In this third article in his series on statistics, Stephen MacDonald moves on to look at precision studies, focusing on issues such as their practical use, terms used, study design, and interpretation. Also included is a worked example and important take-home points for the reader.
Precision studies are among the most familiar statistical exercises in laboratory medicine, but they are also among the easiest to oversimplify. The result is often reduced to a single coefficient of variation, usually reported as ‘the CV’. That is a problem, the CV is the final summary of a precision study, not the statistical method itself.
The more useful question is: where did the variation come from? A well designed precision study uses repeated measurements to separate variation into components. Some variation occurs between replicate measurements within the same analytical run. Some arise between runs, days, reagent lots, operators, analysers or sites, depending on how the study was designed. The final precision estimate only makes sense when those layers of variation are understood.
There is also a terminology point worth making at the outset. Precision describes closeness of agreement between repeated measurements. The results of a precision study are usually expressed as measures of imprecision, such as the standard deviation (SD) or coefficient of variation (CV). Lower imprecision means better precision. This article uses ‘precision study’ for the design and ‘imprecision’ for the numerical output.
Log in or register FREE to read the rest
This story is Premium Content and is only available to registered users. Please log in at the top of the page to view the full text.
If you don't already have an account, please register with us completely free of charge.