Re: "Interpretation and choice of effect measures in epidemiologic analyses".