Background: As the proportion of old and very old people in the population increases, new research on the influence of diet on health and nutritional needs in later life will be needed. Dietary assessment methods that rely on short-term memory or lengthy interviews, such as the 24-h recall and diet history methods, could have some limitations in this age group. There is some support for the use of food frequency questionnaires (FFQs) in older people, although their validity in the older old and in those with more advanced cognitive decline has not been extensively assessed.

Subjects/methods: In a study designed to assess the validity of a modified FFQ in men and women over 65 years, 50 men and 47 women completed two FFQs, and 42 men and 41 women completed one FFQ followed by a 4-day weighed diet diary. Digit span forward (a test of short-term memory) and verbal fluency (a test of executive function) tests were used to assess the possible influence of cognitive function on repeatability and validity of the FFQ.

Results: The FFQ was found to have good repeatability for most nutrients and reasonable validity for some but not all nutrients. Cognitive function assessed prior to the dietary assessment showed no relationship with repeatability, but there was some evidence that validity was lower in those with lower executive function.

Conclusions: Dietary assessment in healthy older people without overt cognitive decline can be achieved, but development and testing of methods of data collection for each target population and nutrient of interest are particularly important in this age group to ensure valid results. The possibility that cognitive decline influences dietary assessment needs to be borne in mind in the interpretation of observational studies of the influence of dietary intake on cognition in later life.