Laboratory Management and Quality Control

General Introduction

Health care delivery is no longer a simple process of examining the patient and giving him a
prescription. Over the years there has been rapid expansion in the various branches of health care
services. As part of this expansion process and explosion of scientific medical knowledge, laboratory
diagnosis has gained tremendous importance in today's practice. Through the use of quality control
(QC) the laboratory can ensure that the results being issued by it are reliable enough to allow decisions
to be taken with confidence.

Quality control in the medical laboratory is a statistical process used to monitor and evaluate the
analytical process that produces patient results. It is the study of those errors which are the
responsibility of the laboratory, and of the procedures used to recognize and minimize them.
When a diagnostic test is performed in the medical laboratory, the outcome of the test is a result. The
result may be a patient result or it may be a quality control (QC) result. The result may be
quantitative (a number) or qualitative (positive or negative) or semi-quantitative (limited to a few
different values).
Why do we need quality control in the laboratory? Lack of quality control can lead to incorrect
test results. Incorrect laboratory results may lead to wrong management decisions with possible fatal
results. The reliability of laboratory results is therefore most important. It is not sufficient to ‘think’
that ‘my’ results are satisfactory. This has to be proved with scientific evidence. Laboratory personnel
must know that QC is an obligation to the patient, that it is designed to give the analyst confidence in
the methods used and that its purpose is not to find scapegoats or to punish those making mistakes.
The impact of an incorrect test result can be very diverse.
 At the individual level an incorrect test result can lead to a wrong diagnosis, wrong treatment
or wrong advice.
 At the population level, this could lead to a wrong evaluation of the nutritional status and to
wrong policy regarding the nutritional problem.
 At the research level, it could lead to disagreement between scientists.
Thus lack of quality control can lead to a loss of credibility of the laboratory and besides that, it is cost
ineffective. The pillars of analytical quality control are based on controlling and monitoring precision
and accuracy of an analytical procedure. Precision (repeatability and reproducibility) can be monitored
and controlled for by analyzing in-house control samples (repeatability) or by participating in
proficiency studies (reproducibility).

Quality Manual
Towards achieving quality, international accreditation programmes strongly recommend the
production of a quality manual by the laboratory.

The quality manual of a laboratory is a document or a set of documents describing the organizational
structure, responsibilities, procedures and processes by which the laboratory achieves its objectives
and gains confidence in its work. The manual is indispensable for achieving and maintaining good
overall quality. Furthermore, the preparation of a quality manual may induce the laboratory to improve
quality. Even a non-mandatory quality manual may be a valuable document for a clinical laboratory in
demonstrating to clinicians and the hospital administration a commitment to quality.

"The laboratory shall define and document its policies and objectives for, and its commitment to, good
laboratory practice. The hospital management shall ensure that these policies and objectives are
documented in the quality manual and communicated to, understood by, and implemented by all
laboratory personnel concerned. The quality manual contents are as follows:

Contents of quality manual

These are:
 Quality Policy and Quality System
 Organization
 Quality Control
 Personnel
 Accommodation and Environment
 Equipment
 Reference Materials
 Test Procedures
 Handling of Reagents
 Sample Collection, Storage and Disposal
 Maintenance of Records
 Laboratory Reports and Despatch of Reports

Quality policy
The aim of the laboratory is to provide clinically useful information through laboratory measurement
of samples from patients, taking into account the allocated resources.
The quality policy is implemented by the following means;
 Proper sample collection, stabilization, transport, sample preparation and identification.
 Reliable analytical work so that systematic and random errors do not exceed specified limits.
 Turn-around time within specified limits for routine and emergency measurements, and for
rare routine measurements.
 Data reported in a clear form and supplemented with relevant information, including reference
intervals to allow reliable clinical interpretation.
 Appropriate communication to the clinicians so that the results will be interpreted correctly
and logically integrated into further (clinical and laboratory) evaluation of the patients, and that
the clinicians become aware of unexpected problems and errors.

Standard operating procedures

The preparation of test procedures comes under the broad heading of Standard Operating Procedures
(SOPs). SOP is a clear, concise and comprehensive written instruction of a method or procedure which
has been agreed upon and authorized as the operating policy of the department.
In general, SOPs, which mainly contain detailed descriptions of each analytical method, are essential
for maintaining the same analytical quality over a long period of time. The procedures are a
prerequisite to correct transfer of methods from one laboratory to another. The contents of SOP are as
follows:
 Introduction
 Principle of method
 Specimen types, collection and storage
 Reagents, standards and control - preparation and storage
 Equipment, glassware and other accessories
 Detailed procedure
 Calculations, calibration curve
 Analytical reliabilities – (QC and Statistical assessment)
 Hazardous reagents
 Reference range and clinical interpretation
 Limitations of method (e.g. interfering substances and troubleshooting)
 References
 Date and signature of authorization

Laboratory errors

Analytical errors are classified into random errors and systematic errors. It is clear that random errors
indicate poor precision while systematic errors indicate poor accuracy. A few examples of random
errors are pipetting error, transcription error, wrong sample numbering and labelling, and fluctuating
readings on the colorimeter. Systematic errors could occur due to wrong procedure, incorrect standard
and calibration procedure.

Errors can occur in any of the limb of the cycle of events taking place in a hospital, starting from the
physician examining the patient and back to the physician (pre-analytical/ analytical/post-analytical).

The physician, after examining the patient, decides and orders a test, and collects and transports the
patient’s samples; this constitutes the pre-analytical limb of the cycle of events. In the analytical limb
the sample is received by the laboratory and analysed. The post-analytical limb consists of the transfer
of the result to the physician and a meaningful interpretation of the laboratory data by the physician,
followed by necessary action.

Definition:

Accuracy: It is the degree of agreement between a measured value and its ‘true/consensus’ value. On
the contrary, inaccuracy, which is represented by analytical bias, is defined as the % of the difference
between the measured value and the ‘true’ value over the true value. Therefore, good accuracy means
least analytical error.

Precision refers to reproducibility. It refers to the agreement between replicate measurements. It is

quantitatively expressed as the standard deviation (SD) or more precisely as percent coefficient of
variation (CV), which is defined as SD times 100 divided by the mean value of the results in a set of
replicate measurements. Therefore, good precision means least CV.

Pre-analytical

The pre-analytical system shall take care of the following aspects, as each can have a major effect on
the accuracy of the result:

 Patient preparation
 Request forms
 Specimen collection, containers, labelling and phlebotomy equipment and procedure
 Specimen transport
 Specimen preparation
 Specimen storage

Analytical

The following aspects shall be monitored, evaluated, implemented and maintained to ensure the
accuracy and precision of the test carried out:

 Quality of distilled water

 Calibration of measuring and testing instruments including balances, thermometers, incubators,
waterbaths, autoclaves, centrifuges and semi-automatic pipettes, and regular servicing and
maintenance of equipment.

It is essential to use a standard calibrator which is traceable to national/international reference

material. The laboratory shall obtain evidence of traceability to the reference material from the
supplier. Precision can be maintained through the use of suitable QC material, either commercial or
prepared in-house. The QC material should be analysed at predetermined intervals along with patient
samples to monitor systematic and random errors. Such QC material shall also be traceable to a
national/international certified reference material so that the accuracy of measurements can be
monitored.

All data relating to the laboratory’s internal QC practices and performance in external quality
assessment schemes (scoring, ranks, etc.) shall be recorded, reviewed and corrective actions
implemented.

Stability of reagents

Laboratory personnel should be aware that the stability of all reagents kept at room temperature will
go down from the stated values if the temperature exceeds 350C.

Use of calibration graphs

A fresh standard curve should be carried out for the analyses described in this manual whenever:

 the calibrator is changed

 new reagents are introduced
  problems with QC are encountered

Post-analytical

In order to avoid transcriptional errors in the results of the test, the reporting/signatory technicians
shall verify the results entered manually or through on-line instrument interfaces before the results are
reported or despatched.

Rectification of laboratory errors

It is therefore essential to continually ask the following questions.

 Is there an analytical error?

 If so, what type of error is this?
 What could have been the causes for this error?

How to rectify this error?

It is important to identify analytical errors and classify them as either random or systematic errors.
Towards this end, the laboratory should implement internal QC procedures. This involves preparation
of a QC pool, either human or bovine, quantification of unavoidable laboratory errors, construction of
Levey-Jennings chart and daily analysis of QC along with every batch of patients’ samples.

