One possibility for quantifying the uncertainty of the cost-effectiveness is focussing on the metric “effect”/mean(“cost”). It ignores the variance of “cost”, but maybe this is not problematic given we want to maxime mean(“effect”) for a fixed (constant) “cost” we are willing to spend?
Edit: mean(x)/mean(y) has some variance, but it’s not quite what we’re after. Thank you Caspar Kaiser for pointing this out.
mean(x)/mean(y) is a contant, and therefore has null variance.
One possibility for quantifying the uncertainty of the cost-effectiveness is focussing on the metric “effect”/mean(“cost”). It ignores the variance of “cost”, but maybe this is not problematic given we want to maxime mean(“effect”) for a fixed (constant) “cost” we are willing to spend?
mean(x)/mean(y) is a contant, and therefore has null variance.
See this interesting comment from CasparKaiser (and my reply to it) for what he means by it.