Evaluating dietary assessment in the Canadian Longitudinal Study on Aging

An improved understanding of the role that nutrition plays in healthy aging will enable researchers to develop effective prevention strategies to improve health and quality of life. However, measuring what people eat in large studies such as the CLSA is a challenge. To reduce the burden on participants, the CLSA developed the Short Diet Questionnaire which takes only 10 to 15 minutes to complete, but is limited in detail. We collected additional information in a sub population of the CLSA to determine how well the Short Diet Questionnaire can measure usual dietary intake and to see if we can improve the accuracy of the SDQ using statistical techniques. We were also interested in determining the feasibility of using a more detailed computer based diet survey in the CLSA. This study will help us to improve our measurement of diet and increase our understanding of the relationship between the foods we eat and our health.

This study successfully collected data from 232 CLSA participants using the Automated Self-Administered 24-hour system, an online method of collecting 24-hour dietary recall data, as well as administered a second short diet questionnaire. All data analyses have been completed.

We found that the short diet questionnaire captures dietary intake with varying reliability and accuracy depending on which dietary component is examined. We observed that, generally, the short diet questionnaire is more reliable for estimating nutrients and food groups compared to the individual food items on the questionnaire. Compared to the results of the 24-hour recalls, the short diet questionnaire provided valid estimates of how participants aged 45 to 64 ranked for their consumption of vitamin D, calcium, and fiber. However, in older adults and for other select nutrients, the short diet questionnaire provided less valid rankings. Our study found that in its current state, the Automated Self-Administered 24-hour system is not feasible to implement in the CLSA. Many participants found the system difficult to use and we had to administer at least one of four questionnaires over the phone to 36% of participants. The recently updated Automated Self-Administered 24-hour system is likely more user-friendly in comparison.