• Variance is a measure of variability about the
mean.

Calculation of Variance (S2)

2

2 (X  X )
S  N 1 2
 mg /dl
1 2

28
Calculation of Standard Deviation

(x  x )
S  mg/dl
2
1

N 1

variance
29
 Calculation of Standard Deviation
• The standard deviation (SD) is the square root of the
variance
 it is the square root of the average squared deviation
from the mean
• SD is commonly used (rather than the variance) since it has
the same units as the mean and the original observations
• SD is the principle calculation used in the laboratory to
measure dispersion of a group of values around a mean

30
 Standard Deviation and Probability
• For a set of data with a
normal distribution, a value X
will fall within a range of:

Frequency
 +/- 1 SD 68.2% of the
time 68.2%
 +/- 2 SD 95.5% of the
time 95.5%
 +/- 3 SD 99.7% of the 99.7%
time -3s- 2s -1s
+3s
Mean +1s +2s

31
 Standard Deviation and Probability
• In general, laboratories use the +/- 2 SD criteria for
the limits of the acceptable range for a test
• When the QC measurement falls within that range,
there is 95.5% confidence that the measurement is
correct
• Only 4.5% of the time will a value fall outside of
that range due to chance; more likely it will be due
to error

32
 Calculation of
Coefficient of Variation
• The coefficient of
variation (CV) is the
standard deviation SD
(SD) expressed as a
CV  x 100
percentage of the
mean
mean
• Ideally should be less
than 5%

33
Monitoring QC Data

34
 Monitoring QC Data
• Use Levey-Jennings chart
• Plot control values each run, make decision
regarding acceptability of run
• Monitor over time to evaluate the precision and
accuracy of repeated measurements
• Review charts at defined intervals, take necessary
action, and document

35
 Levey-Jennings Chart

• A graphical method for displaying control

results and evaluating whether a procedure
is in-control or out-of-control
• Control values are plotted versus time
• Lines are drawn from point to point to accent
any trends, shifts, or random excursions

36
 Levey-Jennings Chart
1 .2
-2 0 -1 5 -1 0 -5 0

+3SD

+2SD
0 .8

+1SD

Mean

0 .4

-1SD

-2SD

-3SD
0

37
 Levey-Jennings Chart -
Record Time on X-Axis and the Control Values on Y-Axis
115
110
Control Values (e.g. mg/dL)

105
100
95
90
85
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time (e.g. day, date, run number)

38
 Levey-Jennings Chart -
Plot Control Values for Each Run
115
110
Control Values (e.g. mg/dL)

105
100
95
90
85
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time (e.g. day, date, run number)

39
 Levey-Jennings Chart
Calculate the Mean and Standard Deviation;
Record the Mean and +/- 1,2 and 3 SD Control Limits
115
+3SD
110
+2SD
105
+1SD
100
Mean
95
-1SD
90
-2SD
-3SD
85
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day

40
 Levey-Jennings Chart -
Record and Evaluate the Control Values
115
+3SD
110
+2SD
105
+1SD
Mean 100
95
-1SD
90
-2SD
-3SD 85

80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day

41
 Findings Over Time
• Ideally should have control values clustered about the mean (+/-
2 SD) with little variation in the upward or downward direction
• Imprecision = large amount of scatter about the mean. Usually
caused by errors in technique
• Inaccuracy = may see as a trend or a shift, usually caused by
change in the testing process
• Random error = no pattern. Usually poor technique,
malfunctioning equipment

42
WESTGARD RULES

43
 When does the Control Value
Indicate a Problem?
• Consider using Westgard Control Rules
• Uses premise that 95.5% of control values
should fall within ±2SD
• Commonly applied when two levels of control
are used
• Use in a sequential fashion

44
 Westgard Rules
• “Multirule Quality Control”
• Uses a combination of decision criteria or
control rules
• Allows determination of whether an analytical
run is “in-control” or “out-of-control”

45
 Westgard Rules
(Generally used where 2 levels of control
material are analyzed per run)
• 12S rule • R4S rule
• 13S rule • 41S rule
• 22S rule • 10X rule

46
 Westgard – 12S Rule
• “warning rule”
• One of two control results falls outside ±2SD
• Alerts tech to possible problems
• Not cause for rejecting a run
• Must then evaluate the 13S rule

47
 12S Rule = A warning to trigger careful inspection of
the control data
+3SD

+2SD

+1SD
12S rule
Mean violation
-1SD
-2SD
-3SD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day

