Very hard to make glaucoma screening for pre-perimetric glaucoma cost-effective even with risk stratified screening targeting pre-known risk groups? How about later stages? Would it make more sense to target disease progression at home from health economics perspective?
You can try to build on this
TODO! Need to tweak the plot and correct the axes!
So to establish first what are the True Positive, False Positive, False Negative, True Negative rates depending on our device specificity, glaucoma prevalence, and clinician variability in glaucoma diagnosis
- Basic Bayesian stats intro: http://sphweb.bumc.bu.edu/otlt/MPH-Modules/BS/BS704_Probability/BS704_Probability6.html
- Bayesian vs. Frequentist Holy War: http://jakevdp.github.io/blog/2014/03/11/frequentism-and-bayesianism-a-practical-intro/