Have you ever wondered how the same exact cup of Starbucks coffee you’ve in D.C. tastes the same as the one in Cairo? Conversely, considering everything is the same in terms of supply chain and manpower, yet the competitors failed to offer consistent taste experience. Thus, how do you thrive in such an environment. There has to be a process that ensures consistency in every cup of coffee being served – you guessed it right, precision. Bar none, when it comes to the analytical methods like Starbucks, the goal is to ensure consistency with less variability to achieve safety and efficacy in every step, with less rework and leave no room for – random errors.
That brings us to another layer to evaluate the efficiency of a method, we must examine the precision. Precision is one of the main parameters that needs to be determined during method validation the ICH Q2 (R1) defines precision as the closeness of agreement between the series of measurements obtained by the replicate measurements on the same homogenous sample under the prescribed conditions. That’s improving the precision typically reduces the degree of variation of results, i.e. improves the consistency of the results.
Precision should not be confused with accuracy. A method can be precise but not accurate and vice versa. Let’s explore this notion by looking at the following figure that illustrates the difference between precision and accuracy using the basketballer practicing 9 shots.
Figure 1: depicts the difference between accuracy and precision
1: Ball scored in the center net 9 times. This represents the optimal scenario, since all shots/results are inside the net around the mean.
2: Ball hit the outer left side of the ring 9 times. This shows a very precise method as all the shots/results are in proximity to each other. However, the method is inaccurate as the shots are outside the target.
3: Ball hit the outer side of the ring from all directions. The method requires optimization.
4: Ball landed inside the net but in different area. This represents an accurate method as all shots are within the net. However, individual shots show a high degree of variation.
Precision studies the degree of random error that could be encountered during measurements through assessing the spread of results.
It consists of four components: system precision, repeatability, intermediate precision, and reproducibility each determines a different kind of variation. Not all four components need to be determined, the extent of the procedure depends on the intended use.
We will see in the following examples when to use what.
Effect of varying conditions:
System Precision: is the first level of precision which investigates the variability of the measurement, it is referred to as the instrument precision.
Repeatability (Sr): expresses the precision of the method under the same operating conditions over a short interval of time.
Intermediate precision (SRW): expresses within laboratories variations; different days, different analysts, different equipment.Reproducibility: is the precision between laboratories.
Figure 2: Depicts the varying levels of precision
Precision have different levels of variability. Generally, the more conditions you change within the method the larger the precision value will become.
Repeatability
Repeatability provides understanding on the degree of variation between analytical results within the same homogenous sample in a short time frame. It’s occasionally referred to as intra-assay precision (Sr). The experiment is performed using the same analyst, same instrument, same set of reagents, and same day at different times. As per ICH Q2 (R1), repeatability should be assessed using:
- A minimum of 9 determinations covering the specified range for the procedure (3 concentrations/3 replicates each).
- A minimum of 6 determinations at 100% of the test concentration.
Figure 3: illustrates the repeatability concept. Same basketballer shots at different times on the same day, same ball, same court.
Intermediate PrecisionIntermediate precision studies the variation within the laboratory over a longer period of time by examining the scatter of analytical results that were obtained when a method is applied on different days, different analysts using different instruments (note, same model and same manufacturer).
Figure 4: illustrates the intermediate precision. Different basketballers shots at different times on different days, different balls, same court.
Reproducibility
Reproducibility illustrates the differences in a method by measuring the variation of analytical results obtained from different laboratories. Contrarily, to intermediate precision, reproducibility does not only involve different analysts, but also different ambient conditions, different manufacturer of instruments. The reproducibility study shows that the performance of the analytical method is location independent. Therefore, it is not a required parameter during method validation within one laboratory. However, it should be considered in case of standardization of an analytical procedure to be used in more than one laboratory.
Figure 5: illustrates the reproducibility. Different basketballers shots at different times in two different courts, and different balls.
The following example illustrates how to calculate precision within a laboratory. The experiment was carried out using SRM of NIST 3280 for determining the precision of Sodium.
Figure 6: NIST 3280 certificate of analysis
Figure 7: precision results.
Precision is a vital aspect of scientific analysis, and to ensure accuracy, results must be scrutinized through statistical analysis. The ICH Q2(R1) guidelines recommend presenting precision test results in standard deviation, relative standard deviation, and confidence intervals for each precision type. Among these, confidence intervals hold great significance in assessing precision, allowing us to gauge an instrument’s ability to deliver consistent data within a specific range of values. Essentially, it enables us to pinpoint the precise range in which an instrument can produce reliable and consistent results. Although the use of precision testing may not always be practical for repeatability, an overall confidence interval that includes both intermediate precision and repeatability results could be worth calculating.
About Author
