QUALITY CONTROL IN CLINICAL

BIOCHEMISTRY

by

Dr. Biswanath Chakraborty

Dept. of Biochemistry and
Medical Biotechnology
STM  
 WHAT IS QUALITY CONTROL?
 

Quality control is a statistical process used to

monitor and evaluate the analytical process
that produces patient results.
What is the aim of Quality Control?

 “The aim of quality control is simply to ensure

that the results generated by the test are
correct.

 However, quality assurance is concerned with

much more: that the right test is carried out on
the right specimen, and that the right result
and right interpretation is delivered to the right
person at the right time”
 Quality Control (QC)
QC is part of the “Big Picture” of Quality
Assurance

 Quality Assurance (QA)

– Focuses on all aspects of laboratory performance
• Analytical results
• Overall laboratory Services
– Specimen labeling, specimen handling,
accurate reporting of test results, TAT,
incident reports, patient, nurse or
clinician surveys
 What is Quality Assessment?
 Quality assessment (also known as proficiency
testing) is a means to determine the quality of the
results generated by the laboratory. Quality
assessment is a challenge to the effectiveness of the
QA and QC programs.

 Quality Assessment may be external or internal,

examples of external programs include NEQAS,
HKMTA, and Q-probes.
 CONTROL OF THE ANALYTICAL QUALITY USING
STABLE CONTROL MATERIALS
• The performance of analytical methods can be monitored by analyzing
specimens whose concentrations are known and then by comparing the
observed values with known values.

• The known values are usually represented by an interval of acceptable

values, or upper and lower limits for control (control limits)

• When the observed values fall within the control limits – analysis is
working properly

• When the observed value fall outside the control limits the analyst should
be alerted to the possibility of problems in the analysis.
 What is Standard Deviation?

Standard deviation is a statistical measure of spread or

variability. The standard deviation is the root mean square
(RMS) deviation of the values from their arithmetic mean. These
parameters give an easy way to summarize data as the sample
becomes large: 68% of the values are within one
standard deviation of the mean. 95% of the values are within
two standard deviations of the mean. More than 99% of the
values are within three standard deviations of the mean value.
 SD

• Standard deviation - extent of random

variation
• SD = d2
n-1

d=difference of individual result from mean

n=number of observations
• SD =√ Σ (X – M)2 / (n – 1)

•     where Σ = Sum of
              X = Individual score
              M = Mean of all scores
              n = Sample size (Number of scores)

      Variance :

            Variance = (SD)2
 Variance Definition

The square of the standard deviation. A

measure of the degree of spread among a set
of values; a measure of the tendency of
individual values to vary from the mean value.
 CV

Co-efficient of variation
 relative
magnitude of variability while
comparing two procedures

CV % = (SD x 100)/mean
 What is Accuracy ?

• Closeness of agreement between true value and the

mean of observed value obtained over large number
of observations. This can be quantitatively expressed
as Bias
Observed value – True value
Bias = ----------------------------------- X 100
True value
Good accuracy means minimum Bias.
 What is Precision ?

 Precision is closeness of results with each other of a large

number of successive observations in a measurement process,
under prescribed conditions.
 This can be quantitatively expressed as percentage of coefficient
of variation ( % CV). Good precision means minimum % CV. It
can be calculated using the formula given below:
SD
% CV = -------------- X 100
Mean
Good Accuracy Good Precision Neither Good precision
Good Precision Only Nor Accuracy
ACCURACY AND PRECISION
• Repeatability :
Closeness of the agreement between the
successive measurements of the same
specimen.

• Reproducibility :
Closeness of the agreement between the
results of successive measurements of the
same specimen.
Ability of a test to
give correct
diagnosis of a
disease/condition.
 Diagnostic sensitivity
The proportion of subjects with disease who
have positive test results. The greater the sensitivity
of a test, the fewer the number of false-negative
results.

True positives
Sensitivity = -----------------------------------
True positives + False Negatives
 Diagnostic specificity
The proportion of subjects without disease who have negative
test results. The ability of a test to give fewer false positive
results.
True Negatives
Specificity = --------------------------------------------
False Positives + True Negatives

• Linearity - It determines the upper and lower limit of

reporting range of a test.
 Sensitivity and specificity are
inversely related.

