I have one variable of continuous data and would like to find a optimal cut-off point for diagnosis a disease. sample dataset below:
ID marker
1 23
2 40
3 55
4 25
5 32
6 56
7 70
8 19
9 38
10 36
I am trying to apply the minimum p-value approach (that is introduced here https://pubmed.ncbi.nlm.nih.gov/8182763/ ) to select this optimal cut-off point (without the use of outcome data).
The method is described as splitting the data into two and calculating the p-value, then repeating this for different splitting points until the lowest p-value is found (this will be corrected for increase in type 1 error later on). The challenge I have is I can't seem to find any package or appropriate code for R that can be used to generate these multiple p-values and would appreciate any suggestions on this.