Quality assurance in medical laboratory

Eisa SalehiTehran University of medical SciencesTehran University of medical Sciences

www.salehilab.org

Pretest?

• Quality assuranceQuality assurance• Q.C.G• GLP

• STD• C.V.• ErrorsErrors

Welcome

Admission

Hem

Sample collectionCSR

matology

Paperwork offices

Molecular Diag

نكنيم چكار آزمايشگاه يم:در ر ي چ ز :ر

حتي اگر مايع مورد نظر آب . ھرگز با دھان پي پت نكنيد•.باشد

.در آزمايشگاه از خوردن و آشاميدن اجتناب كنيد•خودداري كنيد   ̋از مخلوط كردن تصادفي مواد شيميائي جدا•

در كه ترتيبي به را شيميايي مواد محلولھا، ساختن بھنگام ر و يبي ر يي ر ب ي ي و ھ و ن م و بھنقش دانشمند ديوانه .( دستور كار شما آمده است اضافه كنيد

نكنيد بازي آزمايشگاه در .)را ي زي ي ب ز ر .)ر

بكنيم چكار آزمايشگاه :در

مطالب و بخوانيد را استفاده مورد شيميائي مواد بروشورھاي

يم ر ب ي چ ز :ر

ب ي و و ور ر ب ي ي ي و ي ور برو.آنھا را بكار گيريد

كنيد استفاده دستكش و مناسب پوشش از .ھمواره ي ش ب و ش ز پو ر . وتجھيزات ايمني تعبيه شده در آزمايشگاه را بشناسيد و طرز

بگيريد ياد را آن .كار يري ن ر ي ب .ر مواد شيميايي را به روش صحيح و توصيه شده اختصاصي

بريزيد دور .آنھا ور بريزي .ھ در طول حضور در آزمايشگاه اطلاعات بدست آمده و

كنيد يادداشت را شده انجام .كارھاي ي م ر ي ج ي .ر

Material Safety Data Sheets (MSDS)Material Safety Data Sheets (MSDS)• MSDS Databases http://www.msds.com/

P d t Id tifi ti• Product Identification• Composition/Information on Ingredients• Hazards Identification• First Aid Measures• Fire Fighting Measures• Accidental Release Measures• Handling and Storage• Exposure Controls/Personal Protection• Exposure Controls/Personal Protection • Physical/Chemical Properties• Stability and Reactivity Data• Toxicological Information• Ecological Information• MSDS Transport Information• Regulatory Information

Errors:Requesting appropriate tests

i i i iWriting prescription

Reading prescription

Sample collection

Sampling times

environmental factors

D i t fDrug interferences

Patient identification

Sample transfer

Technician errors

Instrumental errors

Method limitationsMethod limitations

Data entry mistakes

Interpretation of the results

How can we trust ??

Quality Assurance vs.Quality Assurance vs.Quality Assurance vs. Quality Assurance vs. Quality ControlQuality ControlQ yQ y

Quality AssuranceQuality Assurance Quality ControlQuality ControlQ yQ yAn overall management plan to guarantee the

Q yQ yA series of analytical measurements used

h li hp gintegrity of data (The “system”)

to assess the quality of the analytical data(The “tools”)( y ) (The tools )

Definitions (1)

• Quality Control - QC refers to the measures that must be includedduring each assay run to verify that the test is working properly.

• Quality Assurance - QA is defined as the overall program thatensures that the final results reported by the laboratory are correct.

• “The aim of quality control is simply to ensure that the resultsThe aim of quality control is simply to ensure that the resultsgenerated by the test are correct. However, quality assurance isconcerned with much more: that the right test is carried out on the rightspecimen, and that the right result and right interpretation is deliveredp , g g pto the right person at the right time”

Definitions (2)

• Quality Assessment - quality assessment (also known asproficiency testing) is a means to determine the quality ofthe results generated by the laboratory. Quality assessmentis a challenge to the effectiveness of the QA and QCprograms.programs.

