Control Charts

These were first applied in clinical chemistry by Levey and Jennings.⁵ They are now widely used in haematology for both automated and manual procedures. Samples of the control specimen are included in every batch of patients’ specimens and the results are checked on a control chart. To check precision, it is not necessary to know the exact value of the control specimen. If, however, its value has been determined reliably by a reference method, the same material can also be used to check accuracy or to calibrate an instrument. If possible, controls with high, low and normal values should be used. It is advisable to use at least one control sample per batch, even if the batch is very small, also at set intervals during a large run, at least once for every 50 patient specimens. Because the controls are intended to simulate random sampling, they must be treated exactly like the patients’ specimens. The results obtained with the control samples can be plotted on a chart as described below.

The mean value and standard deviation (SD) of the control specimen should first be established in the laboratory where the tests are performed. Using arithmetic graph paper, a horizontal line is drawn to represent the mean (as a base), and on an appropriate scale of quantity and unit, lines representing +2SD and −2SD are drawn above and below the mean. The results of successive control sample measurements are plotted. If the test is satisfactory, sequential results oscillate about the mean value and <5% of the results fall outside 2SD. Figure 25.1 illustrates a control chart from an automated system; a similar principle can be used for simple methods where the data are plotted manually (Fig. 25.2).

Figure 25.1 Control chart (Levey–Jennings chart) with automated blood count analyser. The mean value for each component of the blood count is shown, together with the upper limit (UL) and lower limit (LL) for satisfactory performance, which have been set at ± 2SD.

Figure 25.2 Control chart for haemoglobinometry by manual method. The limits for satisfactory performance have been set at ± 2SD.

Any of the following indicates a fault in technique or in the instrument or reagent:

The fault may be in the reagents or the laboratory-ware or it may be caused by incorrect adjustment/calibration of the instrument or other equipment, e.g. pipettes, technical error or even clerical error in transcribing the results. Before an intensive investigation, the test should be repeated with another sample and the possibility must also be considered that the inconsistency may be the result of deterioration or infection of the batch of reagent material or insufficient mixing of the sample. This control process is unlikely to detect an error in an individual specimen, which can only be detected by correlation checks. For haemoglobinometry, it may be useful to use both whole blood and lysate in a quality control check because differences in results obtained with these two preparations help to identify errors resulting from incorrect dilution, inadequate mixing or failure of a reagent to bring about complete lysis. If the control specimen is included with each batch of tests during the course of a day, their measurements should not differ by more than the established CV. A trend of sequentially increasing or decreasing values with the repeated measurements is indicative of drift.

Correlation Check

Correlation check implies that any unexpected result of a test must be checked to see whether it can be explained on clinical grounds or whether it correlates with other tests. Thus, for example, unexpectedly higher or lower haemoglobin might be explained, e.g. by a blood transfusion or by a haemorrhage, respectively. A low MCHC should be confirmed by demonstrating hypochromic red cells on a Romanowsky-stained blood film; a high MCV must correlate with macrocytosis. Similarly, the blood films should be examined to confirm marked leucocytosis or leucopenia or thrombocytosis or thrombocytopenia, to distinguish between platelets and red cell fragments or conversely between giant platelets and normal-sized red cells or to check an erroneously high leucocyte count as a result of incompletely lysed red cells in haemoglobinopathies or liver disease. This morphological procedure can only be used if the blood smear is correctly made and stained; otherwise, blood cell morphology may be misleading.

Recording blood count data on cumulative report forms or charts is good clinical practice and provides an inbuilt quality control system by making it easy to detect an aberrant result when compared with a previously determined baseline. This is especially useful in detecting the occasional wild errors caused by incorrect labelling of the specimen, inadequate suspension of the blood before sampling, partial clotting of a blood sample or deterioration on storage. A discrepant result without apparent clinical reason must be suspect until confirmed by a repeat test on a fresh specimen. The occurrence of a contrasting discrepancy in two different specimens on the same day would suggest that two specimens have been mixed up. It is worthwhile emphasizing the importance of the blood film for quality control, especially as it may tend to be considered obsolete with the increasingly automated blood count systems. Any obvious discrepancy between the count obtained by the analyser and subjective impression of morphology should always be checked. A blood film will confirm or refute an abnormally high or low leucocyte or platelet count. An important cause of artefactual leucopenia and thrombocytopenia is partial clotting of the specimen – this may be revealed by the presence, in the film, of fibrin strands with a mass of aggregated platelets. The film will also identify anomalies due to the presence of cold agglutinins, a high leucocyte count due to incomplete lysis of red cells in haemoglobinopathies, etc.

