I Don’t Regret CI And Test Of Hypothesis For Attributable Risk. But Here’s What I’d Do Differently.