Agreement between dietary interviews.

The nutrient intake for 86 male myocardial infarction patients was calculated from data based on the interview methods studied both one and two years after infarction. The dietary history method resulted in systematically higher mean daily intakes than the other methods. The differences in mean intakes calculated from the recall methods were, in general, rather small. The agreement, measured by the intraclass correlation coefficient, varied from 0.42 to 0.69. The pattern of mean differences between methods was consistent one and two years after infarction.