External Quality Assessment

EQA is based on the evaluation by an outside agency of the analytical performance on specially supplied samples by a number of laboratories. The objective is to achieve between-laboratory and between-method comparability and it does not necessarily guarantee accuracy unless the specimens have been assayed by a reference laboratory alongside a reference preparation of known value.

EQA complements internal quality control and is a basic requirement for clinical laboratory certification and accreditation. Certification is a procedure by which a third party gives written assurance that a product, process or service conforms to specific requirements, whereas accreditation is a procedure by which an authoritative body gives formal recognition that a body or person is competent to carry out specific tasks. Accreditation standards related to clinical laboratories place emphasis on having an effective quality assurance system in place, with a commitment to meeting the needs of patients and their doctors as users of laboratory services and a need for continuous cycle of quality improvement at the centre of all policy-making operational decisions (see also Chapter 24, p. 579)

The term EQA, also known as proficiency testing (PT), was adopted in 1979 by a WHO working group on Quality Assurance of Health Laboratories. It is defined as: ‘a system whereby a set of reagents and techniques are assessed by an external source and the results of the testing laboratory are compared with those of an approved reference laboratory or agency’.⁷ It allows an individual laboratory to compare its performance for one or more tests or techniques against that of other laboratories. Thus, even when all precautions are taken to achieve accuracy and precision in the laboratory, errors arise that are only detectable by objective EQA of the performance of a number of laboratories on material which has been supplied specially for the purpose. EQAS are usually organized, nationally or regionally, as one or PT programmes. National schemes are frequently termed NEQAS (National External Quality Assessment Scheme).

EQA analysis of performance is retrospective and, in addition to inter-laboratory comparability, its main objective is to achieve the harmonization of methods and procedures or educational purposes. EQAS is not necessarily intended to measure accuracy except when the control material is assayed by a reference laboratory alongside a standard reference. The principle is that the same material is sent from a national or regional centre to at least 20 participating laboratories. Results are returned to the EQAS organizer, statistically analysed and a target value with its range of variability (SD) calculated. This provides evaluation of the performance of each individual participant and identifies outliers. It is important that surveys should be performed at regular intervals, although their frequency may vary, depending on the diagnostic importance of the particular test, how frequently they are requested and their technical reliability. It is recommended that at least two specimens should be distributed together for each survey, a minimum of four times each year. Availability of survey material will always influence distribution timings, frequency and quality due to this (e.g. bone marrow films lacking particles).

The main purposes of EQA are to ensure continuous reliable performance by individual laboratories and to achieve harmonization or concordance between laboratories. However, some analysers handle preserved blood differently from routine specimens and, even if correctly calibrated, different types of counter may differ in their responses to EQA samples. It may thus be necessary to analyse results separately for different groups of instruments. When there are unexplained differences in counts on EQA samples with different instruments in the same laboratory, counts should be made on fresh EDTA blood samples with the different instruments to ascertain their true comparability and, if necessary, recalibration of one should be undertaken to achieve concordance. Where appropriate, results from commercial kits and in-house methods should also be analysed in complementary groups. EQAS also have other additional complementary functions:

In summary, EQA is primarily intended to check the technical competence of individual laboratories, but it also provides an overview for assessing the state of the art and for identifying problems that have occurred in instruments, reagents or kits that may be affecting an entire group of users through no fault of their own. This provides a means for verification of manufacturers’ conformity to their claimed specifications and for monitoring product performance in the laboratory, a requirement in Europe in the context of the Directive on In-Vitro Diagnostic Medical Devices.⁸

More information of this type would be useful for determining the priorities for national or international agencies engaged in developing standards. EQA should be used for highlighting shortcomings and for evaluating methods, technologies and reference/standard preparations. It is important for a clinical laboratory to participate in EQAS for accreditation or certification, as clearly stated in ISO 9000, ISO Guides 25 and 58 and also in ISO 15189.⁹

Quantitative Tests

Youden (xy) plot

The Youden plot is a useful method for relating measurements on two samples in a survey to provide a graphic display and, when a participant’s results are unsatisfactory, for distinguishing between a consistent bias and random error. Results for the two samples are plotted on the horizontal (x) and the vertical (y) axis, respectively and the standard deviations (2SD or 3SD) for the two sets are drawn and interpreted as shown in outline in Figure 25.5.

