Quality Assurance and Quality Control

Published on 11/02/2015 by admin

Filed under Allergy and Immunology

Last modified 22/04/2025

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 2 (1 votes)

This article have been viewed 3092 times

Quality Assurance and Quality Control

The introduction of routine quality assurance (QA) programs and quality control (QC) in the clinical laboratory was a major advance in improving the accuracy and reliability of testing. This process ensures the clinician ordering the test that the testing method has been done in the best possible way to provide the most useful information in diagnosing or managing a patient. QA indicators and QC are tools to ensure that reported laboratory results are of the highest quality.

Clinical Laboratory Regulatory and Accrediting Organizations

The U.S. Congress enacted the Clinical Laboratory Improvement Amendments of 1988 (CLIA ’88) in response to the concerns about laboratory testing errors. The final CLIA rule, Laboratory Requirements Relating to Quality Systems and Certain Personnel Qualification, was published in the Federal Register in January 2003. Enactment of the CLIA established a minimum threshold for all aspects of clinical laboratory testing.

Voluntary standards have been set by The Joint Commission (TJC), the Commission on Office Laboratory Accreditation (COLA), and the College of American Pathologists (CAP).

A more recent development in voluntary accreditation aimed at improving quality was the introduction of ISO 15189. The International Organization for Standardization (ISO) is the world’s largest nongovernmental developer and publisher of international standards. ISO standards and certification are widely used by industry, but ISO 15189 has now been formulated for clinical laboratories. ISO 15189 has gained some standing abroad as a mandatory accreditation, such as in Australia, Ontario, and many European countries. In the United States, ISO 15189 accreditation remains optional. Requirements for quality and competence in ISO 15189 are unique because it takes into consideration the specific requirements of the medical environment and the importance of the medical laboratory to patient care. CAP 15189 is a voluntary nonregulated accreditation to the ISO 15189:2007 standard as published by ISO. CAP 15189 does not replace CAP’s CLIA-based Laboratory Accreditation Program, but complements CAP accreditation and other quality systems by optimizing processes to improve patient care, strengthen the deployment of quality standard, reduce errors and risk, and control costs.

Nonanalytical Factors Related To Testing Accuracy

Qualified Personnel

The competence of personnel is an important determinant of the quality of the laboratory result. Only properly certified personnel can perform nonwaived assays (see Chapter 9 for levels of laboratory testing).

Laboratory Procedure Manual

A complete laboratory procedure manual for all analytical procedures performed in the laboratory must be provided. The manual must be reviewed regularly, in some cases annually, by the supervisory staff and updated as needed.

A complete laboratory procedure manual for all procedures performed in the laboratory must be provided. The Clinical and Laboratory Standards Institute (CLSI) recommends that these manuals follow a specific pattern for how procedures are organized (Box 7-1).

Inaccurate Results

Inaccuracies in testing can be systematic or sporadic. Systematic errors can be eliminated by a program that monitors equipment, reagents, and other supplies. Sporadic or isolated errors in technique can produce false-positive and false-negative results, depending on the technique used for testing (Box 7-2).

An important aspect of quality is documentation of results. CLIA regulations mandate that any problem or situation that might affect the outcome of a test result be recorded and reported. These incidents can involve specimens that are improperly collected, labeled, or transported to the laboratory or problems concerning prolonged turnaround times for test results. There must be a reasonable attempt to correct the problems or situation and all steps in this process must be documented.

Errors Related to Phase of Testing

The Institute for Quality Laboratory Medicine has developed measures to evaluate quality in the laboratory based on the preanalytical, analytical, and postanalytical phases of testing.

Errors occurring during the analytical phase of testing in clinical laboratories are now relatively rare. Currently, most laboratory errors are related to the preanalytical and postanalytical phases of testing. To guarantee the highest quality laboratory results and to comply with CLIA regulations, various preanalytical factors need to be considered (Boxes 7-3 and 7-4).

