Has their been a large scale alcohol effect study that exploits the fact that some religions (particularly Islam, Judaism) don't tend to consume alcohol?
It will probably need to be delicately applied since religion correlates with many other socioeconomic/genetic conditions, but I think showing that certain cancers are not as common in for example Indonesia and Turkey as compared to Australia and Greece would be an interesting hypothesis concerning the impact of alcohol.