Prevention of systematic errors

To prevent or minimize systematic errors, the laboratory should adhere to the following
points:

 Use of proper calibration technique. Use of pure chemicals, precision balance, quality
distilled water. Proper preservative and storage
 Regular checking of photometric filter, bulb, tubing, etc.
 Use of recommended analytical methods
 Calibration of pipettes at regular intervals
 Instrument calibration – photometric check
 Use of calibrated cuvette
 Regular preventive maintenance of equipment – daily, weekly and monthly.

While it is easy to identify systematic errors, it is quite difficult to pinpoint random errors. In
order to minimize this possibility, it is important to educate the staff on various aspects that
could lead to random errors.

Preparation of QC pool

Ethanediol stabilized liquid serum QC pool has been established in the authors’ laboratory (4) based on
the WHO method (5). This procedure is applicable to both pooled human serum as well as bovine
serum. This preparation is economical and appropriate for use in the laboratories in developing
countries.
 Use of patients’ s era

A serum pool can be prepared by salvaging the extra serum from leftover patients’ samples after
analysis. Samples that are significantly haemolysed or lipemic or icteric should be excluded.
Similarly, samples that show positive tests for HIV antibodies and Hbs antigen should also be
excluded. In view of the dangers in handling infectious blood samples, use of animal-based QC pool is
recommended.

Use of bovine serum

Collect about 2-3 litres of fresh bovine blood in a 5-litre clean plastic bucket. Allow to clot at room
temperature for about 30 minutes.

Slice the clot into small pieces using a sharp knife and leave the bucket at 2-8 0C for 12 hours to enable
the serum to ooze out. Decant the crude serum into a one litre beaker or flask.

Transfer this crude serum into several glass centrifuge tubes (size 15 x 120 mm) and then centrifuge
for 10 minutes at 3500 rpm and decant the clear serum into a clean bottle.

Transfer one litre serum into a one-litre plastic bottle.

Mix the contents well and store the container at –200 C for 12 hours or until frozen.

While monitoring day-to-day laboratory performance with internal QC, it is preferable to use different
levels of QC materials to cover the entire pathological ranges. Therefore, methods of preparation of
three levels of QC (low, normal and high) are described below.

If the preparation of all three levels of QC pool is not possible, it is essential to make use of at least
one level, viz. normal level.

The procedures described below for the preparation of all three QC levels are applicable to both
human serum and bovine serum.

Preparation of normal-level QC serum

Remove the container from the freezer and fix it upside down over a one-litre plastic measuring
cylinder. Collect the first 830 mg, which will be rich in all constituents. Add 150 ml of ethanediol to
this and mix well. Take an aliquot and measure the levels of various analytes. Use 20 ml distilled
water to dissolve the various substances that will be added to the serum in order to increase the levels
of these to the desired normal levels. Total volume = 830+150+20 = 1000 ml

The QC serum thus prepared will contain 15% (V/V) ethanediol.

Preparation of high-level QC serum

Freeze one litre of clear serum at -200C. Remove the container from the freezer and fix it upside down
over a one-litre plastic measuring cylinder. Collect the first 700ml, which will be more concentrated
and rich in all constituents. Add 127.5 ml of ethanediol to this and mix well. Take an aliquot and
measure the levels of various analytes. Use 22.5 ml distilled water to dissolve the various substances
that will be added to the serum in order to increase their levels to the desired levels. Total volume =
700 +127.5+22.5 = 850 ml.

The QC serum thus prepared will contain 15% (V/V) ethanediol.

Commercially produced liquid or lyophilized QC sera are expensive and may not always be available,
thus, an alternative is to produce ethylene glycol stabilized QC serum in the laboratory, using
commercially QC serum to calibrate and standardize it.

Procedure:

Collect venous blood aseptically into a container from several volunteers that will produce about
100mL of serum who have been tested negative for HIV and hepatitis.

Allow the blood to clot at room temperature. Refrigerate overnight at 2-80C, protecting it from light.

Collect the separated serum and centrifuge for 15 minutes at 2000rpm.

Transfer the clear serum into a screw-cap plastic container and freeze at -20C

Thaw the pooled serum at room temperature the following day, protecting it from light.

Using a 10mL syringe, remove and discard the top 15mL of fluid from the pooled serum (this contains
mainly water and weak concentration of analytes).

Replace the extracted fluid with 15mL of ethylene glycol and mix well for about 30 minutes using a
magnetic votex mixer.

Aliquot the pooled, stabilized serum into 1-2mL eppendorf tubes, date and store at -20C, while some
is left for measuring the parameters.

