Quality Assurance and Quality Control in the Clinical Laboratory



Quality Assurance and Quality Control in the Clinical Laboratory



Objectives


At the end of this chapter, the reader should be able to do the following:


1. Define the following terms: quality control material, outlier, shift, trend, random error, systematic error.


2. Plot quality control results on a Levey-Jennings chart.


3. Evaluate quality control results for shifts and trends.


4. Evaluate quality control results and determine, using Westgard rules, whether the results are acceptable.



BASIC QUALITY ASSURANCE CONCEPTS


In the laboratory, the instruments and methodologies must be monitored to ensure accurate results. “Quality assurance” is the process in which this occurs. Quality assurance is the monitoring of any activity that is associated with a laboratory result. Activities that occur before the sample reaches the laboratory are called preanalytical activities, those that occur in the laboratory that directly deal with the analysis of the sample are called analytical activities; and those activities after the analysis is performed are called postanalytical activities. The laboratory must monitor each of these activity phases for errors to ensure accurate results. For example, if the wrong patient’s sample is drawn, that is a preanalytical error. If the sample is analyzed incorrectly, that is an analytical error. Similarly, if a diluted sample’s result was not multiplied by the dilution factor correctly, that is a postanalytical error. One technique that laboratories can use to check the quality assurance procedures is to follow random samples as they proceed from collection to analysis to charting of the results.



Quality Control Material


Within the laboratory, one method that is used to ensure the quality of the patient results is to use quality control material. Quality control material is material that is analyzed along with patient specimens and should be treated the same as patient specimens. Quality control material should be of the same matrix as the patient sample. The term matrix refers to the chemical and physical characteristics of the material that contains the analytes to be measured. If measurement of analytes in serum is being performed, a serum-based quality control material should be used. Likewise, if measuring analytes in urine specimens, a urine-based quality control material should be used. Quality control material is usually manufactured to contain many different analytes so that it can be used on multichannel analyzers. The concentration of the analytes will vary depending on the “level” of the quality control. When two or more levels of quality control material are included in an assay, one level may have the concentration of the analyte found in the normal population. The second level may have either the elevated or low concentration of the analyte. Whether the second control has an elevated or decreased concentration of the analyte will depend on the medical usefulness of the particular concentration. Some analytes require that three levels of control be analyzed because both the decreased level of the analyte and the elevated level of the analyte are medically useful. For example, when analyzing therapeutic drugs such as theophylline, three levels of control (low, medium, and high or levels 1, 2, and 3) are used. Under the final rules of CLIA’88 published in the Federal Register on January 23, 2003, at least two levels of quality control material must be included for every assay at a minimum of at least once a day unless specialty requirements apply or the method has met the criteria for equivalent quality control. The CLIA’88 final rules in regard to quality control divided the laboratory tests into waived and nonwaived. For waived tests, the quality control rules are very simple: laboratories should follow the manufacturer’s instructions for performing quality control. Equivalent quality control will be discussed later in this chapter. The remainder of this chapter will focus on the quality control that must be performed for nonwaived tests.


The mean and standard deviation (SD) for each analyte is established for each level of quality control material. In this manner, the entire method can be monitored for errors because whatever occurs to the patient samples also occurs to the quality control samples. For example, if the wrong pipette is used in an analysis, and twice as much serum is used in the analysis than was called for in the procedure, the results for the quality control samples will be twice as high as they should be. Without quality control material to check the accuracy of the method, errors can occur. If the quality control results fall outside of the laboratory’s established range of acceptable results, the patient results should not be reported. Instead, the laboratory’s troubleshooting policy for unacceptable quality control results should be followed.


Quality control material can be in one of three forms. Lyophilized, or “freeze-dried,” quality control material must be reconstituted with diluent before use. It is crucial to use the correct diluent and the correct quantity of diluent to ensure accurate quality control material results. Lyophilized controls must be allowed time to be fully reconstituted before use. Often this means that after the diluent is added and the control material swirled gently to mix the diluent and lyophilized material, the control is allowed to be undisturbed for a period of approximately 5 to 20 minutes to allow for all of the lyophilized material to go into a solution. If this time is not allowed and the quality control material is immediately used after the diluent is added, inaccurate results may occur.


The second form of quality control material comes from the manufacturer prediluted and ready for use. No reconstitution is necessary for this type of control. The third type of material is ethylene glycol-based controls. These are prediluted in ethylene glycol. Ethylene glycol-based controls should not be used on analytes measured by ion-specific electrodes because the ethylene glycol may damage the electrodes. Ethylene glycol controls are stored in liquid form at 0° C.



