Epidemiological data for smoking, blood pressure, and dyslipidemia show larger hazard ratios and consistent directionality compared to nutrition studies.

“but especially when it's always in the same direction I mean I think that's what always that's what differentiates the epidemiology around for example exercise and blood pressure from the epidemiology around nutrition the epidemiology around Nutrition a it has very low Hazard ratios and it's always changing the direction it's moving in suggesting that whatever's being studied probably doesn't matter yet when you look at you know the epidemiology of smoking blood pressure dyslipidemia exercise much bigger Hazard ratios virtually always pointing in the same direction so it strikes me that the latter is signal the former is noise”