Measure each analyte in duplicates in the serum spectrophotometrically, including the commercial QC
serum as a control. List the analytes measured and their values, date and keep for routine use.

Construction of Levey Jennings Chart

On each day when analyses are performed a fresh sample is thawed, thoroughly mixed and analysed.
(Remember: Ethanediol-treated serum may not freeze completely at –20 0C; however, the constituents
are quite stable). The QC serum is analysed for a period of 20 days or so. [ Important to note:
Analysis should not be carried out by only one person; all staff should participate in this exercise to
determine the true unavoidable error in the laboratory]. From these data, mean and SD are calculated.
Levey Jennings chart is then constructed with x + 2SD as warning limits and x + 3 SD as control
limits.

Calculate the %CV for each analyte to ascertain whether this is within the acceptable limit (Ideal = <
5%. Must be definitely < 8%). If % CV is found to be high, this will indicate that between-day
laboratory precision (variation) is high and the data cannot be used to construct a Levey Jennings
chart. It is then essential to identify the causes for this, correct these and then repeat the whole exercise
and confirm that the %CV is well within the acceptable limit.

Table 1 shows precision data obtained by a WHO trainee from a developing country for routine
biochemical analytes through analysis of ethanediol-treated QC serum for 20 days, while undergoing
training at the author’s laboratory recently.

Table 1

Autoanlyser Hitachi 912 Manual Method

Analyte
Mean %CV Mean %CV
Glucose 167 1.2 166 4.2
Total protein 9.8 1.2 9.6 3.3
Albumin 3.5 1.4 3.6 4.2
Urea 50 1.7 55 3.8
Calcium 10.3 1.9 10.5 4.0
Cholesterol 185 1.4 180 4.5
Creatinine 1.7 0.7 1.8 3.0

Interpretation of QC data

A. According to WHO(6)an analytical system is ‘out of control’ if one of the four criteria is met.
That is:

a value lies entirely outside the control limits

seven consecutive values show a rising tendency

seven consecutive values show a falling tendency

seven consecutive values lie on the same side of the mean

If one of these situations arises, the patients’ results must be discarded, the cause of the error sought
and removed, and then the batch repeated with a QC serum.
B. Use of two different levels of QC simultaneously in every batch of analysis provided valuable
information on the type of errors – whether these are random (precision) errors or systematic
(accuracy) errors.

Westgard’s rules for interpreting QC data obtained using a single-level QC as well as two different-
level QCs are schematically presented in Fig.1.(7)

QC results are used to validate whether the instrument is operating within pre-defined specifications,
inferring that patient test results are reliable. Once the test system is validated, patient results can then
be used for diagnosis, prognosis, or treatment planning. For example, when a patient’s serum is
assayed (tested) for potassium, the test result tells us how much potassium (concentration) is present in
the blood. This result is then used by the physician to determine whether the patient has a low, normal
or high potassium.

How Often Should Controls be Run?

Ideally, controls should be assayed with each analytical run and placed randomly through the run to
detect analytical imprecision. Controls should also have assay values within clinically significant
ranges.
Factors to Help Determine QC Frequency
••Instrument, reagent, and method reliability
••The clinical application of the test result (i.e.will incorrect patient results pose risk?)
••Amount of time available to retrieve and correct a result in error (i.e.will result be acted upon
immediately after it's reported)
••Training and competency of test operators
Use of multiple levels of control allows for better laboratory decisions regarding analytical error and
validity of the run.

Basic Quality Control Statistics

The expected range of values for a control is calculated by use of relatively simple statistics.
These statistics include mean, standard deviation, coefficient of variation, and the standard
deviation index.

Mean [x]

The mean is defined as the arithmetic average of a set of data points.It is expressed in Formula 1.
The mean describes the “central tendency” of the data set.In the clinical lab, the mean identifies the
“target value” of a set of data points, usually QC or patient data.
The mean is the fundamental statistic used for comparison or for calculation of other statistics. The
Clinical and Laboratory Standards Institute's Guidance for statistical quality control recommends that
at least 20 data points collected from 20 or more “separate” runs be used to establish laboratory target
values for control materials. Laboratories should establish their own target values using manufacturer
assay values only as guides. Provisional target values may be established by running 20 replicates in
less than 20 runs, but the provisional values must be replaced after data from 20 separate runs is
accumulated.3 However, for purposes of this discussion, only five data points will be used in the
following example.Using LDH values, find the sum of the data {120, 115, 110, 119, 123}.The sum [Σ]
is 587 U/L.The number of values is 5 (n = 5).Therefore, the mean for the LDH values is 117.4 U/L (or
587 U/L divided by 5).

Standard Deviation [s]

The standard deviation (s) quantifies the degree of dispersion of data points about the mean and is
used to set limits upon which control result acceptability is determined.Quality control data often
exhibit a “normal” or Gaussian distribution around the mean.
In a Gaussian distribution:
••68.3% of values are within ± 1.0 standard deviation of the mean
••95.5% of values are within ± 2.0 standard deviations of the mean
••99.7% of values are within ± 3.0 standard deviations of the mean

Formula 2: Calculating a Standard Deviation [s] For a Set of QC Values

Where:
s = standard deviation
x = mean (average) of the QC values
Σ(xn - x)2 = the sum of the squares of differences between individual QC values and the mean
n = the number of values in the data set

To calculate the standard deviation for the set of values in the previous LDH example, begin by
calculating the mean [ x ]:
x = (120 + 115 + 110 + 119 + 123 U/L) ÷ 5 x = 587 U/L ÷ 5 x = 117.4 U/L

Calculate the standard deviation [SD or S] of the LDH

s = 5.03 U/L (Rounded

Limits for data acceptability are defined using the standard deviation statistic.The range for the 1s
limit would be calculated as Mean ± 1s.
Consequently, the 1s range (limit) for our LDH example is calculated as:
117.4 U/L – 5.03 U/L = 112.4 U/L 117.4 U/L + 5.03 U/L = 122.4 U/L The 1s range is 112.4 to 122.4
U/L
Approximately 68% of future data should be between 112.4 and 122.4 U/L.Approximately 32%
should be less than 112.4 U/L or greater than 124.4 U/L.
The 2s range (limit) is calculated as:
Mean ± 2s
117.4 U/L – (2 x 5.03 U/L) = 107.3 U/L
117.4 U/L + (2 x 5.03 U/L) = 127.5 U/L
The 2s range is 107.3 to 127.5 U/L.
Only about 4.5% of future data should be less than 107.3 U/L or greater than 127.5 U/L (i.e., only one
result in 20 should be beyond these limits).

The 3s range (limit) is calculated as:

Mean ± 3s
117.4 U/L – ( 3 x 5.03 U/L) = 102.3 U/L
117.4 U/L + ( 3 x 5.03 U/L) = 132.5 U/L
The 3s range is 102.3 U/L to 132.5 U/L.
Only about 0.3% of future data should be less than 102.3 U/L or greater than 132.5 U/L.It would be
very unusual to obtain a result beyond these limits.

In the medical laboratory, these ranges (limits) are used to determine the acceptability of a test
run not only on the basis of a single data point but on groups of data points as well.
Standard deviation is also valuable when comparing methods or evaluating new instruments. A
method or instrument with a low standard deviation produces consistent results. The lab using an
instrument or method which has high standard deviations will have less certainty about the accuracy of
diagnosis or the effectiveness of treatment because of test result variability. In other words, high
standard deviations (poor precision, greater variability) can affect the integrity of all results. The
method or instrument selected should provide a standard deviation which is medically acceptable.

Coefficient of Variation [CV]

The coefficient of variation (CV) is a measure of variability. The CV for our LDH example using the
CV formula would be (5.03 U/L ÷ 117.4 U/L) (100) = 4.3%.
The CV is useful for comparisons of precision at different concentrations as long as the materials used
are similar and CVs are determined under similar conditions. This statistic is commonly used to
compare manufacturer claims. It can also be used as a part of the internal quality control system when
performing patient precision testing.

Standard Deviation Index [SDI]

Another statistic which is helpful to evaluate performance is the standard deviation index (SDI).This
statistic, which is usually obtained by participation in an interlaboratory or proficiency testing
program, is used to compare a laboratory’s results to its peer group.
Formula 4: Calculating the Standard Deviation Index [SDI]
SDI= Mean (Lab) – Mean (Group)/ s (group)
Mean (Group) = peer group mean
S (Group) = standard deviation
Mean (Lab) = laboratory mean

The target SDI is 0.0.This would indicate that the lab’s performance is identical to the peer group
average. Acceptable SDI values are between ±1.0. Any test/method/instrument which has an SDI
between ±1.0 and 1.5 may have a problem and the lab should investigate. The lab must troubleshoot
and correct any test/method/instrument which has an SDI of ±2.0 or greater. The relative importance
of the SDI statistic does depend, however, on the size of the peer group.