• Quality Assessment may be external or internal,

Good Laboratory PracticesGood Laboratory PracticesGood Laboratory PracticesGood Laboratory Practices

GLP: GOOD LABORATORY GLP: GOOD LABORATORY PRACTICEPRACTICE

GLP GLP is an FDA regulation.is an FDA regulation.DefinitionDefinition: GLP embodies : GLP embodies a set of principles thata set of principles thata set of principles that a set of principles that provides a framework provides a framework within which laboratory within which laboratory studies arestudies are plannedplannedstudies are studies are planned planned performed, monitored, performed, monitored, reported and archivedreported and archived..GLP i ti f dGLP i ti f dGLP is sometimes confused GLP is sometimes confused with the standards of with the standards of laboratory safety like laboratory safety like

i f t li f t lwearing safety goggles.wearing safety goggles.

HISTORYHISTORYGLP is a formal regulation that was created by the GLP is a formal regulation that was created by the FDA (United states food and drug administration) in FDA (United states food and drug administration) in 1978197819781978..Although GLP originated in the United States , it had Although GLP originated in the United States , it had a world wide impact.a world wide impact.NonNon--US companies that wanted to do business with US companies that wanted to do business with the United states or register their pharmacies in the the United states or register their pharmacies in the United States had to comply with the United StatesUnited States had to comply with the United StatesUnited States had to comply with the United States United States had to comply with the United States GLP regulations.GLP regulations.They eventually started making GLP regulations in They eventually started making GLP regulations in y y g gy y g gtheir home countries.their home countries.In In 1981 1981 an organization named an organization named OECD (organization OECD (organization for economic cofor economic co operation and development )operation and development )for economic cofor economic co--operation and development ) operation and development ) produced GLP principles that are international produced GLP principles that are international standard.standard.

WHY WAS GLP CREATED?WHY WAS GLP CREATED?In the early In the early 7070’s FDA became ’s FDA became aware of cases of poor laboratory aware of cases of poor laboratory practice all over the United States.practice all over the United States.practice all over the United States.practice all over the United States.FDA decided to do an inFDA decided to do an in--depth depth investigation on investigation on 40 40 toxicology toxicology labs.labs.They discovered a lot fraudulent They discovered a lot fraudulent activities and a lot of poor lab activities and a lot of poor lab practices.practices.E l f f thE l f f thExamples of some of these poor Examples of some of these poor lab practices found were lab practices found were

1. Equipment not been calibrated to standard form thereforeto standard form , therefore giving wrong measurements.

2. Incorrect/inaccurate accounts of the actual lab studyy

3. Inadequate test systems

FAMOUS EXAMPLEFAMOUS EXAMPLEOne of the labs that went One of the labs that went under such an investigation under such an investigation made headline news.made headline news.The name of the Lab was The name of the Lab was Industrial Bio Test. This was a Industrial Bio Test. This was a big lab that ran tests for big big lab that ran tests for big companies such as Procter andcompanies such as Procter andcompanies such as Procter and companies such as Procter and Gamble.Gamble.It was discovered that mice It was discovered that mice that they had used to testthat they had used to testthat they had used to test that they had used to test cosmetics such as lotion and cosmetics such as lotion and deodorants had developed deodorants had developed cancer and died.cancer and died.

d l l b hd l l b hIndustrial Bio Test lab threw Industrial Bio Test lab threw the dead mice and covered the dead mice and covered results deeming the products results deeming the products good for human consumption.good for human consumption.good for human consumption.good for human consumption.Those involved in production, Those involved in production, distribution and sales for the distribution and sales for the lab eventually served jail time.lab eventually served jail time.

OBJECTIVES OF GLPOBJECTIVES OF GLPOBJECTIVES OF GLPOBJECTIVES OF GLP

GLP makes sure that the data submitted are a GLP makes sure that the data submitted are a true reflection of the results that are obtained true reflection of the results that are obtained during the study.during the study.GLP also makes sure that data is traceableGLP also makes sure that data is traceableGLP also makes sure that data is traceable.GLP also makes sure that data is traceable.Promotes international acceptance of tests.Promotes international acceptance of tests.

