The observed associations of vegetable intake with cardiovascular disease risk and all-cause mortality are likely primarily due to residual confounding, such as healthy user bias, rather than the vegetables themselves being beneficial.

“When they actually looked for confounders, such as healthy user bias, financial status, socioeconomic status, almost all of the benefit of raw vegetables was erased. ... the study indicates that observed associations of vegetable intake with CVD risk and all cause more like mortality are likely to be mostly accounted for by residual confounding. Now, this is what I'm talking about, healthy user bias rather than the vegetables actually being beneficial for people.”