Quality Control Analysis


After quality control materials are used, their results must be analyzed before patient results are reported to the physician. There are many different methods of analysis of quality control results; however, it is beyond the scope of this book to list all of them. The basic statistical concepts discussed in Chapter 13 form the basis for quality control analysis. Recall that 68.2% of the time a control result will fall within +/−1 SD of the mean, 95.5% of the time the results will fall within +/−2 SD of the mean, and 99.7% of the time the results will fall within +/−3 SD of the mean. On the other hand, approximately 32% of the time the results will fall outside of the +/−1 SD interval, almost 5% of the time they will fall outside of the +/−2 SD interval, and 0.3% of the time the result will fall outside of the +/−3 SD interval. These statistical probabilities are used when establishing the acceptable limits of our quality control results.


The question that the laboratorian has to ask and answer is, “What is an acceptable limit for my quality control material result?” Remember, if the quality control result is outside of the acceptable limit established by the laboratory, the patient results should not be reported until the problem is solved. If a laboratory uses the 2 SD range, 5 out of 100—or 1 out of 20—quality control results will fall outside of the 2 SD range. This means that if a result falls outside of 2 SD, there is a 95% chance that the result is invalid and only a 5% chance that the result is valid and simply fell outside of the range by chance. Figure 14–1 shows a frequency distribution of 30 quality control results for glucose. The frequency of each range of SD is noted. If a result falls outside the laboratory’s established quality control result range, it is called an outlier. The method is termed out of control, and action must be taken to determine the problem. No patient results can be reported until the method is “in control.”


image
FIGURE 14–1 Probabilities associated with standard deviations.

Whether or not outlier control values should be included in the laboratory’s quality control statistics depends on each individual laboratory’s quality assurance protocol. In general, if the outlier value can be directly traced back to a problem with the control itself—for example, an outdated control—the outlier value is not included in the statistics.


In the hematology laboratory, the quality control technique of moving averages may be used on automated analyzers to establish the control limits for the erythrocyte indices. Erythrocyte indices tend to be stable within a given population. In the technique of moving averages, 20 consecutive patient samples are batched, and the mean is calculated by the instrument. An overall mean is established for every 20 batches of patient samples (400 different patients). This technique uses a complex mathematical formula to “smooth” the individual results within the batches, thereby allowing means and standard deviations to be calculated for each index.



Errors That Cause a Method to Be Out of Control


There are two main reasons why a method is out of control. The first is due to random error. Random error is error that occurs solely by chance. In the 2 SD range, random error is why 5 out of 100 (or 1 out of 20) times the quality control results will be outside of the range, although there is really nothing wrong with the method. Laboratorians strive to keep the occurrence of random error as low as possible. An example of random error may be a 100-μL pipette that delivers 100 μL of sample in most samples to be analyzed but delivers only 90 μL of sample in a few samples to be analyzed. Those few samples will have inaccurate results because insufficient sample was pipetted. Routine maintenance and calibration can reduce the chance of this random error occurring. Random error is related to the precision of the method. Other examples of random error include a bubble in the sample or reagent that results in inaccurate pipetting by the instrument, an electrical surge or transient power reduction, improperly mixed reagents, pipette tips that do not fit properly, or a clot in the sample that results in a “short sample” being pipetted.


The second type of error is systematic error; that is, all samples are affected, not just a few. Systematic error may produce a “bias” in the method. One reason for systematic error may be a refrigerator that does not keep the reagents at the proper temperature. Reagents that contain enzymes will be affected as the enzymes may lose their strength if not stored properly. The results obtained for all of the samples and quality control material used in a method that uses reagents not stored properly may be falsely lowered and inaccurate. Another example of a systematic error is a method that is not calibrated properly. All results will be adversely affected either in a positive or negative direction. The chance of systematic error can be reduced by proper calibration and routine maintenance of all laboratory equipment and instruments. Systematic error is related to the accuracy of the method. Other causes of systematic errors icclude a change in the lot number of the reagents or calibrators, wrong calibrator or quality control values being used, incorrectly prepared reagents and controls, incorrect storage of reagents and controls, change in temperature of reaction blocks and incubators in the instruments, deterioration of the light source, and a change in procedure from one technologist to another.



