Evaluation of methods used to assess dietary intake: simulation analyses

The choice of dietary methodology can affect the ability to detect and describe the relationship between dietary sodium intake and blood pressure. This is illustrated in this paper through the use of simulation modelling of the effect of using different dietary methods (food recalls or records covering different numbers of days, food frequency questionnaire estimates of a single diet component) and using urinary excretion as a proxy for intake. Both epidemiologic studies and experimental interventions are simulated. Although the data base used was simulated rather than real, an attempt was made to keep it realistic in relation to what might be seen in actual populations. From these analyses it can be inferred that with appropriate choice of methodology and study design, even low order relationships between sodium intake and blood pressure should be detectable. At a more general level, it may be concluded that while there is no perfect dietary methodology, there are preferred methodologies for defined purposes.