Marker concentration

- Disease +
 Evaluating Clinical Performance of
laboratory tests
 The sensitivity of a test indicates the likelihood that it
will be positive when disease is present
 The specificity of a test indicates the likelihood that it
will be negative when disease is absent
 The predictive value of a test indicates the probability
that the test result correctly classifies a patient
 Do you prefer a sensitive test or a specific
test?
• Choice of test characteristic is a function of
diagnostic need
– Discovery tests are often applied to asymptomatic patients.
• requires good sensitivity

– Confirmatory tests require high specificity

• establish presence of disease

– Test to exclude a disease require high sensitivity

• little chance of false negatives
 Defining the value of a Diagnostic Test

 The clinical performance of a laboratory test defines how

well it predicts disease.
 The sensitivity of a test indicates the likelihood that it will
be positive when disease is present
 The specificity of a test indicates the likelihood that it will
be negative when disease is absent
 Distinction between a test being a true positive (TP) or
true negative (TN) is usually determined with reference to
a selected “normal range” based on 95% confidence. This
range is referred to as a Reference Range (RR). Test
focuses on the upper or lower limit often use a Reference
Value (RV).
 False Positive (FP) Normal people falling outside RR
 False Negative (FN) Diseased individual falling inside RR
Defining the value of a Diagnostic Test
 What is normal???
• Normal Health
– choice between defining an ‘ideal state’ or determining
an ‘average’ state in people considered healthy
– more practical to analyze ‘average’ state --with respect
to a biological parameter results in a symmetric
(Gaussian) distribution
 Do you prefer a sensitive test or a
specific test?
• Choice of test characteristic is a function of
diagnostic need

– Discovery tests are often applied to asymptomatic patients.

 requires good sensitivity

– Confirmatory tests require high specificity

 establish presence of disease

– Test to exclude a disease require high sensitivity

 little chance of false negatives
 QC in the Laboratory involves
Evaluating method performance

Constant
Long term
Short term

So what are we evaluating?

 Evaluating method performance
• Accuracy
• Precision
• Sensitivity
• Linearity
 The Expectation
• Clinicians depend on the the laboratory to provide
results that are
– Accurate
– Precise
– Timely

• Result compared to reference interval and patient

is treated accordingly
– Result also compared to previous patient results

How does the laboratory meet this expectation?

Evaluating method performance
Accuracy
The closeness of the agreement between the
measured value of an analyte and True Value
 Gold standard
 Comparison to Reference Laboratory
 Average of a number of labs
Evaluating method performance
 Precision
The ability of an analytical method to
produce the same value for replicate
measurements of the same sample
With-in run
Between run
 Method Sensitivity
• The analytical sensitivity of a method refers
to the lowest concentration of analyte that
can be reliably detected.
• The most common definition of sensitivity
is the analyte concentration that will result
in a signal two or three standard deviations
above background.
 Clinical Sensitivity Vs Analytical Sensitivity

Signal/Noise threshold
Signal

time
 Application of QC to Laboratory Methods

We base the validity of our patient test result by

comparison to other samples analyzed by the same
method
• Sample Types Analyzed by the Lab
– Patient Samples
– QC samples
– Proficiency samples (PT)
 The Application of QC to Laboratory Methods
• QC Samples
• Samples selected by the laboratory and purchased from an external
company
• Tested each day of patient testing to mimic patient and proficiency
samples
• Indicates if the lab’s analytical processes are performing in an acceptable
manner, thereby producing clinically acceptable patient and proficiency
results
• QC samples provide data about accuracy and precision of each methods
at the level of analyte present in the control
• We interpret the data when making daily decisions about the
acceptability of each batch of patient or PT samples
 The Application of QC to Laboratory Methods

• Running QC and PT samples reflects how well the lab

is doing but does not modify or correct or control
analytical problems
• Require QC system to “drive” the system:
– QC data
– Rules of acceptability
– Tools for monitoring and documenting daily QC and
corrective actions
 The Application of QC to Laboratory Methods

Rules of Acceptability are based upon:

• Performance specifications
– Numerical limits established by each laboratory for each
analyte and each testing system, which often include
accuracy, precision, analytical sensitivity (minimum
reportable amount), analytical specificity (interfering
substances), the reportable range of patient test results and
the reference range.
• Action Limits
– Ranges set for QC samples that if exceeded signal a possible
deterioration of the quality of the testing system and
requires an investigation by a technologists.
 The Application of QC to Laboratory Methods
• QC Tools
– Levy-Jennings Control Chart
• A simple graphical display in which the observed values are plotted
versus an acceptable range of values as indicated on the chart by lines
for upper and lower control limits
– Control Rules
• A decision criterion used to interpret QC data and make judgment on
the control status
– Statistical QC program
• Software part of specific analyzer and/or laboratory information system
(LIS) which analyzes QC data and provides appropriate feedback
– warnings
– flags
 Levey-Jennings Control Chart

