Perhaps the blood tests will/already do include this analysis. Asking people about their behaviours/habits/diets is usually a very poor data source. Not only do people have poor memories, but they are motivated to recall only positive behaviours, etc.

Many studies on health interventions (diet, exercise, sleep) have had to handle these problems and usually have to resort to much more controlled studies to demonstrate the effect beyond the "very large numbers, very long time" post-hoc analyses.