Quality Descriptors

Quality control activities include monitoring the performance of laboratory instruments, reagents, other testing products, and equipment. A written record of QC activities for each procedure or function should include details of deviation from the usual results, problems, or failures in functioning or in the analytical procedure and any corrective action taken in response to these problems. All solutions and kits used in testing must be carefully checked before actually being used for testing patient samples.

Definitions

Quality control consists of procedures used to detect errors that result from test system failure, adverse environmental conditions, and differences between technologists, as well as the monitoring of the accuracy and precision of test performance over time. Accrediting agencies require monitoring and documentation of quality assessment records. Documentation of QC includes preventive maintenance records, temperature charts, and QC charts for specific assays.

Quality control monitors the accuracy and reproducibility of results through the use of control specimens. The diagnostic usefulness of a test and its procedure is assessed by using statistical evaluations, such as descriptions of the accuracy and reliability of the test and its methodology.

The terms accuracy and precision are often used to describe quality. Accuracy describes how close a test result is to the true value. Precision describes how close the test results are to one another when repeated analyses of the same specimen are performed. It is possible to achieve great precision, with all laboratory personnel who perform the same procedure arriving at the same answer, but without accuracy if the answer does not represent the actual value being tested. Accuracy can be improved by the following:

The precision of a test, its reproducibility, may be expressed as the standard deviation (SD) or derived coefficient of variation (CV). A procedure may be extremely accurate, yet so difficult to perform that individual laboratory personnel are unable to arrive at values that are close enough to be clinically meaningful.

Precision can be ensured by the proper inclusion of standards, reference samples, and/or control solutions, statistically valid, replicate determinations of a single sample, or duplicate determinations of sufficient numbers of unknown samples. Day-to-day and between-run precision are measured by the inclusion of blind samples and control specimens.

Sensitivity and Specificity

Laboratory results should provide medically useful information, including the specificity and sensitivity of the tests being ordered and reported. Both specificity and sensitivity are desirable characteristics for a test, but in different clinical situations, one is generally preferred over the other.

Assessing the sensitivity and specificity of a test requires four factors: tests positive, tests negative, disease present (positive), and disease absent (negative). True positives are subjects who have a positive test result and who also have the disease in question. True negatives represent those who have a negative test result but do not have the disease. False positives are those who have a positive test result but do not have the disease. False negatives are those who have a negative test result but do have the disease.

Predictive Values

To assess the predictive value (PV) for a test, the sensitivity, specificity, and prevalence of the disease in the population being studied must be known. The prevalence of a disease is the proportion of a population who have the disease. The incidence is the number of subjects found to have the disease within a defined period, such as 1 year, in a population of 100,000.

A positive predictive value for a test indicates the number of patients with an abnormal test result who have the disease compared with all patients with an abnormal result:

Positive PV=Number of patients with disease   and with abnormal test resultsTotal number of patients with        abnormal test results

image

Positive PV=True positivesTrue positives+False positives

image

A negative predictive value for a test indicates the number of patients with a normal test result who do not have the disease compared with all patients with a normal (negative) result:

Negative PV=True negativesTrue negatives+False negatives

image

Monitoring Quality

Proficiency Testing

Proficiency testing (PT) is incorporated into the CLIA requirements. In addition to the use of internal QC programs, each laboratory should participate in an external PT program as a means of verification of laboratory accuracy. Periodically, a laboratory tests a specimen that has been provided by a government agency, professional society, or commercial company. Identical samples are sent to a group of laboratories participating in the PT program. Each laboratory analyzes the specimen, reports the results to the agency, and is evaluated and graded on those results compared with results from other laboratories. In this way, quality control between laboratories is monitored.

Control Specimens

A QC program for the laboratory uses a control specimen, a specimen with a known value that is similar in composition to the patient’s blood. A control specimen must be carried through the entire test procedure and treated in exactly the same way as any unknown specimen; it must be affected by all the variables that affect the unknown specimen. Control specimens are used because repeated determinations on the same or different portions (or aliquots) of the same sample will not give identical values. Many factors can produce variations in laboratory analyses. With a properly designed control system, it is possible to monitor testing variables.

