From 1949, the American College of American Pathologists (CAP) first began to study the problem of indoor laboratory quality control (QC) in clinical laboratories. American scholars Levery and Jenning published the first laboratory ventricular quality control chart in 1950. Controls and clinical laboratory laboratories began to formally kick off the work of indoor quality control.
By the 1970s, laboratory quality control entered a new stage - total quality management, and implementation of Good
Laboratory Parctice (referred to as GLP). In the late 1980s, a unified standard of GLP was created and developed into the "certified laboratory" management stage.
The purpose of Total Quality Management is to prevent errors. The purpose of statistical control of quality control charts is to detect errors.
Statistical laboratory laboratory quality control is an important part of total quality management.
This chapter mainly introduces statistical internal control methods for immunological tests. Because ELISA is the most commonly used immunological test method in clinical practice, the ELISA test is used as an example to introduce related issues.
The Quality Control Laboratory of the Clinical Laboratory Center of the Ministry of Health began the nationwide evaluation of the quality of hepatitis B marker tests since 1988. It has always adopted this set of quality assessment methods, and has practiced, improved and improved it in practice, hoping to find it. An effective quality management road that suits China's national conditions.
5.1 Basic Concepts
5.1.1 Quality Control (QC)
Quality control is a management process that monitors the entire process, eliminates errors, prevents changes, and maintains the status quo in standardization. This process is done through a feedback loop.
1) Determine the object of control;
2) Specify the criteria for the control object (expected value);
3) Formulate or select control methods and methods;
3) measuring actual data;
4) Compare or compare the difference between the actual data and the expected value and explain the reason for this difference. Beyond the predetermined error range, the alarm signal is issued and the feedback channel is interrupted.
5) Take action to resolve differences. The means of restitution (the original standard state) works.
Quality control mainly uses quality control charts. The quality control chart is a graph that compares the performance data of a test with the calculated expected "control limit." This kind of performance data is selected in chronological order when it is conducted according to the regulations. Its purpose is to detect "retrievable" causes of variation in the inspection process. The reason for the "recoverable" nature of errors is to exclude other causes other than random errors. The "control limit" is calculated by statistics. We will introduce it in detail later (see
5.3. Indoor Quality Control Procedures).
5.1.2 Error experimental errors are divided into three types: systematic errors, random errors and errors.
Systematic error refers to the error that a series of measurement results have the same tendency as the true value or the target value. It has obvious regularity and can be repeated under certain conditions. It can be prevented and corrected through quality control.
Random error, also known as accidental error, is an accidental, unpredictable error that is difficult to avoid and correct. The distribution of random errors in inspection work conforms to the normal distribution law.
Fault error is human error. By strengthening laboratory management and quality control work can be avoided.
5.1.3 In the normal distribution and standard deviation ELISA test, when the same sample is tested more than 20 times, it will be found that this set of data (referred to as the absorbance of the measurement results) is distributed on both sides of the mean, most of which are concentrated around the mean value. If the measured value is the abscissa and the frequency of occurrence is plotted on the ordinate, a bell-shaped graph can be drawn. As shown in Figure 5-1, the top of the clock is the mean, and the other values ​​are symmetrically distributed around the mean. This is the normal distribution.
The area called probability below the normal curve is expressed by the mean (X) and standard deviation (SD) of the commonly used samples. The calculation method is as follows:
The relationship between the mean, standard deviation, and probability is as follows:
X±1SD, probability 0.68 X±2SD, probability 0.95 X±3SD, probability 0.99
In other words, when ELSIA detects the same sample for a specified number of times, the data within ±1 SD that is close to the mean (X) account for 68% of the data, and the data distributed within the range of X±2SD occupies the population. 95%
The data distributed within the range of X±3SD accounted for 99% of the total. When we require the test results to be qualified within the range of X±2SD, 95% of the data may be qualified.
5.1.4 The value of a true value measured using the exact, most ideal deterministic method is called the true value. The true value is generally undetectable. The value measured by a reliable deterministic method, called the target value, is usually expressed as a target value.
