Analytical Errors in Clinical Laboratory,

Analytical Errors in 

Clinical Laboratory

1. Random error 

2. Systematic error

Method Selection Using 

Statistical Equations

Analytical Errors in Clinical Laboratory

An error, in simple terms, is the difference between the result obtained and the result expected (true value). Laboratory errors which occur during analysis can be minimised if attention is paid to proper laboratory procedures and techniques. Such analytical errors are of two types, namely, random error and systematic error. Random error may arise from factors such as technique and temperature fluctuations. Systematic errors may be caused by poorly made standards or reagents, poorly written procedures or improper instrumentation. These may directly affect the mean.

1. Random error 

Random error reflects absence of precision. When aliquots of a specimen with a constant concentration of a substance are repetitively analysed for its measurement, the absence of precision (imprecision) is indicated by variation in the value obtained by each analysis, i.e., imprecision is the dispersion of repeated measurements about the mean. Precision represents reproducibility of a result and can be determined by calculating the standard deviation between multiple measurements. Random error, present in all measurements, is due to chance, and can be either positive or negative. When measurements are plotted on a graph, random error is indicated by a scatter of points around a line drawn to show the expected values (Fig. 4.4).

2. Systematic error 

Systematic error reflects accuracy of a method. Inaccuracy in the analysis is the difference between a measured value and the true value of the analyte the substance being analysed. A result is said to be accurate when it matches the expected result. Accuracy is determined by the mean.

Systematic error can be constant or proportional. In a graph, a constant error causes the line to shift in one direction or the other while the slope remains unaffected and the line does not pass through the origin. Results showing a proportional error show a graph which has a a slope different than that of the expected line, and may or may not pass through the origin of the graph (Fig. 4.5).

Sensitivity for a qualitative test, sensitivity of a test is the test 's ability to give a true positive result when a certain disease is present. In other words, it can detect even a very small amount of the substance which is diagnostic. The analytical sensitivity or detection limit refers to the smallest concentration that can be measured accurately. Sensitivity can be expressed as:

TP Sensitivity (%) =

where TP is true positive and FN is false negative. 

Specificity Qualitative specificity is the ability of a test to give a true negative result in the patient who does not suffer from that specific disease. Analytical specificity is the ability of a test to de

Method Selection Using Statistical Equations

Both the random error and the systematic error must be minimal before a test methodology is accepted in a laboratory. Both the errors cause the result to differ from its true value, even though they are independant of each other, and are caused by different reasons. There are statistical equations that can be used to evaluate these errors. They are: 

1. I test-The i test is used to determine whether there is a statistically significant difference between the means of two groups of data. 

2. F test-The F test is used to determine whether there is a statistically significant difference between the standard deviations of two groups of data.

For both the F and the t test, a statistic is calculated and then compared with critical values found in the t and Ftables of statistics books. The critical values are used to derive the significance level or the probability that the differences in the means or standard deviations are due to chance.

By convention, if the probability of a difference occuring due to chance is less than 5 per cent, then the difference evaluated by the t or F test is statistically significant. On the other hand, if the probability of the difference occuring due to chance exceeds 5 per cent, then the difference is said to be statistically insignificant

1. t test In its ismplest application, the i test is used to detrmine whether the mean of one method is different from the mean of another. The equation for calculation of r is:

where SD, is the standard error for that difference. The standard error is calculated as follows:

Where n₁ is the number of observations in the first group: n₂ is the number of observations in the second group: SD₁ is the standard deviation of the first group: and SD₂ is the standard deviation of the second group.

In a clinical laboratory, the t test is usually applied to method-comparison data when patient specimens are measured by both the test and reference methods. The mean and standard deviation of measured values in each group is determined; and tested for statistically significant differences using statistical tables. 

2. F test The second statistical test, the F test, is used to compare the sizes of the standard devia. tions of two methods. To calculate the F statistic, the square of the larger standard deviation (SD) is divided by the square of the smaller standard deviation (SDs)

Like the test, the F value is then compared with critical F values in the statistical table. Because the F test provides information about statistical significance but not clinical significance, some workers are of the view that it should not be used as an indicator of acceptibility of a test. The acceptibility should depend more on the size of the random error.