Although SDI is a statistic usually generated by participation in an interlaboratory or proficiency

testing program, it can be used as a tool in monitoring internal quality control performance, as well.
For example, if the laboratory suspects that a trend is occurring, this suspicion can be validated by use
of the SDI statistic. In this case, rather than using peer group data in calculating the statistic, the
laboratory’s cumulative mean and cumulative standard deviation are used.The SDI formula is
modified as follows: (Mean of suspect data points – lab cumulative mean) ÷ lab cumulative standard
deviation. An SDI value greater than ±1.0 indicates a possible problem with the test.

QUALITY CONTROL POOL PREPARATION, CHARACTERIZATION, AND USE

1. Reasons for using quality control (QC) pools: QC gives confidence and validity to results,
helps monitor the performance and output (especially drift) of a process, indicates when
problems/deficiencies exist, helps make decision about results (i.e., acceptable criteria), and provides a
statistical basis from which the results can be judged.
2. Accuracy and precision: Accuracy is the extent to which the average of several measurements
agrees with the true value; precision is the ability of an instrument or method to reproduce its own
measurement.
3. Statistical parameters: The mean of a QC material assists in the indication of method
performance. The mean provides an estimate of the central tendency of the distribution that is
expected when method performance remains stable. Any change in accuracy, such as a systematic
shift or drift, would result in a change in the mean value of the control, which would be shown by a
shift in the distribution of control results.

Selection of materials for QC pools

1. When selecting and using QC materials, note the following:

a. Composition should be the same as patient samples. In case of analysing serum or
plasma, this implies using serum or plasma samples for quality control.
b. The laboratory procedure manual or manufacturer’s manual should specify appropriate
QC material for analysis.
c. QC materials should be handled as patient samples.
d. Mean and standard deviation of control material must be established before use.
e. Control materials with different concentrations must have different lot numbers.
2. Sources of QC materials:
a. Residual specimens
b. Expired blood-bank materials
c. Collections specifically for use
d. Laboratory preparations
e. Commercial sources

Preparation of serum bench QC pools

1. Preparing sufficient amounts of bench QC pools at one time for 2–3 years is recommended; this
will provide long-term quality assurance for your laboratory and make the operation as efficient as
possible.
2. Use one of the previously mentioned sources to obtain serum materials.
3. Prescreen the collected serum materials to determine the analyte concentrations of each
individual sample and how to pool the serum materials to obtain one low-, one medium-, and one
high-concentration pool.
4. Pool the materials designated to make up each pool and mix thoroughly.
5. Filter each pool through several layers of sterile 4x4 gauze into a sterile, glass Wheaton bottle and
thoroughly mix on a stir plate before and during dispensing. Clean stir bar with alcohol then water
before adding to the pool.
6. For each pool, generate aliquots of a prespecified amount of serum into a prelabeled 2-mL cryovial
and close the vial.
7. Place capped vials in 9x9 or 10x10 boxes labeled with pool name, assay, and box number (e.g.,
Box 1 of 7, 2 of 7, etc.).
8. Freeze boxes at -40C or colder. Typically, they can be kept at -70C for several years (depending
on the analyte).
Homogeneity testing of freshly prepared bench QC pools

1. Randomly select ~5% of the total number of vials prepared for each pool. Selection is best
done by taking a specific number of vials from different locations of each rack of pools.
2. Analyze those samples in one run to determine whether pool is homogenous. Not more than
10% variability should occur from vial to vial.

