Hello fellow researchers. This is my first time asking question on ResearchGate, and I would be grateful if someone could provide me with suggestions and help. There are two questions that I want to ask:

1. Two AST methods with large range of MICs

Currently, I am working on an alternative AST method and want to do comparison with another method by my collaborator. I notice that Bland-Altman plot is a way to analyze agreement, but there is a problem when I use Bland-Altman to analyze my results: the collaborator provides bacteria with MICs ranging from 0.5 to >256 µg/mL (mostly 0.5, but some are over 200), and most of the points cluster nearby zero while several points scatter, as the differences between means of MICs in each bacterium are very large.

There are several options I am considering, but I am unsure which one is the best.

(1) Analyze locally, e.g. One plot for bacteria ranging from 0 to 2 µg/mL; another for bacteria over 64 µg/mL.

(2) Analyze using log(mean), turning the >256 µg/mL into log(256) = 2.41.

(3) Do not use Bland-Altman plot at all. Values like Essential Agreement (EA) are already sufficient.

2. Overlapping of MIC differeces

Also, as my method will give a distinct MIC, I find out that the points on graph sometimes overlap. For example, if I have 5 samples for comparison between two methods, and the results are 1.5 vs 2.0, 1.5 vs 2.0, 1.5 vs 2.0, 0.4 vs 0.8 and 0.5 vs 0.7, then there are 3 identical differences in means. The resultant plot will only have 3 points. Will it be problematic to present my data with only 3 points, but indeed there are 5 comparing values?

Thank you for reading my questions and helping!