MISSION OF GLPMISSION OF GLPMISSION OF GLPMISSION OF GLP

Test systemsTest systemsArchiving of records and materials.Archiving of records and materials.ggApparatus, material and reagent facilities.Apparatus, material and reagent facilities.Quality assurance programs.Quality assurance programs.Qua ty assu a ce p og a sQua ty assu a ce p og a sPerformance of the study.Performance of the study.Reporting of study results.Reporting of study results.Reporting of study results.Reporting of study results.Standard operating procedures (SOP)Standard operating procedures (SOP)Personnel and test facility organizationPersonnel and test facility organizationPersonnel and test facility organizationPersonnel and test facility organization

Standard Operating Procedures Standard Operating Procedures (SOP)(SOP)(SOP)(SOP)

Written procedures for a laboratories program.Written procedures for a laboratories program.They define how to carry out protocolThey define how to carry out protocol--specified specified activities.activities.Most often written in a chronological listing of Most often written in a chronological listing of action steps.action steps.action steps.action steps.They are written to explain how the procedures They are written to explain how the procedures are suppose to workare suppose to workare suppose to workare suppose to work

SOPSOPSOPSOP

Routine inspection, cleaning, maintenance, Routine inspection, cleaning, maintenance, testing and calibration.testing and calibration.Actions to be taken in response to equipment Actions to be taken in response to equipment failure.failure.Analytical methodsAnalytical methodsDefinition of raw dataDefinition of raw dataKeeping records, reporting, storage, mixing, and Keeping records, reporting, storage, mixing, and retrieval of dataretrieval of data

Instrumentation ValidationInstrumentation ValidationInstrumentation ValidationInstrumentation Validation

This is a process necessary for any analyticalThis is a process necessary for any analyticalThis is a process necessary for any analytical This is a process necessary for any analytical laboratory.laboratory.Data produced by “faulty” instruments mayData produced by “faulty” instruments mayData produced by faulty instruments may Data produced by faulty instruments may give the appearance of valid data.give the appearance of valid data.The frequency for calibration reThe frequency for calibration re validationvalidationThe frequency for calibration, reThe frequency for calibration, re--validation validation and testing depends on the instrument and and testing depends on the instrument and extent of its use in the laboratoryextent of its use in the laboratoryextent of its use in the laboratory.extent of its use in the laboratory.Whenever an instrument’s performance is Whenever an instrument’s performance is

id h “ l li i ” bid h “ l li i ” boutside the “control limits” reports must be outside the “control limits” reports must be discontinued discontinued

I t t V lid ti ( t)I t t V lid ti ( t)Instrument Validation (cont)Instrument Validation (cont)

Equipment records should include:Equipment records should include:Name of the equipment and manufacturerName of the equipment and manufacturerModel or type for identificationModel or type for identificationSerial numberSerial numberDate equipment was received in the Date equipment was received in the laboratorylaboratoryC f f t ti i t tiC f f t ti i t tiCopy of manufacturers operating instruction Copy of manufacturers operating instruction (s)(s)

Reagent/ Mate ials Ce tificationReagent/ Mate ials Ce tificationReagent/ Materials CertificationReagent/ Materials Certification

This policy is to assure that reagents used This policy is to assure that reagents used are specified in the standard operating are specified in the standard operating p p gp p gprocedure.procedure.

Purchasing and testing should be handled Purchasing and testing should be handled by a quality assurance programby a quality assurance programby a quality assurance program.by a quality assurance program.