If the control value in a determination is out of the acceptable range (out of control), one or more of the following factors may be responsible:

Reference Range Statistics

In analytical immunology and serology testing using methods such as enzyme immunoassay, quantitative reference range statistics can be used. Statistically, the reference range for a particular measurement is usually related to a normal, bell-shaped curve (Fig. 7-1). This Gaussian curve has been shown to be correct for almost all types of biological, chemical, and physical measurements. A statistically valid series of individuals who are thought to represent a normal healthy group are measured and the average value is calculated. This mathematical average is defined as the mean (X¯image, called the X-bar). The distribution of all values around the average for the particular group measured is described statistically by SD.

Use of the mean, median, and mode is explained in the following example:

A series of results reported for a laboratory test on seven different specimens = 7, 2, 3, 6, 5, 4, and 2

The mean is the mathematical average and is calculated by taking the sum of the values (29) and dividing by the number of values (7) in the list. The mean is 4.1 (rounded off to 4).

The median equals the middle value. To find the median, the list of numbers must first be ranked according to magnitude: 2, 2, 3, 4, 5, 6, 7. There are seven values in the list, and the median is the middle value, 4.

The mode is the value most frequently occurring value, or 2 in this example.

The standard deviation is the square root of the variance of the values in any one observation or in a series of test results. In a normal population, 68% of the values will be clustered above and below the average and defined statistically as falling within the first standard deviation (±1 SD; see Fig. 8-1). The second standard deviation (±2 SD) represents 95% of the values falling equally above and below the average, and 99.7% is included within the third standard deviation (±3 SD). (Again, variations occur equally above and below the average value [or mean] for any measurement.) Thus, in determining reference values for a particular measurement, a statistically valid series of people are chosen and assumed to represent a healthy population. These people are then tested and the results are averaged.

The reference range is the range of values that includes 95% of the test results for a healthy reference population. The term replaces normal values, or normal range. The limits (or range) of normal are defined in terms of the standard deviation from the average value. Thus, normal or reference values are stated as a range of values in terms of SD units.

Testing Outcomes

Before physicians can determine whether a patient has a disease, they must know what is acceptable for a representative population of similar patients (e.g., same age, same gender, same ethnicity), as well as the analytical method used for an assay. Furthermore, an individual may show daily, circadian, and physiologic variations.

Biometrics, the science of statistics applied to biologic observations, has been a rapidly expanding field that attempts to describe these variations. The selection of a group on whom to base reference groups is another problem confronting the individual laboratory. To develop reference values (normal values), the proper statistical tools of sampling, selection of the comparison group, and analysis of data must be used by the manufacturer of testing kits or individual laboratories.

Although generally accepted values are published, reference values will vary, especially between laboratories and between geographic locations. Each laboratory must give the physician information concerning the range of reference values for that particular laboratory.

Validating New Procedures

The QC program also determines how new procedures are validated before being included as one of the methods routinely used by the laboratory. Each laboratory must determine the reproducibility (or confidence limits) for each procedure used and establish acceptable limits of variation for control specimens. The QC program includes calculation of the mean (or average value) and standard deviation and the preparation of control charts for each procedure.

Parallel Testing of Test Kits

The requirements for the parallel testing of test kits differ depending on your accreditation agency. For example, the CAP and Clinical Laboratory Improvement Acts (CLIAs) have slightly different requirements. There is also a difference in the requirements, depending on the circumstances. Are you changing manufacturers and tests kits, or are you only changing lot numbers for the same kit?

The CAP asserts that CLIA-waived assays are not recognized and the laboratory must treat all tests the same way. It is best to check the immunology checklist at www.cap.org for the latest revisions to questions related to kits. Currently, CAP checklist question IMM.33150 (phase II) is “Are new reagent lots checked against old reagent lots, or with suitable reference material before, or concurrently with, being placed in service?”

