Precision is a measure of agreement between independent test results obtained with identical samples under certain experimental conditions. It is typically expressed numerically in terms of the standard deviation of the independent test results. Generally, there are two types of experimental conditions. The first type is where independent test results are obtained with identical samples by the same analyst using the same method and the same device in one laboratory over a relatively short, typically less than one day, period of time. The standard deviation of the independent test results obtained under the first type of conditions is known as the repeatability standard deviation.
The second type is where independent test results are obtained with identical samples by the same method using multiple devices in multiple laboratories and multiple operators over a long, typically many days, period of time. The standard deviation of the independent test results obtained under the second type of conditions is known as the reproducibility standard deviation. Both repeatability and reproducibility are estimated at multiple levels of analytes including levels that are of special clinical importance.
The relationship between the precision of a medical device and the amount of analyte is very important clinically and statistically and must be characterized. Repeatability and reproducibility can be estimated using appropriate experimental designs and calculations. An example is a fully nested and balanced random effect model analysis of variance.