Reagents and Solutions contReagents and Solutions contgg

Requirements:Requirements:Reagents and solutions shall be labeledReagents and solutions shall be labeledReagents and solutions shall be labeled Reagents and solutions shall be labeled Deteriorated or outdated reagents and Deteriorated or outdated reagents and solutions shall not be usedsolutions shall not be usedInclude Date openedInclude Date openedStored under ambient temperatureStored under ambient temperatureStored under ambient temperatureStored under ambient temperatureExpiration dateExpiration date

Anal st Ce tificationAnal st Ce tificationAnalyst CertificationAnalyst Certification

Some acceptable proof of satisfactory Some acceptable proof of satisfactory training and/or competence with specific training and/or competence with specific g / p pg / p plaboratory procedures must be established laboratory procedures must be established for each analyst.for each analyst.yyQualification can come from education, Qualification can come from education, experience or additional trainings but itexperience or additional trainings but itexperience or additional trainings, but it experience or additional trainings, but it should be documentedshould be documentedSufficient peopleSufficient peopleSufficient peopleSufficient peopleRequirements of certification varyRequirements of certification vary

Labo ato Ce tificationLabo ato Ce tificationLaboratory CertificationLaboratory Certification

Normally done by an external agencyNormally done by an external agencyEvaluation is concerned with issues such as Evaluation is concerned with issues such as

Adequate spaceAdequate spaceVentilationVentilationStorageStorageHygieneHygiene

Specimen/Sample T ackingSpecimen/Sample T ackingSpecimen/Sample TrackingSpecimen/Sample Tracking

Vary among laboratoriesVary among laboratoriesMust maintain the unmistakable Must maintain the unmistakable connection between a set of analytical connection between a set of analytical data and the specimen and/or samples data and the specimen and/or samples from which they were obtained.from which they were obtained.Original source of specimen/sample (s) Original source of specimen/sample (s) must be recorded and unmistakably must be recorded and unmistakably connected with the set of analytical data.connected with the set of analytical data.

Documentation and Maintenance of Documentation and Maintenance of RecordsRecords

Maintenance of all records provide Maintenance of all records provide documentation which may be required in the documentation which may be required in the event of legal challenges due to repercussions ofevent of legal challenges due to repercussions ofevent of legal challenges due to repercussions of event of legal challenges due to repercussions of decisions based on the original analytical results.decisions based on the original analytical results.General guidelines followed in regulatedGeneral guidelines followed in regulatedGeneral guidelines followed in regulated General guidelines followed in regulated laboratories is to maintain records for at least laboratories is to maintain records for at least five yearsfive yearsLength of time over which laboratory records Length of time over which laboratory records should be maintained will vary with the situationshould be maintained will vary with the situation

Important questions to be answered for any Important questions to be answered for any l ti l i t tl ti l i t tanalytical instrumentanalytical instrument

What is the equipment being used for?What is the equipment being used for?

Is the instrument within specification and is the Is the instrument within specification and is the documentation to prove this available?documentation to prove this available?documentation to prove this available?documentation to prove this available?

If the instrument is not within specifications, howIf the instrument is not within specifications, howIf the instrument is not within specifications, how If the instrument is not within specifications, how much does it deviate by?much does it deviate by?

If the instrument is not within specifications what If the instrument is not within specifications what action has been taken to overcome the defect?action has been taken to overcome the defect?

Can the standards used to test and calibrate the Can the standards used to test and calibrate the instrument be traced back to national standards?instrument be traced back to national standards?

Q lit i L IdQ lit i L IdQuality is a Lousy IdeaQuality is a Lousy Idea--

If it’s Only an IdeaIf it’s Only an IdeaIf it s Only an IdeaIf it s Only an Idea

Ex: balance ??

True Value vs. MeasuredTrue Value vs. MeasuredTrue Value vs. Measured True Value vs. Measured ValueValue

True ValueTrue Value Measured ValueMeasured ValueThe known, accepted value of

The result of an individual’s p

a quantifiable property

measurement of a quantifiableproperty quantifiable property

A P i iA P i iAccuracy vs. PrecisionAccuracy vs. PrecisionAccuracyAccuracyH ll

PrecisionPrecisionH ll iHow well a

measurement agrees ith an

How well a series of measurements agree ith eachagrees with an

accepted valueagree with each other

A P i iA P i iAccuracy vs. PrecisionAccuracy vs. Precision

Systematic vs. Systematic vs. Random ErrorsRandom Errors

Systematic ErrorSystematic ErrorAvoidable error due

Random ErrorsRandom ErrorsUnavoidable errorsAvoidable error due

to controllable variables in a

Unavoidable errors that are always present in anyvariables in a

measurement. present in any measurement. Impossible toImpossible to eliminate

Q li C l MQ li C l MQuality Control MeasuresQuality Control Measures

• Standards and Calibration• Blanks• Recovery Studies• Recovery Studies• Precision and Accuracy Studiesy• Method Detection Limits• NJQLs