A CLIA inspection focuses on the following:

Validation of a New Procedure Write-Up

Each student should develop a procedure checklist following the CLSI procedural format. A manufacturer’s package insert or book should be used as a source of information. After completing the CLSI procedural protocol, a fellow student should rate the write-up.

Procedure Validation Checklist Example: Traditional Screening Test for Infectious Mononucleosis

Format Procedure Details Evaluation of
Write-Up
Acceptable: Yes/No
(add comments as needed)
Title Paul-Bunnell Screening Test for Infectious Mononucleosis Is the title defined and specific?  
Purpose or principle of assay The Paul-Bunnell test is a hemagglutination test designed to detect heterophil antibodies in patient serum when mixed with antigen-bearing sheep erythrocytes. Dilutions of inactivated patient serum are mixed with sheep erythrocytes, incubated, centrifuged, and macroscopically examined for agglutination. Positive reactions are preliminarily associated with the manifestation of infectious mononucleosis. Is the principle or purpose of the assay clearly stated?  
Specimen Collection and Preparation
Preliminary specimen preparation No special preparation of the patient is required before specimen collection. The patient must be positively identified when the specimen is collected.
The specimen should be labeled at the bedside and include the patient’s full name, date the specimen is collected, patient’s hospital identification number, and phlebotomist’s initials.
Blood should be drawn by aseptic technique.
The required specimen is a minimum of 2 mL of clotted blood (red-topped evacuated tube).
Centrifuge the tube of blood and remove an aliquot of clear serum.
The presence of hemolysis makes the specimen unsuitable for testing.
Inactivate the serum at 56° C for 30 min before testing.
Are the specimen collection requirements clearly stated?
Are there any special specimen processing requirements stated?
 
Reagents, supplies, and equipment 2% suspension of washed sheep cells in normal saline (prepared by pipetting 0.2 mL of packed erythrocytes into 9.8 mL of saline)
0.9% sodium chloride (normal physiologic saline)
12- × 75-mm test tubes
Note: The cell should be no more than 1 wk old.
Graduated serologic pipettes
Centrifuge
37° C incubator (optional)
Are all of the necessary reagents, supplies, and equipment listed?  
Quality Control
Positive control serum; negative control serum A known positive control should be run concurrently. Are the QC requirements stated?  
Procedural steps

1. Label two sets of test tubes. Each set should consist of 10 tubes.

2. Pipette 0.5 mL of saline into tube 1 and 0.25 mL of saline into each of the remaining nine tubes.

3. To the first set of tubes, add 0.1 mL of patient’s inactivated serum to the first tube; mix and transfer 0.25 mL of the dilution to the second tube; mix and transfer 0.25 mL of the dilution to the third tube. Repeat this process to tube 10. Discard 0.25 mL from the final tube, tube 10.