Levey-Jennings Charts


CLIA’88 and good laboratory practice require the use of at least two quality control materials per day for each nonwaived method to ensure accurate and reliable patient results (assuming equivalent quality control (EQC) is not how the laboratory performs their quality control). The results of the quality control material must be analyzed to determine if the method is “in control” before patient results are reported. One mechanism for quickly analyzing each control value is to plot the value on an individual Levey-Jennings control chart. Figure 14–2 is a Levey-Jennings chart of level 1 glucose control and consists of a graph in which the mean and SD ranges are plotted on the y axis, and the days of the month are plotted on the x axis. Each level of quality control material for a particular analyte has its own Levey-Jennings chart. For example, although CLIA’88 requires the use of two levels of quality control material for automated hematology analyzers each day, many laboratories use three levels instead and spread them out over the three shifts. This also satisfies the CLIA’88 requirement that the QC is rotated among all staff who perform the tests. If the analyte was hemoglobin, the hemoglobin results obtained from the three different levels of controls would be plotted on three different Levey-Jennings charts, one for each level of control.


image
FIGURE 14–2 Levey-Jennings chart for level 1 glucose control.




Example 14–1

Three levels of control are run daily on an automated hematology analyzer in a physician’s office laboratory. The mean for the “normal” level is 15 mg/dL with a SD of 1.5 mg/dL. What will the Levey-Jennings chart look like if a technician plotted hemoglobin results obtained from the “normal” level of quality control material over a 5-day period? The quality control hemoglobin results that were obtained are as follows:


Day1=13mg/dL,Day2=12.5mg/dL,Day3=16.0mg/dL,Day4=15.5mg/dL,Day5=14mg/dL.


image

The Levey-Jennings chart consists of the days of the week or times of the run on the x axis and the mean and SD intervals for the particular level of quality control material to be charted on the y axis. The 1, 2, and 3 SD intervals must be calculated first. The mean for the level to be plotted is 15 mg/dL, and the SD is 1.5 mg/dL. Therefore the +/−1 SD interval will be from 13.5 to 16.5. The +/−2 SD interval will be from 12 mg/dL to 18 mg/dL, whereas the +/−3 SD interval will be from 10.5 to 19.5 mg/dL. These interval values are plotted on the y axis of the chart. The days of the run are plotted on the x axis. Figure 14–3 illustrates the Levey-Jennings chart up to this point.


image
FIGURE 14–3

Next, the five values obtained are plotted on the chart by placing a dot or circle at the intersection of where the value is found on the y axis and the day analyzed on the x axis, as demonstrated by Figure 14–4.


image
FIGURE 14–4

Last, each result obtained from the same lot number is connected to the next by a line as illustrated in Figure 14–5. When a new lot number is used, the mean and SD may be different and should be recalculated before use. Then either a new Levey-Jennings chart is used or the same Levey-Jennings chart is used, with a new labeled y axis reflecting the mean and SD of the new lot. The results from the two different lot numbers are not connected by a line. A notation should be placed where the results from the new lot number begin to be charted.


image
FIGURE 14–5 Levey-Jennings chart for normal level hemoglobin control.


Shifts and Trends


By plotting quality control results on a Levey-Jennings chart, shifts and trends in the quality control results can be quickly discovered. A shift occurs when the quality control results are all distributed on one side of the mean or the other for 5 to 7 consecutive days. Shifts occur because of systematic error. A new lot of reagent might have inadvertently been used, or a method that is not calibrated can cause a shift to occur. Figure 14–6 demonstrates a shift on a Levey-Jennings chart of the low level of hemoglobin quality control material. When a shift occurs, the cause must be found and corrected because the method is “out of control.”


image
FIGURE 14–6 Levey-Jennings chart demonstrating a shift.




Example 14–2

The following quality control results obtained on days 6 through 11 for a “normal” level of control material for aspartate transaminase (AST) must be plotted on the following Levey-Jennings chart (Figure 14–7). Days 1 through 5 have already been plotted.


image
FIGURE 14–7

Day6=60U/L,Day7=65U/L,Day8=62U/L,Day9=61U/L,Day10=61U/L,Day11=62U/L.


image

Is there anything wrong with these quality control results? If there is, what are some of the predominant causes?


The quality control results demonstrate a shift. All the results fall on the same side of the mean, which is statistically unlikely, and there is an abrupt change in the pattern of the quality control results. The most likely cause of the shift may be a new lot of reagent used without the assay being recalibrated, or another type of systematic error has occurred.


A trend occurs when the quality control results either decrease or increase consistently over a period of 5 to 7 days. A trend is also due to systematic error, but the type of error tends to occur more slowly. For example, reagents stored in a refrigerator that is unable to keep the correct temperature may slowly deteriorate, or the light source in the instrument is slowly deteriorating. Figure 14–8 demonstrates a trend occurring in level II of a quality control material for automated white blood cell counts. As with shifts, when trends occur, the cause must also be found and corrected.


image
FIGURE 14–8 Levey-Jennings chart demonstrating a trend.

Tags:
Nov 18, 2017 | Posted by in PHARMACY | Comments Off on Quality Assurance and Quality Control in the Clinical Laboratory

Full access? Get Clinical Tree
Get Clinical Tree app for offline access