Standards and CalibrationStandards and CalibrationStandards and CalibrationStandards and CalibrationP d P h d St d d• Prepared vs. Purchased Standard

• Signals: Peak Area, Beer’s LawSignals: Peak Area, Beer s Law• Calibration Curves• Continuing Calibration Checks

I t l St d d• Internal Standards• Performance Testing.e o a ce es g

C lib i CC lib i CCalibration CurvesCalibration CurvesGraphical representation of the Graphical representation of the relationship between:relationship between:relationship between:relationship between:

• The concentration of the analyteand

• The analytical signaland

y g

Spectrophotometry: Luminescence• Can you tell the difference between how many marks are in each box?

Spectrophotometry: Luminescence

400 360Sensitivity of Absorbance Measurements

Spectrophotometry: Luminescence• Can you tell the difference between how many marks are in each box?

Spectrophotometry: Luminescence

0 40

Sensitivity of Luminescence Measurements

Calibration Curve for DDTCalibration Curve for DDT500

y = 9.3005x + 4.3313400

x 10

6

200

300

k ar

ea

R2 = 0 9989

100Peak

R = 0.9989

00 10 20 30 40 50 60

Parts per trillion DDT

Continuing CalibrationContinuing CalibrationContinuing Calibration Continuing Calibration VerificationVerification

• Many methods don’t require that daily calibration curves are preparedprepared

• A “calibration verification” is analyzed with each batch of samples

Sample BatchSample BatchSample BatchSample Batch• 10 - 20 samples (method defined)10 - 20 samples (method defined)

or less• Same matrix• Same sample prep and analysis• Same sample prep and analysis• Contains a full set of

QC samples

Internal StandardsInternal StandardsInternal StandardsInternal Standards• A compound chemically similar to p y

the analyteN d b i h• Not expected to be present in the samplep

• Cannot interfere in the analysis• Added to the calibration standards

and to the samples in identicaland to the samples in identical amounts.

Internal StandardsInternal StandardsInternal StandardsInternal Standards• Refines the calibration processRefines the calibration process• Analytical signals for calibration

standards are compared to those for internal standardsinternal standards

• Eliminates differences in random and systematic errors between samples and standardssamples and standards

Performance TestingPerformance TestingPerformance TestingPerformance TestingBlind samples submitted toBlind samples submitted to laboratories ?

??

?Labs must periodically analyze with acceptableanalyze with acceptable results in order to maintain accreditation

Bl k Bl k Bl kBl k Bl k Bl kBlanks, Blanks, BlanksBlanks, Blanks, Blanks• Laboratory Reagent Blanks

• Instrument Blanks

• Field Reagent Blanks

• Trip Blanks

Laboratory Reagent BlanksLaboratory Reagent BlanksLaboratory Reagent BlanksLaboratory Reagent Blanks• Contains every reagent used in theContains every reagent used in the

analysis• Is subjected to all analytical procedures• Must give signal below detection limit• Must give signal below detection limit• Most methods require one with every q y

batch

I Bl kI Bl kInstrument BlankInstrument Blank• A clean sample (e.g., distilled water)

processed through the instrumentalprocessed through the instrumental steps of the measurement process; used to determine instrument contaminationcontamination.

Field Reagent BlanksField Reagent BlanksField Reagent BlanksField Reagent BlanksP d i h l b k h• Prepared in the lab, taken to the field

O d t th li it• Opened at the sampling site, exposed to sampling equipment, p p g q p ,returned to the lab.

Trip BlanksTrip BlanksTrip BlanksTrip Blanks• Prepared in the lab, taken to thePrepared in the lab, taken to the

field

• Not opened

• Returned to the lab

• Not always required in EPA methods

R S diR S diRecovery StudiesRecovery Studies

• Matrix Spikes

• Laboratory Control Samples

• Surrogates .