4. To the second set of tubes, add 0.1 mL of the control serum and proceed to dilute it as in step 3.

5. Add 0.1 mL of 2% sheep cells to each tube.

6. Gently shake the tubes until mixed.

7. Incubate the tubes at 37° C for 1 hr or overnight at room temperature.

8. Centrifuge the tubes for 1 min at 1500 rpm.

9. Gently shake each tube and examine macroscopically for agglutination.

10. Record the results.

Are the steps in the procedure understandable?
Can the procedure be performed as described?   Reporting Results Positive reaction A titer >1:56 is considered to be a positive presumptive test.     Negative reaction The antigens on sheep erythrocytes are associated with infectious mononucleosis, serum sickness, and the Forssman antigen.     Procedural notes   Are the criteria for acceptable results clearly defined?   Sources of error False-positive reactions have been observed in conditions such as hepatitis infection and Hodgkin’s disease. An improperly inactivated serum will produce hemolysis.     Limitations The test is only indicative of the presence or absence of heterophil antibodies.
Demonstrating agglutination by using sheep erythrocytes does not make a distinction between antibodies associated with infectious mononucleosis, serum sickness, or the Forssman antigen.
Heterophil antibody assay lacks sensitivity as a diagnostic criterion for infectious mononucleosis.
Sheep erythrocytes are less sensitive than erythrocytes from other species such as the horse.
A patient may take as long as 3 mo to develop a detectable heterophil titer.     Clinical applications The Paul-Bunnell test is a useful screening test for the presence of heterophil antibodies because it is simple and inexpensive. Although the specificity of the heterophil assay is rated as good, negative results are demonstrated in individuals who do not produce infectious mononucleosis heterophil antibody. If negative results are displayed, however, Epstein-Barr virus (EBV) serology may be indicated.     General question: Are all necessary fields of the CLSI format addressed?
Additional general comments:
Evaluation of write-up validation by: _____________________________ Date _________
Supervisory reviewer: _________________________________________ Date _________

image

Adapted from Paul JR, Bunnell WW: The presence of heterophil antibodies in infectious mononucleosis, Am J Med Sci 183:90–104, 1932; and Sumaya CV: Infectious mononucleosis and other EBV infections: diagnostic factors, Lab Manage 24:37–45, 1986.

Chapter Highlights

• Quality assurance indicators and quality control (QC) are tools to ensure that reported laboratory results are of the highest quality.

• The Clinical Laboratory Improvement Amendments of 1988 (CLIA ’88) established a minimum threshold for all aspects of clinical laboratory testing.

• Voluntary QC standards have been set by TJC, COLA, and CAP.

• Nonanalytical factors related to testing accuracy include the following: qualified personnel; established policies; procedure manual; test requisitioning; patient identification, specimen procurement, and labeling; preventive maintenance of equipment; appropriate testing methods; and inaccurate results.

• The Institute for Quality Laboratory Medicine has developed measures to evaluate quality in the laboratory based on the phase of testing: preanalytical, analytical, and postanalytical.

• Quality control monitors the accuracy and reproducibility of results through control specimens. Accuracy describes how close a test result is to the true value. Precision describes how close the test results are to one another when repeated analyses of the same specimen are performed. It is possible to have great precision, but without accuracy if the answer does not represent the actual value tested.

• The precision of a test, its reproducibility, may be expressed as a standard deviation (SD) or the derived coefficient of variation (CV); it is used to compare SDs of two samples. A procedure may be extremely accurate but so difficult that values are not clinically meaningful.

• Assessing sensitivity and specificity of a test involves tests positive, tests negative, disease present (positive), and disease absent (negative). Sensitivity is the proportion of subjects with a specific disease or condition who have a positive test result. Specificity is the proportion of subjects without the specific disease or condition who have a negative test result.

• Assessing the predictive value (PV) requires knowledge of the sensitivity, specificity, and disease prevalence. Prevalence is the proportion of a population who has the disease. Incidence is the number of subjects who have the disease within a defined period per 100,000 population.

• Proficiency testing (PT) is incorporated into the CLIA requirements. In addition to internal QC programs, each laboratory should participate in an external PT program to verify laboratory accuracy.

• A control specimen has a known value and is similar in composition to the patient’s blood. A control value out of the acceptable range (out of control) may result from the deterioration of reagents, faulty equipment, dirty glassware, lack of attention to timing or temperature, use of inappropriate methods, or poor technique.

• Reference range for a particular measurement is usually a normal bell-shaped curve.

• Mean is the mathematical average of the values. Median is the middle value. Mode is the most frequently occurring value. Standard deviation (SD) is the square root of the variance of the values.

• Reference range is the range of values that includes 95% of the test results for a healthy reference population, formerly referred to as normal values or normal range.

• Biometrics attempts to describe statistical variations in biological observation.

• Each laboratory must determine the reproducibility for each new procedure and establish acceptable limits of variation for control specimens.