Detailed explanation of precision

Precision performance is one of the basic analytical performance of the detection system. It is also the basis for other methodological evaluations. If the precision is poor, other performance evaluation experiments cannot be performed. Precision is widely used as a measure of the good or bad of a test method. Many statistical methods have been established to determine precision, some of which have been established as standards due to their widespread adoption.

Precision refers to the consistency between independent test results under specified conditions. Characterizes the magnitude of random errors in the assay. The size of the precision, the range, the mean deviation, the relative mean deviation, the standard deviation, and the relative standard deviation can all be used to indicate the precision, and the standard deviation is more commonly used. Explain that the higher the precision, the more accurately measure the precision statistically, generally expressed by the standard (bias) difference. Precision refers to the repeatability and reproducibility of the test. High precision is a prerequisite for ensuring good accuracy. Generally speaking, precision is not high and it is impossible to have good accuracy. On the contrary, the precision is high and the accuracy is not necessarily good. This situation indicates that the random error in the measurement is small, but the systematic error is large.

First, related terms and definitions

When discussing precision, you often encounter the following terms:

1, parallelism

Parallelism refers to the degree of conformity between two or more replicates of the same sample using the same analytical method in the same laboratory when the analyst, analytical equipment, and analysis time are the same.

2, repeatability

Reproducibility refers to the agreement between two or more independent determinations of the same sample using the same analytical method when at least one of the three factors of the analyst, analytical equipment, and analysis time are different in the same laboratory. degree.

3. Reproducibility

In different laboratories, different operators use different equipment, and the same test method is used to test the consistency of test results independently of the same sample (or component). The difference in reproducibility and repeatability of the test results is obvious. Although they all refer to the consistency between the test results of the same sample, the premise is different. Repeatability is the consistency between successive test results in the case where the test conditions remain the same; reproducibility refers to the consistency between the test results in the case where the test conditions are changed. Reproducibility, also known as reproducibility, reproducibility, is expressed by the standard deviation of reproducibility. The standard deviation of reproducibility is sometimes referred to as the standard deviation between groups. When reproducibility is given, the conditions of the test conditions should be specified in detail, including: test personnel, test equipment, location, conditions of use, and time. These contents can change one, many or all of them.

4, measuring precision during the period

The precision of measurement during period, referred to as period precision, refers to the precision under precision conditions during a period of time, similar to the “inter-assay precision” often referred to in daily work. The period precision condition refers to the same measurement procedure and the same location, and repeatedly measures a set of measurement conditions of the same object to be measured over a long period of time. The instrument can be recalibrated or replaced during this time.

Second, the precision calculation method

First, the experimental design is carried out. The suitable concentration of the sample to be used in the experiment is generally related to the concentration level or medical decision level declared by the manufacturer. Usually, the specimens of low, medium and high levels are selected. Evaluate the precision within the batch or within the day, usually within 20 to 30 times within one batch or within one day; evaluate the precision between batches or during the day, generally 20 to 30 batches of measurement or 20 to 30 days Measurement (measured once a day), directly calculate s and CV to obtain precision evaluation data. Intra-assay precision and inter-assay precision can be calculated simultaneously if the measurement is repeated several times in a day, within a day (or within a batch). Some scholars refer to the sum of intra-batch and inter-batch precision as the total precision of the laboratory, or simply indoor precision. The calculation formulas for s and CV are as follows: