Equivalence Margin Development
Hi, it's me again, Sarah. I am a member of the bioassay group of Stegmann Systems GmbH and today’s topic will be: The EMD and especially its Visualization Tool. Equivalence Margin Development within PLA is far from new, but its true beauty has only unfolded with the release of Bioassay 25. With the beauty, however, the size and complexity of the document have also grown, so it would be too much of a good thing to treat it all at once and in a single blog article. That's why two or three interesting contributions await you now...
First things first:
What is this document type good for?
The EMD document supports our well-tried and widely used QRA document. The processes of many of our customers require a multitude of assay runs, which have to be analyzed and validated individually. As mentioned above, we have invested a lot of time and energy in the development of Bioassay 25 to improve the EMD document. It is possible to systematically record findings from these assays and develop targeted test strategies. The document can also be used descriptively to aggregate assay results and visualize the data. Thus, this document type enables our clients to learn from the past and improve the future – .
Step by Step:
What are the requirements for the use of the EMD document?
At the beginning of our small EMD journey is the successful selection of suitable test margins. No matter what purpose our further actions are intended to serve, we have two possibilities to complete this first stage. In case we do not work with manually entered limits, we need several calculated and valid QRA documents that will be used to develop our margins. And as so often in life - the more the better… or let me put it this way: the more assays (development data) available, the more applicable the margins will be. Okay, maybe that was a little too euphoric. Sooner or later we will probably come up against the limits of computing power here. I think a good benchmark is about 100 assays.
In addition, the QRAs themselves must meet the following conditions: Each of the documents consists of only two assay elements: One standard element and one test element. The calculation of the EMD document also runs with more than one test sample but would (in most cases) not produce meaningful statistical results. Besides, all referenced QRAs work with the same analytical model, the same logarithm base and the same preparation scheme settings. Only in this way can we guarantee comparability that supports the development of meaningful test limits. With a large amount of documents, this can be quite annoying, can it? Be excited … Soon we will delight you with some new developments!
How to use the EMD?
Now that our QRAs have been selected, prepared and calculated, we can finally get started. First, we create a new EMD document and assign a significant name.
Below the Settings element we can do some general settings for our equivalence margin development. The diagram above may suggest that the branches "Verification of QRAs" and "Visualization via Plots" are mutually exclusive. This is not the case! Initially Margin Development, Test Strategy Verification and Test Strategy Visualization are assigned the value “yes”, the confidence interval and the tolerance level are set at 90% and 95% respectively and all available tests are created - both for the margin development block and for the initially created test strategy. That is, even if we do nothing at all, we already have an executable document. It's up to us now to make it fit our needs.
Then we switch to the PLA Reference Editor to enter our selected QRAs into the document. (Just a small hint: At this point it's especially handy if you've already sorted your source QRAs into a shared folder, because then you can add all references at once via multiselection!) With the referencing of the QRAs, certain information from the documents is transferred into the dataset of our EMD.
This also holds two columns for the assignment of development assays and verification assays, which are both initially specified with “yes”. This is where the tolerance level comes in. If it is set to 100%, each development assay passes the verification for every test, as it is effectively used in the calculation of the margins. With the tolerance level we can play around a bit to see the margins changing and influencing which assays pass the test system successfully and which assay is disqualified by the strategies. On vice-versa, with a tolerance level of 80%, for every test the highest and the lowest 20% of the values will not be considered when the margins are determined. You're wondering if that's true? It was the same with me! However, there is a simple explanation: each test margin is developed individually. Thus the tolerance level always refers to one side only. This becomes logical when you remember that not every test necessarily has both an upper and a lower limit.
The next step is to devise appropriate testing strategies. In the margin development part, at least all tests that are later used in the test strategy should be covered.
And that brings us to the point where I leave you alone! Play around, try it out! And if you've done your homework well, I'll be happy to see you again next time for some details about our visualization tool.