AOAC Guidance on FA Immunoassay Validation (August 2023)

1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429

plate to estimate well-to well ELISA variance, since the 2 wells that are used for extract #3 cannot be reused for extract #4, again the factor ELISA is nested within the TP factor. Statisticians will say that for a factor to be nested, there needs to be a significant “separation” in that factor across the different levels of the factor one level higher in the hierarchy. Separation is achieved because the test portion is destroyed and can’t be recovered. If a factor is not nested then we say (some authors use this terminology) that the factors are “crossed”, meaning they need to be treated as a factorial design, such as a “2x2” factorial. It does not mean to imply there are interactions fitted in the model. To avoid this confusion, some authors refer to these 2 factors as “Main Effects.” The area where this will be difficult in these validation designs is the level that includes Analyst/Day/Calibration. For each method and experimental design, we will need to determine if the Analyst/Day factor can be considered nested within the Lot factor, or if there is inadequate separation between Analyst/Days for one lot to another and so will have to be considered as 2 main effects. To make this easy, Lot will always be a Main Effect at the top of the hierarchy, and TP and ELISA (if replicated) will always be nested. The other easy thing is that the ANOVA calculations in R are simple, and R can do the

analysis either way, with a minor change to the code.

1.3.2. Proposed Decision Rules for Determining Nested Variables

Design

# Analysts

# Days

# Calibrations

Adequate Separation?

Factor Is:

1a or 2a 1a or 2a 1a or2a 1a or 2a 1a or 2a 1b or 2b 1b or 2b 1b or 2b 1b or 2b

2 2 2 4 4 2 2 2 4

2 2 2 4 4 2 2 2 4

1 2 4 2 4 1 2 6 6

No No Yes Yes Yes No No Yes Yes

Not Nested Not Nested

Nested Nested Nested

Not Nested Not Nested

Nested Nested

1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443

1.3.2.1.

The idea here is if you just have 2 analysts and 2 days, you can only have enough separation for nested if you have a different calibration for each day/lot combination. If you have 4 trained analysts in the lab and you can spare them, then you can get separation that way. This is assuming calibration is the significant source of variation, which is usually the case in ELISA methods. In fact, Day is usually always confounded with calibration for a traditional ELISA. The case where there is a common calibration might be if there is a pre-calibrated kit and the calibration is associated with the lot at the factory. If you can’t get separation, it is not a problem. You just need to differentiate before the analysis happens so you get the correct

ANOVA estimates.

1.4. Model Statements in R

1.4.1. For the nested ANOVA analysis, we will be using R package VCA, which was developed by CLSI for doing method validation on clinical analyses. Information can be found at