5.1.5 Accuracy
It refers to the degree to which the measurement result is close to the true value (or target value). Accuracy cannot be expressed numerically and is often measured in terms of inaccuracy. The degree of deviation between the measurement result and the target value is called deviation, which indicates the inaccuracy of the test.
Absolute Deviation = Mean of Test - True Value (or Target Value)
Relative Deviation = Absolute Deviation True Value (or Target Value) × 100%
5.1.6 Precision
It refers to the degree of closeness between each measurement result and the average value when the same sample is repeatedly measured, that is, the degree of coincidence between repeated measurement values.
5.1.7 Standards
1. Internationally certified materials that are calibrated by the WHO or a corresponding organization using determined, accepted, accurate physical or chemical methods.
2. International biological activity standards The materials of international active units calibrated by WHO or corresponding organizations according to biological reactions.
3. Reference Standards Serum The statutory material produced by the National Standardization Organization according to international standardization. Can be used to identify the accuracy of instruments and methods.
5.1.8 Clinical decision leuel
When the concentration of a test substance reaches a certain level, clinicians must adopt medical measures. This concentration of the analyte is referred to as a clinically decisive level.
5.2 Quality Control Serum quality control serum is the target serum, add one or several parts to each routine test, and get the results to understand the situation of this test. If the result of the quality control serum test can control the error to be within a certain range, it means that no error of the test is allowed. If there is an abnormal result that exceeds the allowable error range, it indicates that the test is unqualified, and the cause should be found. After correction, the sample to be tested shall be re-examined. Therefore, quality control serum plays an important role in quality control work.
5.2.1 Use of quality control serum The serum of hepatitis B marker substance prepared by the Ministry of Health Clinical Inspection Center can be kept at the constant value of -20°C for half a year.
Melting in the frozen state should be mixed first. The unused part can be stored at 4°C for 5 days. Should not be repeated ice melt or self-packing. The quality control serum required for carrying out an internal quality control work for an inspection is generally prepared at a dosage of 3-6 months. The self-made indefinite quality control serum should be prepared for the next batch of quality control serum before a batch of quality control serum will be used up. Quality control serum requires stable performance, long-term potency remains unchanged, and its physical and chemical properties should be similar to the patient sample, so as to effectively play a monitoring role.
5.2.2 Threshold Quality Control Serum quality controls both serum and undefined values. If only one quality control serum is used, it is generally set at the point where the normal value and the abnormal value meet. When the qualitative measurement is at a weakly positive level, it is called the critical value. The establishment of the critical value of hepatitis B markers should be based on clinical requirements to provide a unified standard for the determination of weak positives.
The critical value control serum can be used as a positive control in the kit and a third control other than the negative control. It can sensitively reflect the detection level of the kit and ensure that the weak positive reaction specimen is not missed.
5.2.3 Preparation of quality control serum Each laboratory can use the quality control serum provided by the clinical center according to its own conditions, or prepare the quality control serum used in the laboratory by the following methods (for example, hepatitis B quality control serum).
1) Collect fresh positive serum without hemolysis, jaundice, or bacterial contamination.
2) Heat at 56°C for 10 hours.
3) Centrifuge or filter to remove precipitate.
4) Dilute the collected serum to the desired concentration with 10% calf serum or normal human serum (PBS buffer). If it can be diluted with normal human serum better, because its composition is closer to the test sample.
5) Remove bacteria by suction filtration. Dispense the ampoules in a single use, seal them, and store them at 20°C for later use. Do not freeze and thaw repeatedly.
The level of detection required by the seized object is often considered to be the level that the quality of control serum should select. If there are other requirements for the test, the required concentration of controls should be added.
6) Calibration content. The mean value of the 20-30 test results deletion >±2 SD data was used as a target value, and it was measured in comparison with a known fixed serum.
