Quality assurance Quality assurance
Accuracy § Close of the estimated value to the true value § Only checked with use of reference materials Precision § Reproducibility of the assay § Can be controlled by replicate tests and repeated tests on previously measured specimens
§ Causes of inaccuracy / imprecision § Pre analytic phase § Specimen not representative of the in vivo state (leakage of sample) § Pre analytical deterioration of specimens (time of receipt should be registered) § Sample mix up § Analytic phase § Bias in the instrument § Sampling not representative of the specimen § Inconsistent dilution § Unreliable reagents (e.g. ISI variation) § Incomplete reactions § Carry over from a previous test § Inherent approximations / assumptions in the analyser’s function § Environmental effects on the analyser § Poor technique (staff training) § Internal quality control § Failure to undertake necessary IQA (Duplicate patient samples, collection and analysis of patient data) § Unreliable standards § Post analytic phase § Transcription errors § IT errors
Summary of Quality assurance procedures 1. Calibration with reference standards a. Instruments – i. Minimum 6 month intervals ii. If control chart or EQA indicate bias or fluctuation of results iii. After servicing b. Diluting systems i. Minimum 1-2 week intervals 2. Control chart with control material a. Done daily or with each batch of specimens 3. Delta check a. Daily on selected patients 4. Duplicate tests on two or three patients samples a. Done if the control chart or delta check show discrepancies 5. Analysis of constancy of MCV, MCHC, MCH a. Done daily 6. Correlation assessment of test reports a. Cumulative results – following previous tests b. Blood film – if unusual test results and/or counter flags present c. Clinical state 7. EQAS performance
Internal quality control § Repeat measurements on routine specimens § Measurements on specially prepared samples § Daily statistical analysis of data - continual evaluation of the reliability of the work of the laboratory
External quality assessment § Retrospective analysis of performance § NEQAS § WHO EQA scheme
Reference method § Exactly defined technique that provides sufficiently precise and accurate data for assessing the validity of other techniques
Selected method § A method that is directly comparable with and traceable to the international reference method – should be used for evaluation and validation of a proposed routine method
Working method § Intended for routine practice taking account of economy of labour and materials
Controls § Preparations used for either internal quality control or EQA. § Although some are assigned values they should not be used as standards because the values are usually only approximate and they are often only stable for a limited time
Reference Preparations § Preparations used for calibration - conform to national or international standards § Haemoglobin - Haemiglobincyanaide – national authorities certify preparations that conform to the national standard which is be compared to the international standard which can be obtained from the WHO
Blood cells § Need to use fixed cells (EDTA has too short a half life) § E.g. glutaraldehyde fixed erythrocytes
Quality Control materials § Obtained commercially or prepared locally § Blood collected into ACD or CPD anticoagulant rather than EDTA § It can then be manipulated by centrifugation to increase or decrease the RCC § Can be fixed with formaldelhyde and will last for 3 weeks § Animal blood can be used § Horse blood can be used for red cell QC - will last longer and is microcytic § Chicken and turkey blood can be used for testing leukocytes (the red cells are nucleated)
Analysis of data Standard deviation of controls § The control material should be dispensed into vials (using a mixing system to ensure it is homogenous) and the intertube variability should me measured in 3-10 vials § The SD or CV should be calculated
Control charts § A sample of control specimen should be included in every batch of patients specimens and the results should be checked on a control chart – will check precision of the result § If the control specimen’s value has been determined reliably then this can also be used to check accuracy § The results should be plotted and should be within +/- 2 SD of the mean of the control sample (<5% should lie outside the SD)
Internal QA checks § “Delta” check § Blood count from the same patient checked repeatedly over 2-3 weeks § Values should not change significantly providing the patients condition has not altered significantly § Patient data § Need enough through put (i.e. >100 samples /day) § MCV, MCHC, MCH should not vary providing the patient population is stable and samples from different sets of patients are not batched together. § Mean of 20 successive patients can be plotted and any drift in the calibration of the analyser will be recognised § Duplicate tests on patient specimens § Test of precision can be performed by repeat analysis on samples from two or three patients
§ Unexpected test results must be checked and explained § Clinical grounds § Blood film.