Characterization of bench QC pools

1. Analyze the newly prepared bench QC pools throughout a 20-day period to establish QC limits
for each pool.
2. Calculate the mean and standard deviation (SD) for the 20 days, and calculate the mean ± 2 SD
and mean ± 3 SD to set the limits.
3. Do not analyze patient samples for which results will be reported during this characterization
phase.
4. A minimum 10 days worth of data should be collected to establish preliminary limits before
any data are reported for patient samples.
5. As the end of the bench QC pools (~40 vials left) approaches, prepare a new lot of bench QC
pools and characterize it while still using the old lot. This will provide overlapping data from one lot
of QC pools to the next.
Setting up in-house quality control
In each run, the precision of the analytical process is determined by the analysis of replicates
(duplicates) of control material. Here we assume that "day" and "run" are similar. Control limits for
each level of the control pool are calculated from the average of the daily (run) means, standard
deviations of the daily (run) means, and the average of the daily ranges of the control material.
Preliminary control limits are calculated from a minimum of 20 acceptable runs. Final limits are
determined after 50 runs
Evaluation of quality control limits
The results of the control pool should be reviewed to determine that the data have been collected
during a stable analytical period. Care should be taken to look for outliers, for periods of questionable
or unstable performance, and for evidence of excessive bias. An outlier will distort the limits if
incorporated into the calculations. An outlier is any value of x which falls outside the control limits
(x± 3sx) or any value of which exceeds the control limit for. Outliers and values from any
questionable period of performance should be rejected.

Monitoring intra-laboratory performance

Internal quality control using statistical control charts, is a simple and powerful tool to monitor the
stability of an analytical procedure. The concept is simple: a homogeneous sample (control sample)
with a matrix similar to the "real" (subject) sample is analysed in the same run as the subject samples.
Most often, the control sample is analysed in duplicate (sometimes in triplicate or quadruplicate).
How to use bench and blind QC pools

1. Bench QC pools are known to the analyst and inserted at the start of each run (after analyzing
the calibration curve and before analyzing any patient samples) and at the end of each run so that
judgments can be made on the day of analysis.
2. Blind QC pools are prepared, tested for homogeneity, and characterized in the same way as
bench QC pools. Typically, only two levels of blind QC pools are used: a level representing normal
concentrations and a level representing abnormal concentrations (either elevated or depressed). Blind
QC pools are labeled so that they are indistinguishable from the patient samples. Because blind QC
pools are not known to the analyst, they are processed exactly like patient samples. Only the
supervisor reviews the results of the blind QC pools and has the key to decode them. This procedure
for testing blind QC pools is difficult when only one person works in the laboratory. One alternative
solution would be to have the supervisor pull five random samples to be re-run by the analyst.
3. Together, the bench and blind QC system helps assess the range of analyte concentrations by
taking these samples through the complete analytical process. The data from these materials can then
be used to estimate methodological imprecision and assess the magnitude of any time-associated
trends.

Using Westgard Rules

The Westgard rules are simple to use routinely. They improve capacity for detecting real analytical
errors, give a low level of false assay rejections, indicate the type of error, and help with problem
solving. There are six basic rules in the Westgard scheme. Some are designed to detect random error;
others detect systematic error which may indicate a bias in the system.
Westgard rules are used individually or in combination to evaluate the quality of analytical runs. Rule
combinations are selected by the laboratory and should be based upon the quality required and the
laboratory performance for each analytical method. The overall objective is to obtain a high
probability of error detection and a low frequency of false rejection of runs.
Laboratories which use the 12s rule alone in performing their quality control will frequently reject
runs which are valid. Following are recommendations for acceptance/rejection criteria when three QC
pools (low, medium, high) are used per run:
1. If all three QC pools (mean of replicate measurements) are within 2s
limits, accept the run.
2. If one of the three QC results is outside 2s limits, apply rules below and reject if any condition
is met:
12S: This rule is violated when a single control observation is outside the ±2s limits.Remember that in
the absence of added analytical error, about 4.5% of all quality control results will fall between the 2s
and 3s limits.This rule merely warns that random error or systematic error may be present in the test
system.The relationship between this value and other control results within the current and previous
analytical runs must be examined.If no relationship can be found and no source of error can be
identified, it must be assumed that a single control value outside the ±2s limits is an acceptable
random error.Patient results can be reported.
Laboratories should never use this rule as a run rejection rule without reason.This rule may be used as
a run rejection rule when the lab wishes to tightly control poor performing tests.

If a Westgard rule is violated, the technologist must review the performance of the test, consult
troubleshooting guides, perhaps perform maintenance, correct any identified problems or
departure from protocol, and notify the supervisor who will make decisions about reporting
results and rerunning the test.
13S: One QC pool (mean of replicate measurements) is outside 3s limit;
This rule identifies random error or possibly the beginning of a large systematic error.Any QC result
outside ±3s violates this rule.

