These studies are notorious for being worthless. Very few people can remember what they ate the previous day, much less the prior year. In addition even if the data is recorded correctly, it can barely show a correlation, much less a connection between a diet and a given outcome.<p>"Diet data were collected using food frequency questionnaires or by taking a diet history. Each participant was asked a long list of what they’d eaten for the previous year or month. The data were collected during a single visit.The study had up to 31 years of follow up (median: 17.5 years), during which 5,400 cardiovascular events and 6,132 all-cause deaths were diagnosed."