LAUSR

BACKGROUND AND OBJECTIVES This study aimed to determine the accuracy of a diet quality measurement tool, the Total Diet Score (TDS) using two validation methods; firstly the TDS calculated from a food frequency questionnaire (FFQ) was compared to the TDS calculated from weighed food records (WFRs); secondly the TDS was compared to a number of dietary biomarkers. METHODS AND STUDY DESIGN Data were collected from a population based cohort study located in the Blue Mountains region of Sydney, Australia. To compare dietary assessment tools, a sub sample of 75 subjects (aged 63 to 83 years) completed the FFQ and three, four-day WFRs at baseline. Fasting blood samples were collected from 2897 subjects at the first follow up in 1997-1999. TDS scores were calculated from both WFRs and FFQs. Methods to compare FFQ TDS scores to WFR TDS scores included paired t-tests, Pearson correlations, Bland-Altman plots, joint classification quartiles and weighted kappa scores. Linear regression analyses were used to assess the relationship between TDS and biomarkers. RESULTS No significant mean difference was found between FFQ TDS and WFRs TDS (p=0.63) with a significant positive correlation seen between the two methods (r=0.75, p<0001). The Bland-Altman method found no linear trend between the differences and means of TDS scores between the FFQ and WFR (p=0.38). A significant trend for higher serum vitamin B-12, serum folate, homocysteine and lower total cholesterol was found with increasing TDS. CONCLUSIONS These findings suggest that the TDS is a useful tool for assessing diet quality in an older population.