+3sd

+2sd

+1sd

mean

-1sd

-2sd

-3sd
 Levey-Jennings Control Chart

+3sd

+2sd

+1sd
mean

-1sd

-2sd

-3sd
 Levey-Jennings Control Chart

+3sd

+2sd

+1sd

mean

-1sd

-2sd

-3sd
Levey-Jennings Control Chart
 The Application of QC to Laboratory Methods

• Control Rules
– Westgard Multirule
• A control procedure that uses a series of control
rules to test the control measurement.
– A 12s rule being used as a warning, followed by the use of
13s, 22s, R4s, 41s and 10x as rejection rules
 Westgard Rules
13s refers to a control rule that is
commonly used with a Levey-Jennings
chart when the control limits are set as
measurement exceeds the mean plus 3s or
the mean minus 3s control the mean plus
3s and the mean minus 3s. A run is
rejected when a single control limit.

12s refers to the control rule that is

commonly used with a Levey-
Jennings chart when the control
limits are set as the mean plus/minus
2s. In the original Westgard multirule
QC procedure, this rule is used as a
warning rule to trigger careful
inspection of the analyte
 Westgard Rules
22s - reject when 2 consecutive
control measurements exceed the

 
same mean plus 2s or the same
mean minus 2s control limit.
                    

R4s - reject when 1 control

measurement in a group exceeds the
                      

mean plus 2s and another exceeds the

mean minus 2s.
 Westgard Rules

41s - reject when 4

consecutive control
measurements exceed the
same mean plus 1s or the
same mean minus 1s
                       

control limit. 

10x - reject when 10

consecutive control
 
 
                   
measurements fall on one
side of the mean.
QC algorithm utilizing Westgard Rules
 LABORATORY TESTING PROCESSES AND THEIR POTENTIAL ERRORS
  PRE ANALYTICAL ERRORS
 
PROCESS POTENTIAL ERRORS
 Test orderinginappropriate test
Handwriting not legible
Wrong patients ID
Special requirements not specified
 Specimen acquisition
Incorrect tube or container
Incorrect patient ID
Inadequate volume
Invalid specimen
(hemolysed or diluted) Collected at wrong time
Improper transport conditions
ANALYTICAL ERRORS

 Analytical Instrument not calibrated Measurement correctly

Specimens mix – up
Incorrect volume of specimen
Interfering substances present
Instrument precision problem

Test reporting Wrong patient ID

Report not legible
Report delayed
Transcription error
POST ANALYTICAL ERRORS
 
Test interpretation Interfering substance not recognized

Specificity of the test not understood

Precision limitation not recognized
Analytical sensitivity not appropriate
Previous values not available for comparison
HOW TO CONTROL THESE ERRORS?
 PRE ANALYTICAL VARIABLES
  It is very difficult to establish effective methods for monitoring and controlling
preanalytical variables because may of the variables are outside the laboratory
areas.
  Requires the coordinated effort of many individuals and hospital departments
 Patient Identification
  The highest frequency of errors occurs with the use of handwritten labels and
request forms. The use of bar code technology has significantly reduced ID
problems.
Turnaround time
Delayed and lost test requisitions, specimens and reports can be major problems
for labs. Recording of the actual times of specimen collection, receipt in the lab and
reporting of results with use of computers will solve these problems.
Transcription error
  A substantial risk of transcription error exists from manual entry of data even with the
double checking of results, computerization will reduce this type of transcription error.
 Patient preparation
 Lab tests are affected by many factors, such as,
recent intake of food, alcohol, or drugs
smoking
exercise
stress
sleep
posture during specimen collection
The lab must define the instructions and procedures compliance with these
instructions can be monitored directly efforts should be made to correct non
compliance
Specimen Collection
  Prolonged tourniquet application causes local anoxia to cells and excessive venous backpressure,
venous stasis and hemoconcentration.
  Blood collection from an arm into which an intravenous infusion is running can be diluted or
contaminated.
 Hemolysis during blood collection
Improper containers with incorrect preservatives

  To monitor and control these problems, specially trained lab team assigned to specimen collection
The identification of the person collection a specimen should be maintained
 Clinicians should be encouraged to report clinically inconsistent results.
Pride of workmanship should be encouraged and quality performance should be rewarded.
 
Specimen transport
The stability of specimens during transport from the patient to the lab is seldom monitored;
 CONTROL OF ANALYTICAL VARIABLES

There are many analytical variables that must be carefully

controlled –

Water quality
Calibration of analytical balances
Calibration of volumetric glassware and pipets
Stability of electrical power
Stability of temperature of heating baths, refrigerators, freezers
and centrifuges
The procedure Manual should contain the following

Procedure name
Clinical significance
Principle of method
Specimen of choice
Reagents and equipments
Procedure
Reference values
Comments
References