48
 Westgard – 13S Rule
• If either of the two control results falls outside
of ±3SD, rule is violated
• Run must be rejected
• If 13S not violated, check 22S

49
13S Rule = Reject the run when a single control measurement
exceeds the +3SD or -3SD control limit
+3SD

+2SD

+1SD

Mean
13S rule
-1SD violation
-2SD
-3SD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day

50
51
 Westgard – 22S Rule
• 2 consecutive control values for the same
level fall outside of ±2SD in the same
direction, or
• Both controls in the same run exceed ±2SD
• Patient results cannot be reported
• Requires corrective action

52
 22S Rule = Reject the run when 2 consecutive control
measurements exceed the same
+2SD or -2SD control limit
+3SD

+2SD

+1SD

Mean 22S rule

-1SD violation
-2SD
-3SD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day
53
 Westgard – R4S Rule
• One control exceeds the mean by –2SD, and
the other control exceeds the mean by +2SD
• The range between the two results will
therefore exceed 4 SD
• Random error has occurred, test run must be
rejected

54
 R4S Rule = Reject the run when 1 control
measurement exceed the +2SD and the other exceeds
the -2SD control limit
+3SD

+2SD

+1SD

Mean
R4S rule
-1SD violation
-2SD
-3SD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day
55
 Westgard – 41S Rule
• Requires control data from previous runs
• Four consecutive QC results for one level of
control are outside ±1SD, or
• Both levels of control have consecutive results
that are outside ±1SD

56
 41s

• 41s - Reject when 4 consecutive control measurements

exceed the same mean plus 1s or the same mean
minus 1s control limit.
57
 Westgard – 10X Rule
• Requires control data from previous runs
• Ten consecutive QC results for one level of
control are on one side of the mean, or
• Both levels of control have five consecutive
results that are on the same side of the mean

58
 10x Rule = Reject the run when 10 consecutive control
measurements fall on one side of the mean
+3SD

+2SD

+1SD

Mean

-1SD
-2SD
10x rule
-3SD
violation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Day

59
Westgard Multirule QC

60
 When a rule is violated
• Warning rule = use other rules to inspect the
control points
• Rejection rule = “out of control”
 Stop testing
 Identify and correct problem
 Repeat testing on patient samples and controls
 Do not report patient results until problem is
solved and controls indicate proper performance

61
 Solving “out-of-control” problems
• Policies and procedures for remedial action
• Troubleshooting
• Alternatives to run rejection

62
 Types of Error
• Random error: the error that occurs
without any real pattern - 1-3s, R 4s
• Systematic error: the error that is
continuous and affects all results
equally - 2-2s, 4-1s, or 10x

63
 Summary
• Why QC program?
 Validates test accuracy and reliability

64
 Summary:
How to implement a QC program?
 Establish written policies and procedures
 Assign responsibility for monitoring and reviewing
 Train staff
 Obtain control materials
 Collect data
 Set target values (mean, SD)
 Establish Levey-Jennings charts
 Routinely plot control data
 Establish and implement troubleshooting and corrective
action protocols
 Establish and maintain system for documentation
65
 Problem-solving

• When a run is out-of-control, investigate

the process & correct the problem:
determine the type of error occurring on
the basis of the rule violated.
• Random error: the error that occurs
without any real pattern - 1-3s, R 4s
• Systematic error: the error that is
continuous and affects all results equally -
2-2s, 4-1s, or 10x
66
• Refer to trouble shooting guide to identify possible
causes for type of error indicated by the control rule
violated

• Inspect the testing problem. Occurring with only one

test or with many tests on chemistry analyzer? Related
causes to recent changes? Systematic errors are most
often due to reagent or calibration problems. So inspect
reagent, calibration, & maintenance records for clues.

• If shift occurred immediately following reagent

replacement, verify that the lot number is correct & has
been checked out or calibrated, that it has been
prepared properly, that it is the correct reagent.

67
• A systematic trend – check for slowly deteriorating reagent, a
deteriorating filter or lamp. Use a systematic logical
troubleshooting approach in isolating the causes, making only
one change at a time & documenting each action taken.

• Correct the problem, and then analyze control samples again

to assess control status.

• Repeat or verify the results on the patient samples once the

method has been demonstrated to be in-control.

• Consult a supervisor for any decision to report patient results

when a run is out-of-control.

68
69