Matrix SpikesMatrix SpikesMatrix SpikesMatrix Spikes• Sample spiked with a known• Sample spiked with a known

amount of analyte• Subjected to all sample prep and

analytical proceduresanalytical procedures• Determines the effect of the matrix

on analyte recovery• Normally one per batch• Normally one per batch

L b t C t l S lL b t C t l S lLaboratory Control SampleLaboratory Control Sample

S bj t d t ll l d• Analyte spiked into reagent water• Subjected to all sample prep and analytical procedures y p

L b C l S lL b C l S lLaboratory Control SampleLaboratory Control SampleAlso known as:Also known as:

• Laboratory Fortified Blank (LFB)• Quality Control Sample (QCS)

SurrogatesSurrogates• Similar to an internal standard• Added to all analytical samples,

d ll QC l iand to all QC samples to monitor method performance, usually p , yduring sample prepM h d f h ifi• Methods often have specific surrogate recovery criteriag y

• Most common in Organic methods

Q li C l MQ li C l MQuality Control MeasuresQuality Control Measures• Standards and Calibration

Bl k• Blanks• Recovery Studiesy• Precision and Accuracy Studies• Method Detection Limits• NJQLs• NJQLs

Precision and AccuracyPrecision and AccuracyPrecision and AccuracyPrecision and AccuracyR i d f i iti l tifi ti• Required for initial certification and annually thereafter

• A series of four laboratory control lsamples

• Must meet accuracy (recovery) and ust eet accu acy ( ecove y) a dprecision (standard deviation)

i t ft i th drequirements, often in method

P i i d AP i i d APrecision and AccuracyPrecision and Accuracy• Required with a change in

instrumentation or personnelinstrumentation or personnel• Specific to the analystp y

ititMethod Detection LimMethod Detection Limitit“The minimum concentration of a

b h b d dsubstance that can be measured and reported with 99% confidence that the panalyte concentration is greater than zero”zero

N.J.A.CN.J.A.C 77::1818 -- 11..77N.J.A.C N.J.A.C 77::18 18 11..77

M h d D i Li iM h d D i Li iMethod Detection LimitMethod Detection Limit• MDLs are determined according to

40 CFR part 136 Appendix B40 CFR, part 136, Appendix B• Seven replicate laboratory control p y

samples, analyzed for precisionM l i l d d d i i b• Multiply standard deviation by 3.14 (Student’s t- value)( )

Method Detection LimitMethod Detection LimitMethod Detection LimitMethod Detection Limit• Must be performed initially for

certificationM t t it i ifi d i• Must meet criteria specified in method

• Must be performed with change in instr mentation or test methodinstrumentation or test method

• Annually with ELCPy

New Jersey New Jersey Quantitation LimitsQuantitation Limits

(NJQLs)(NJQLs)(NJQLs)(NJQLs)• The minimum concentration of anThe minimum concentration of an

analyte that can be quantified with t ti ti l fidstatistical confidence

• 5 x MDL, for the NJ Lab5 x MDL, for the NJ Lab Certification Program

Variables that affect the quality of results

The educational background and training of the laboratorypersonnel

The condition of the specimensThe controls used in the test runsReagentsEquipmentThe interpretation of the res ltsThe interpretation of the resultsThe transcription of resultsThe reporting of resultsThe reporting of results

Errors in measurement

True value - this is an ideal concept which cannot beachievedachieved.

Accepted true value - the value approximating thetrue value, the difference between the two values istrue value, the difference between the two values isnegligible.

Error - the discrepancy between the result of ao e d sc epa cy be wee e esu o ameasurement and the true (or accepted true value).

