search

mrc-ide / DRpower

Sample size calculation and power analysis for drug- and diagnostic resistance studies
MIT License
1 stars 0 forks source link

Counter-intuitive result for power threshold with a single site #38

Closed alimanfoo closed 3 months ago

alimanfoo commented 3 months ago

Thanks so much for this amazing package!

I'm using DRpower to investigate power to detect new resistance variants during the early stages of emergence, when variants may not yet be widely spread and hence only detectable at a single site or small number of sites.

For interest I was looking at power to detect a variant with prevalence 10% above a threshold of 5% at a single site, given different sample sizes, and got a counter-intuitive result where increasing sample size seemed to decrease power:

image

Is this expected or am I doing something wrong or inappropriate?

alimanfoo commented 3 months ago

Sorry somehow my browser opened the same issue twice! See #37.