Figure 25.5 Youden (xy) graph. The range of standard deviations (SDs) calculated from the overall results with sample x and sample y, respectively, are drawn on the x axis and the y axis; the individual paired results are then plotted on the graph. Results in the central square are satisfactory; those in B demonstrate a consistent bias with measurements that are too low (B1) or too high (B2), whereas results in other areas indicate random errors (inconsistency) in the two samples.

Semi-Quantitative Tests

Reactions such as lysis, agglutination or colour change are recorded as 0, 1+, 2+, 3+ or 4+. Assessment of performance should be based on extent of divergence from the target value, which might be a consensus of participant results or a referee’s results. Account must be taken of the diagnostic and clinical significance of an incorrect or confusing result.

EQA schemes should assess not only technical reliability but also professional competence in interpretation of the measurements. Thus, participants should be required to report on the technical significance of their results (i.e. whether within normal reference values for the specified method) and also on the clinical significance, taking account of any clinical information provided. Incorrect interpretation of a correct quantitative result is often due to lack of understanding of the concept of reference values and to the use of inappropriate reference ranges. It is important in the post-analytical phase of quality management for each laboratory to establish its own reference values for normal and for specific groups (e.g. smokers, pregnancy). With qualitative tests, too, account must be taken of the clinical significance of the answer. For each test it is necessary to decide if it is as serious to err by reporting a feature which does not exist as to miss an abnormality which is present and to penalize the score accordingly. Scoring may be either as a credit for correct answers or as a penalty for an incorrect answer. Thus, using the penalty method in G6PD screening, a correct answer scores zero, while an error will be graded depending on the implications of misclassification for the subsequent management of the patient: it is more serious to report a low level as normal (score 5) than a normal value as low (score 3), and when the true value is intermediate, it is more serious to report it as normal in a female (score 3) than in a male (score 1).

Interpretive Tests

In assessing qualitative or interpretive tests (e.g. stained blood smears for morphology), participant results are compared with the consensus obtained from a panel of referees or by concordance of at least 80% of the participants. Some schemes may use a panel of experts to produce a ‘Comments’ section in the report. The features reported are graded on their clinical significance, taking account of the specific medical condition, as follows:

All correct observations are given a positive score and false-positive observations are subtracted in accordance with the grading. The result is expressed as a percentage of the total possible score established by the referees. The more stringent limits have been proposed to provide greater sensitivity in diagnostic discrimination, but they may be impractical, resulting in an overestimate of the numbers of poor performers.^12,13

Similarly, scores on parasite films will depend on clinical significance; a negative result on a film containing Plasmodium falciparum or a microfilaria, will be penalized more than a Plasmodium ovale, reported as a Plasmodium malariae.

While few EQA schemes provide a bone marrow slide on a regular basis, due to the difficulties of producing large numbers of films from the same donor, the introduction of digital morphology EQAS may be useful in the future. Preparation and reporting on these films should be standardized from conception using guidelines.¹⁴

Control of Non-Compliance

When non-compliance has been detected, it is essential to ensure the reliability of the laboratory service. This requires the following steps:

Method^4,15

Method

Stored at 4°C, the product should retain its assigned value for at least several months or for 1–2 years if kept at −20°C.

Method

Simple Method for Producing QC Preparations for Individual Types of Blood Cell Material

The following method for control material preparation provides a suitable procedure for red blood cell, leucocyte and platelet counting by some semi-automated blood cell counters, but is not suitable for more complex automated systems. If kept at 4°C, this material is stable for a maximum of 3 weeks.

Method

Preparation of Surrogate Leucocytes

Chicken and turkey red blood cells are nucleated and, when fixed, their size is within the human leucocyte range as recognized on electronic cell counters. They are thus suitable to serve as surrogate leucocytes.⁴ This material may not be suitable for counting systems that are based on technologies other than impedance cell sizing.

Method

Reagent

Alsever’s solution. (A) Trisodium citrate, 16 g; NaCl, 8.2 g to 1 litre with water; (B) dextrose, 41 g to 1 litre with water. Store at 4°C. Immediately before use, mix equal volumes of A and B; filter through 0.2 μm micropore filter. EDTA solution. 100 g/l of K₂ EDTA in the Alsever’s solution; stable for 6 months at 4°C.⁴

Method

A simpler method for preserving platelets by adding prostaglandin E₁ to blood in ACD has been reported to provide a control preparation with stability of about 14 days.¹⁷