Error

Error Error is the discrepancy between the resultError is the discrepancy between the result obtained in the testing process and its ‘True Value’ / ‘Accepted True Value’

Error

Sources of Error ReagentsReagentsStandardsTechniqueTechniqueEnvironmentSpecimen collection handling etcSpecimen collection, handling etc.Many more

ErrorTypes of Error

Pre-Analytical error yIncludes clerical error, wrong patient, wrong specimen drawn, specimen mis-handled, etc.Th h Q li A h l bThrough Quality Assurance measures, the laboratory tries to maintain control over these factors

Well trained phlebotomy staffp yUse of easy patient & specimen identification methods, such as bar code identification.Willingness to be information resource and / or trainers forWillingness to be information resource and / or trainers for physicians and floor personnel often involved with specimen collection.

Error

Types of ErrorAnalytical errorAnalytical error

Random or indeterminateHard or impossible to trace, ie fluctuations in elect. temp, effects of light, etc

Systematic or determinantHave a definite cause, ie piece of equipmentHave a definite cause, ie piece of equipment that fails to function properly, poorly trained personnel, contaminated reagent

Through Quality Control measures such asThrough Quality Control measures, such as always running controls, the laboratory limits these errors.

Error

Types of ErrorPost-Analytical errorPost Analytical error

Errors that occur after the testing process is complete.

Clerical errors very possible here as well.Test result fails to get to the physician in a timely manner

Quality Assurance measures must be implemented if problems identified.

(M i i th t b th h d t t(My opinion – these seem to be the hardest to control. )

Sources of error

Input data required - such as standards used, calibration values, andvalues of physical constants.

Inherent characteristics of the quantity being measured - e.g. CFTand HAI titre.

Instruments used - accuracy, repeatability.y, p y

Observer fallibility - reading errors, blunders, equipment selection,analysis and computation errors.

E i t t l i fl ff ti th tEnvironment - any external influences affecting the measurement.Theory assumed - validity of mathematical methods andapproximations.

Random Error

An error which varies in an unpredictable manner, in magnitudeand sign, when a large number of measurements of the samequantity are made under effectively identical conditions.Random errors create a characteristic spread of results for any testmethod and cannot be accounted for by applying corrections.y pp y gRandom errors are difficult to eliminate but repetition reduces theinfluences of random errors.Examples of random errors include errors in pipetting and changesExamples of random errors include errors in pipetting and changesin incubation period. Random errors can be minimized by training,supervision and adherence to standard operating procedures.

Random Errors

x

x x

x xx x

True x x x x

Value x x x

x x x

x

x

x

Systematic Error

An error which, in the course of a number of measurements ofthe same value of a given quantity remains constant whenthe same value of a given quantity, remains constant whenmeasurements are made under the same conditions, or variesaccording to a definite law when conditions change.

Systematic errors create a characteristic bias in the test resultsand can be accounted for by applying a correction.

Systematic errors may be induced by factors such as variations inSystematic errors may be induced by factors such as variations inincubation temperature, blockage of plate washer, change in thereagent batch or modifications in testing method.

Systematic Errors

x

x x x x x x x

True x

Value

Internal Quality Control ProgramInternal Quality Control Program for Serological Testing

An internal quality control program depend on the use ofinternal quality control (IQC) specimens, Shewhart ControlCharts, and the use of statistical methods for interpretation.

Internal Quality Control SpecimensInternal Quality Control Specimens

IQC specimens comprises either (1) in-house patient sera(single or pooled clinical samples) or (2) international serum(single or pooled clinical samples), or (2) international serumstandards with values within each clinically significant ranges.

Shewhart Control ChartsA Shewhart Control Chart depend on the use of IQC specimens and isdeveloped in the following manner:-

h i f l d dPut up the IQC specimen for at least 20 or more assay runs and recorddown the O.D./cut-off value or antibody titre (whichever is applicable).Calculate the mean and standard deviations (s.d.)Make a plot with the assay run on the x-axis, and O.D./cut-off orantibody titre on the y axis.Draw the following lines across the y-axis: mean, -3, -2, -2, 1, 2, and 3s.d.Plot the O.D./cut-off obtained for the IQC specimen for subsequent assayrunsMajor events such as changes in the batch no. of the kit and instrumentsused should be recorded on the chart.