External quality assessment § Important to complement internal quality control § Same material is sent to numerous participating labs (at least 20) § The results are then sent to the EQAS centre and a target value and an acceptable range is established § Frequency of testing varies with the diagnostic importance of the test involved (e.g NEQAS produces specimens 4 weekly for blood counts and 3 monthly for most other tests)
Main purpose of EQAS § Ensure reliable performance of individual laboratories § Achieve concordance between different laboratories (but different analysers will handle preserved blood differently – and it may be necessary to handle results from different analysers differently) § Other functions § Collect information on the reliability of different methods, materials and equipment § Identify problems with some techniques requiring reaction from the manufacturer § Provide information for laboratory accreditation § Identify laboratories who provide a benchmarking standard § Recommending new procedures and training staff on best practice.
Analysis of EQA data § Deviation Index – same as the z score (i.e. = (result – mean)/SD) – if non-Gaussian distribution then can be calculated from the median
Consecutive monitoring § One method is to add 6 consecutive results together (maximum score rounded down to 3.5) and multiply by 6 - if score over 100 = persistent unsatisfactory performance (mean not allowed to be over 2.7)
Outwith consensus method § Refinement of the non-Gaussian procedure used by UK NEQAS for clotting results
§ Performance in any particular test is assessed from the grades obtained in two consecutive exercises. § Unsatisfactory performance is designated by D-D or E – with C, D, or E
Target values and bias § Target values can be obtained by either using the result from the labs with best performance or from experts using reference methods § The percentage bias can be calculated (Result-TV)/TV x100)
Youden (xy) plot § Two samples analysed and the results sample A plotted on x axis and B on y axis. § Lines then drawn demarcating the SD (or 2SD). § The results in the central block are satisfactory § Results in top R (B) and bottom L blocks (A) indicate consistent over or under reading § Results elsewhere indicate random errors
 Methodology checks § Sometimes useful to check separate components of a method (e.g. a survey might include identical whole blood or lysates from the same samples togherther with a prediluted haemoglobin sample)
Clinical significance § Using limits based on the SD may be too rigid in some cases and too lenient in others – depending on the clinical use of the results § Different limits are allowed for different tests e.g.
Semiquantitative tests § e.g. Agglutination (0, 1+, 2+……) can be assessed for example by divergence from a target value
Qualitative tests § e.g blood film analysis § Participant results are compared with the consensus obtained from a panel of referees or by concordance of 75% or more of participants
NEQAS – for clotting § Acceptable deviation is +/- 15% § Persistently out with consensus….. § Grading system A-E (A = best) § Unexpected results § Sample mix up § Carry over from a previous test § ISI variation
§ Further investigations § Who? How? Why? § Is it a systematic error § Check SOPs, and ensure no deviation from methodology with the individual responsible § Check pipette calibration / reagents etc § Retrospective check on the tests in which the NEQAS samples were run to ensure internal quality control satisfactory
§ Beware of INR reagents § May need to check the ISI and even perform calibration to determine the ISI § Risk of over dosing warfarin § If similar results to everyone else on NEQAS, may be that everyone else is making the same mistake as you.
§ PT variables § Thromboplastin reagent § Calculation of ISI and INR § Heparin § Normal range § APTT variables § Phospholipid § Heparin / warfarin § Normal range § Lupus
§ Requirements of the instrument § Space § Power § Water § Compressed air § Waste disposal § Storage of reagents § Accuracy / precision § Linearity § Effect of diluting the sample and remeasuring different dilutions § Precision § Samples with high, medium and low values analysed 10 times consecutively – SD or CV determined § Accuracy § Comparison with reference methods § Carry over § High concentration measured three times followed by a low concentration thee times – carryover % can be calculated = (low1-low3)/(high1-low3) x 100 § Sensitivity (determination of the smallest change in the analyte that gives a detectable result) § Specificity (extent of errors caused by interfering substances) § Throughput time and number specimens analysed per day § Assessment of safety § Cost § Fixed costs § Reagent costs § Reliability § Servicing costs § Emergency support § Need for staff training
SOPs § Written instructions intended to maintain consistent quality performance in the laboratory § Should cover all aspects including § Specimen collection (including handling urgent requests) § Laboratory safety § Data storage § Telephone reporting § Should be reviewed once a year
|
Coagulation >