5.3 Test Results of Clinical Tests for Indoor Quality Control Procedures It is unlikely that there will be no error between each time or day. To determine the allowable error range, clinically do not cause misdiagnosis and missed diagnosis, the following steps to determine the scope of quality control.
1) Measurement error under optimal conditions.
2) Errors of known values ​​of serum under normal test conditions.
3) The error of the unknown serum under normal test conditions.
4) Clinical application requirements. For any test, an allowable error range should be determined. The precondition is to meet the clinical requirements. If the allowable error is set too small, there is no clinical significance, but it takes a lot of manpower, material resources, and time to comply with this rule. On the contrary, if the allowable error is set too large, the monitoring system will not detect the clinically detected error and lose the significance of quality control.
5.3.1 Determination of optimal conditional control serum variation (abbreviated as OCV) under best conditions 20-30 times of the quality control serum in the best conditions of the laboratory (including operators, reagents, instruments, etc.) Measured results were calculated and the mean and standard deviation (SD) of this set of data were calculated to indicate the best quality of work in the laboratory.
An example is the determination of OCV in the HBsAg ELISA assay. The quality control serum used is critical serum,
The HBsAg concentration is 5 ng/ml. In this laboratory, the best-selected and most skilled technicians are selected for serious and specific measurement. The best kit is selected. Before testing, the constant box, sampler, etc. are carefully corrected, adjusted, and debugged. The pipette tips, etc., are tested under the best and most ideal conditions. In addition to control serum, negative controls and positive controls were measured simultaneously. The two measurements were performed to obtain two absorbance values ​​(A values), and X was determined. For 20 consecutive times, find 20
X, that is X1 ... X20. From these 20 data, X and SD of OCV are found.
5.3.2 Determination of routine conditions variance-known value (RCVK) under known conditions.
The technicians doing routine inspections will place the quality control serum in the routine test samples under the conditions of routine inspections and perform 20 tests. The results are calculated using the same OCV method. It is generally accepted that SD of RCV is acceptable within twice the SD of OCV.
If it is too large, you should look for the cause and bring it closer to the OCV's SD value. After the improvement of laboratory conditions (such as correcting the sampler, correcting the plate washing operation, adjusting the temperature of the incubation, etc.), the RCVK measurement is performed again. If the SD value of RCVK is smaller, it means that OCV is not measured under optimal conditions and OCV should be measured again. Under normal conditions, RCVK is definitely better than
OCV is big. Through the quality control of each condition, the RCVK data is as close as possible to the OCV value. RCVK's data reflect the quality of the daily work of the laboratory, used as a quality control chart, to control the results of indoor inspections, the results of daily inspections, reports can be issued.
5.3.3 Under normal conditions, the measurement of unknown conditions (control equation variancl-unknown value, abbreviated as RCVU) is sometimes performed for RCVU measurement to avoid subjectivity.
The measurement procedure is the same as RCVK, but the operator of the test does not know the quality of the control serum, or performs routine tests under the condition that the operator does not know which of the controls is blood-throttled blood to exclude the operator's subjectivity. This is no longer an illustration.
5.3.4 Quality Control Chart Through the above three steps, you can start the indoor quality control chart, according to RCVK and SD for quality control block diagram. The quality control chart can be used to monitor the results of each test, when there is no replacement of another batch of reagents and another lot of quality control serum.
The quality control chart can be continued in succession.
The S/CO value of the quality control serum is lower than the range of -2SD, and it is an "alarm". You should look for the cause and record the reason for the detection on the quality control chart.
In the ELISA test, the error allowable range of various test items is to be concluded in practice. The above is only an example to illustrate the quality control method, but it is not conclusive. 2SD is a generally accepted tolerance limit. When each batch of control is given a quality control serum, more than 2SD at one time should be used as an "alarm" and the second exceeding 2SD is "out of control". When the quality control process is out of control and a runaway occurs, the cause should be found, usually caused by the failure of the kit or the quality control serum. Replace the kit or replace the quality control serum, find out the reason and re-test. If the test results still do not meet the requirements or the cause cannot be found, the OCV test should be repeated. If the OCV test results are still good, there is a problem with the regular operation.