Westgard rulesThe formulation of Westgard rules were based on statisticalmethods. Westgard rules are commonly used to analyse data inShewhart control chartsShewhart control charts.Westgard rules are used to define specific performance limits for aparticular assay and can be use to detect both random and systematicerrorserrors.There are six commonly used Westgard rules of which three arewarning rules and the other three mandatory rules.Th i l ti f i l h ld t i i f t tThe violation of warning rules should trigger a review of testprocedures, reagent performance and equipment calibration.The violation of mandatory rules should result in the rejection of the

lt bt i d ith ti t ’ l i th tresults obtained with patients’ serum samples in that assay.

Shewhart Chart

80

90

100

A

50

60

70

80+3 sd+2 sd+1 sd

Antibody U Target value

20

30

40

-3 sd

-2 sd-1 sd

Units

0

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

VZV IgG ELISA: Target Value = 49 U/ml

Assay Run

Warning rules

Warning 12SD : It is violated if the IQC value exceeds themean by ±2SD It is an event likely to occur normally in lessmean by ±2SD. It is an event likely to occur normally in lessthan 5% of cases.

Warning 22SD : It detects systematic errors and is violatedWarning 22SD : It detects systematic errors and is violatedwhen two consecutive IQC values exceed the mean on thesame side of the mean by ±2SD.

Warning 41SD : It is violated if four consecutive IQC valuesexceed the same limit (mean ± 1SD) and this may indicate theneed to perform instrument maintenance or reagent calibration.p g

Mandatory rules

Mandatory 13SD : It is violated when the IQC value exceedsthe mean by ±3SD The assay run is regarded as out of controlthe mean by ±3SD. The assay run is regarded as out of control.

Mandatory R4SD : It is only applied when the IQC is tested induplicate. This rule is violated when the difference in SDduplicate. This rule is violated when the difference in SDbetween the duplicates exceeds 4SD.

Mandatory 10x : This rule is violated when the last 10consecutive IQC values are on the same side of the mean ortarget value.

Westgard Rules: 1 3SD

80

90

100

A

50

60

70

80+3 sd+2 sd+1 sd

Antibody U Target value

20

30

40

-3 sd

-2 sd-1 sd

Units

0

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

VZV IgG ELISA: Target Value = 49 U/ml

Assay Run

Westgard Rules: 10X

80

90

100

A

50

60

70

80+3 sd+2 sd+1 sd

Antibody U Target value

20

30

40

-3 sd

-2 sd-1 sd

Units

0

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

VZV IgG ELISA: Target Value = 49 U/ml

Assay Run

Follow up action in the event of aFollow-up action in the event of a violation

There are three options as to the action to be taken in the event of aviolation of a Westgard rule:

Accept the test run in its entirety - this usually applies whenonly a warning rule is violated.R j t th h l t t thi li l hReject the whole test run - this applies only when amandatory rule is violated.Enlarge the greyzone and thus re-test range for that particularassay run - this option can be considered in the event of aviolation of either a warning or mandatory rule.

Relative sensitivities of tests (approx)

Usual operating range [Ab] or [Ag]

precipitationimmunoelectrophoresis 10 μg/ml - 1 mg/mlimmunoelectrophoresisdouble/radial diffusion

10 μg/ml 1 mg/ml

immunofluorescence 0.1 - 10 μg/ml

ELISA (colour) 0.1 - 10 ng/ml ELISA (colour) 0.1 10 ng/ml (chemiluminescence) 0.01 - 10 ng/ml

radioimmunoassay 0 01 10 ng/ml radioimmunoassay 0.01 - 10 ng/ml

CALCULATING THE RATES

A test is used in 50 people with disease and 50 p ppeople without.  These are the results:

Disease

+ -+ 48 3 51

Test+ 48 3 51

- 2 47 49

50 50 100

Disease

+ -+ 48 3 51

Test+- 2 47 49

50 50 100

Sensitivity = 48/50 = 96%Specificity = 47/50 = 94%Specificity = 47/50 = 94%Positive predictive value =  48/51 = 94%Negative predictive value = 47/49 = 96%Negative predictive value = 47/49 = 96%