iRoeMetz provides options to do fully-crossed or split-plot study and paired or unpaired data study.
Fully-crossed study
"Fully crossed" refers to the data collection: all the readers rate all the cases in both modalities with respect to confidence of disease.
Split-plot study
In split-plot study, we separate reader, non-disease cases and disease cases into several groups. Each group of readers only read specific groups of non-disease and disease case.
Paired data study
In paired data study, readers and cases are same across two modalities.
Unpaired data study
In unpaired study, readers or cases are different across two modalities.
Example Study
Figure 1 is an example of paired split-plot group study design. We design a split plot study with 4 groups of readers, non-disease cases and disease cases. Each group of readers only read a quarter of whole non-disease and disease cases. And readers and cases in both modalities are same.
Figure 1. Paired split plot study with 4 groups of readers, non-disease cases and disease cases
Figure 2 is an unpaired reader paired case study design with 4 split plot groups. Two modalities have different readers, but same diseased and non-diseased case.
Figure 2. Unpaired readers, paired cases study with 4 split plot groups