It is generally considered that: 1 exceeds 3 SD at a time; 2 exceeds 2 SD continuously; 2 33-5 times continuously falls within 2 SD of one side;
45 to 7 consecutive sideways to the horizontal axis are out of control.
In cases 3 and 4, reliance on records alone is often not easy to detect, but it can be clearly seen on the quality control chart.
5.3.5 Statistical Methods - "Immediate" Quality Control The quality control method described above is basically the same as the quality control method for clinical chemistry determination, but ELISA has its own particularities.
The most appropriate quality control method is yet to be studied and established. Some laboratories do not perform daily ELISA tests, but ELSIA
The efficacy of the kit is short, and it is difficult to test it 20 times with a batch of kits. Using the “immediate method†quality control statistical method, only three consecutive measurements are needed to perform quality control on the third test result.
The specific calculation method for establishing the "immediate method" is as follows:
1) Sort the measured values ​​first to second
2) Calculate X and SD.
3) Calculate the SDI upper limit and the SDI lower limit value.
4) Compare the SDI upper limit, SDI lower limit value with the number in the SDI value. When SDI upper limit value and SDI lower limit value are less than n2SD, it indicates that it is within the control range, it can continue to measure down and continue to repeat the above calculations; when SDI upper limit and SDI lower limit have a value between n2SD and n2SD, It indicates that the value is in the "alarm" status in the 2SD~3SD range. When the SDI upper limit and the SDI lower limit have values ​​>n2SD, it means that the value is outside the 3SD range and is "runaway". Values ​​in the "alarm" and "out of control" states should be discarded and the control serum and patient samples should be re-measured. What is left is only the number of out-of-control values, and other measured values ​​can still be used.
Instantly controlled statistical methods for quality control of ELISA assays.
When the detected value is more than 20 times, it is no longer necessary to use the "immediate method" quality control statistical calculation, and can be transferred to the quality control of the conventional quality control chart. The SD obtained from the first 20 numerical values ​​is used as a quality control framework diagram, and the 21st numerical value is clicked.
5.4 External quality assessment (EQA)
Inter-room quality assessment is referred to as inter-office quality assessment. The quality control center uses a series of methods to continuously and objectively evaluate the test results of each laboratory, and finds that the inaccuracy of indoor quality control is not easy to find. The difference between the results and help to make corrections are comparable. The laboratory test results were reported to the quality control center. After statistical analysis, the results were compared with each other. This evaluation cannot control the test reports issued by each laboratory every day, but is a retrospective evaluation. The indoor quality control mainly monitors the precision of the test results, while the inter-chamber quality control mainly controls the accuracy of the test results and cannot replace each other. The laboratory participating in the quality assessment should first perform indoor quality control.
5.4.1 Method of Inter-Quality Assessment
1. Conducting quality control investigations This is a common form of quality assessment at home and abroad. Department of clinical examination center for the evaluation of hepatitis B markers ELISA quality assessment between the use of regular quality control materials to the various laboratories, each laboratory on the date of the inspection, and report the inspection results to the Ministry of Health Center. The Department's clinical inspection center has statistically analyzed and sent the evaluation results back to the laboratories. Through the evaluation, each laboratory understands the quality of the work in the room, finds gaps, and tries to improve them so as to continuously improve the quality of the tests.
This evaluation method has certain drawbacks, that is, each laboratory often treats the quality control materials in a special way. During the test, special kits are used to select special technicians for inspection, and some laboratories modify the results with each other. This makes the results of EQA not reflect the daily work level of the laboratory.
2. Send observers to the laboratory to conduct reagent surveys. This survey does not require advance notice. Temporary observers are sent to the laboratory to designate routine methods to test a set of specimens for evaluation.
This kind of survey method can easily find practical problems in the laboratory, and can directly give guidance and help, solve problems, and improve inspection quality. This type of survey can often use real samples to avoid some of the disadvantages of using quality controls.