STATISTICS


sufficiently consistent for their intended clinical use. In day-to-day work, imprecision affects whether small differences between serial patient results are credible, whether results close to a decision limit can be interpreted confidently, and whether apparent instability reflects patient change or analytical scater. These are practical questions, but they depend on statistical design. The key point is that a precision


estimate is conditional. A single run of replicate measurements estimates short-term repeatability. A multi-day study estimates a broader form of within- laboratory imprecision. A multisite study estimates something broader again. If the study did not include a source of variation, that source cannot be represented in the final estimate.


Practical use cases in clinical laboratories Precision studies should be planned around the laboratory question. In many routine settings, the question is whether or not a commercial measurement procedure performs locally as expected. This may arise after installation of a new analyser, introduction of a new reagent system, or transfer of a method into a different laboratory environment. Here, the laboratory is usually verifying an established claim rather than establishing precision from first principles. A different question arises when the


laboratory has developed, modified or substantially adapted a method. In that seting, a short verification study may be insufficient. The study may need to include


Term Repeatability Within-run precision What it describes


Variation under closely controlled conditions


Variation within one analytical run


Within-laboratory precision Variation within one laboratory over time


Reproducibility Variation across laboratories or sites Table 1. Precision terms and what they mean in practice.


several days, runs, operators, reagent lots, instruments or sites, depending on which sources of variation are relevant to routine use. Precision studies are also used for


troubleshooting. An assay may show acceptable agreement between replicate measurements in a single run, but greater variation between days or reagent lots. A structured study can help identify whether the main source of variation sits at the level of replicate measurement, run, day, operator, instrument or material. A further use is to assess


concentration-dependent imprecision. A normal level control may give a reassuring CV, but it may not describe performance at low concentration, in an abnormal range, or close to a clinical decision point.


Precision terms Repeatability is the narrowest form of precision. It describes variation under closely controlled conditions, such as repeated measurement of the same


material within the same analytical run. It is useful, but it is a best case estimate. Within-laboratory precision is


broader. It describes variation when measurements are repeated in the same laboratory over a longer period, allowing routine sources of variation such as different days, runs, calibrations and operating conditions to contribute. This is usually more relevant to routine service than repeatability alone. Reproducibility is broader again


and usually refers to variation across different laboratories or sites. It is important for transferability of a measurement procedure, but it is not usually the main objective of a local precision verification study (Table 1, Fig 1).


Operator / reagent lot / analyser / site (broader routine sources of variation)


From replicate results to variance components The statistical centre of a precision study is not the CV. It is the separation of total variation into components. In a simple balanced design, repeated results are arranged into groups. A group might be a day, run, series, operator, reagent lot, analyser or site, depending on the question. The analysis then separates variation within groups from variation between groups. Within-group variation describes the


Day (between-day variation)


Analytical run (within-day variation)


Replicate measurement (repeatability variation)


Precision estimates become broader as more sources of variation are included in the study design.


Fig 1. Layers of variation in a precision study. A precision estimate reflects only the sources of variation included in the study design.


scater of replicate measurements under the most similar conditions included in the design. This is the repeatability component. Between-group variation describes additional variation in the group means. If the groups are days, this reflects day-to-day effects. If the groups are runs, it reflects run-to-run effects. If the groups are reagent lots, it reflects lot- to-lot effects. This is where analysis of variance


(ANOVA – see article 2 in series) is useful. The ANOVA table partitions variation into a within-group mean square and a between-group mean square. The within-group mean square estimates the repeatability variance. The between-group


August 2026 WWW.PATHOLOGYINPRACTICE.COM 17


Practical interpretation Best case short-term


imprecision


Short-term variation during a batch or run


More realistic estimate of routine local imprecision


Broader transferability of method performance


Broader study design (more sources intentionally included)


More sources of variation included


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56