22S: Two of the three QC pools (mean of replicate measurements) in current run are outside 2s limit
on the same side of the mean;
This rule identifies systematic error only.The criteria for violation of this rule are:
••Two consecutive QC results
••Greater than 2s
••On the same side of the mean
There are two applications to this rule: within-run and across runs.The within-run application affects
all control results obtained for the current analytical run.For example, if a normal (Level I) and
abnormal (Level II) control are assayed in this run and both levels of control are greater than 2s on the
same side of the mean, this run violates the within-run application for systematic error.If however,
Level I is -1s and Level II is +2.5s (a violation of the 12s rule), the Level II result from the previous
run must be examined.If Level II in the previous run was at +2.0s or greater, then the across-run
application for systematic error is violated.
Violation of the within-run application indicates that systematic error is present and that it affects
potentially the entire analytical curve.Violation of the across-run application indicates that only a
single portion of the analytical curve is affected by the error.

R4S: Two of the three QC pools (mean of replicate measurements) in current run are outside 2s limit
on opposite sides of the mean.This rule identifies random error only, and is applied only within the
current run.If there is at least a 4s difference between control values within a single run, the rule is
violated for random error.For example, assume both Level I and Level II have been assayed within the
current run.Level I is +2.8s above the mean and Level II is -1.3s below the mean.The total difference
between the two control levels is greater than 4s (i.e.[+2.8s – (-1.3s)] = 4.1s).

41S: This rule detects systematic error and is applied both within and across control materials.This
rule is violated within the control material when the last four control values of the same control level
exceed the “same” (mean +1s) or (mean –1s) limit.The rule is violated across control materials when
the last four consecutive control values for different control levels exceed the “same” (mean +1s) limit
or (mean –1s) limit.

10x sequential - Previous 9 QC results (mean of replicate measurements) for this pool were on the
same side of the mean. This rule detects systematic error and is applied both within and across control
materials.This rule is violated within the control material when the last ten values for the same control
level are all on the same side of the mean.The rule is violated across control materials when the last
ten consecutive values, regardless of control level, are on the same side of the mean.

Patient Precision Testing

Although the purpose of controls is to validate analytical runs, they also identify potential problems
with the analytical system which includes the instrument or kit, technologist, ancillary equipment and
reagents.Sometimes, people working in the laboratory have difficulty deciding whether the control is
at fault when an out of control situation occurs, especially if it is a repeated occurrence.Patient
precision testing is a useful mechanism to distinguish between analytical system performance and
control performance.It also increases sensitivity to random errors.
Patient precision testing is relatively easy to implement.The laboratory chooses an abnormal or normal
patient sample for repeat testing on the next analytical run or next day.It is a form of duplicate testing.

Limits for the maximum allowable difference between duplicate results may be derived by either of
two methods.One method requires the laboratory to perform a series of replicate tests on normal
and/or abnormal patient samples, defining the absolute difference between each replicate, and finally
establishing an acceptable range of performance based on these replicate differences.9,10
The other system requires the use of normal and/ or abnormal patient samples and the CV for the
corresponding control level.In this system, the CV is applied to the initial patient result to define an
acceptable replicate range.10 When the sample is retested, the replicate value is compared to the
predefined range.Calculated limits for allowable differences are applicable only over a narrow range
over which the standard deviation (or CV) is fairly constant.Thus, specific concentration ranges must
be defined for selecting patient specimens for use as controls.

Patient precision test results can be interpreted as follows:

1. If the control and the patient precision replicate are within acceptable limits, then there probably is
no problem with the control or the system.
2. If the control and the patient precision replicate are both outside acceptable limits, then there
probably is a problem with the system.
3. If the control is within acceptable limits, but the patient precision replicate is outside acceptable
limits, then there is possible random error occurring or a possible problem with the integrity of the
precision sample.
4. If the control is outside acceptable limits, but the patient precision replicate is within acceptable
limits, then there is possible random or systematic error occurring or a possible problem with the
integrity of the control.

5.2.7 Corrective actions

Analysts should be alert for indicators of pending loss of control. If an out-of-control condition occurs,
the analyst should:
1. Stop analysis of specimens
2. Alert the laboratory supervisor
3. Check readings and calculations for errors
4. Review control charts and records, including records of instrument maintenance or of reagent and
standard solution preparation
5. If an obvious cause is found, apply the appropriate remedial action
6. Recalibrate and analyse control samples to determine if performance is now acceptable
7. When corrective actions have been successful, resume analysis; repeat the analysis of any samples
analysed during the out-of-control period
8. Document the problem and the remedial actions taken; record all actions on control charts and
records.