439Reproducibility and relative validity of dietary patterns in older New Zealand adults

Abstract

Abstract Background Valid and reproducible dietary patterns (DP) are necessary to assess relationships between diet and disease. This study, Researching Eating, Activity, and Cognitive Health (REACH), assessed the reproducibility and relative validity of DPs derived from a food frequency questionnaire (FFQ) in older New Zealand adults. Methods Participants (n = 294, 70±3yrs, 37% male), completed two identical 109-item FFQs to assess reproducibility. A 4-day food record (4DFR) was used to assess relative validity. After collapsing data into 57 food groups, DPs were derived using principal component analysis. Agreement of DP loadings were assessed using Tucker’s congruence coefficient. Agreement between derived DP scores from the FFQs and 4-DFR were assessed using Spearman correlation coefficients, weighted kappa statistic, and Bland-Altman analysis. Results Three similar DPs (‘Mediterranean style’, ‘Western’ and ‘prudent’) were identified from each dietary assessment tool. Congruence coefficients between DP loadings ranged from 0.66-0.80 (reproducibility) and 0.54-0.75 (validity). Correlations of DP scores ranged from 0.47-0.59 (reproducibility) and 0.33-0.43 (validity) (all P<0.001); weighted kappa scores ranged from 0.40-0.48 (reproducibility) and 0.27-0.37 (validity); limits of agreement from ±1.79 to ± 2.09 (reproducibility) and ±2.09 to ± 2.27 (validity); a negative slope of bias was seen in the ‘prudent’ pattern for reproducibility and validity (P<0.001). Conclusions The FFQ derived DPs demonstrated acceptable reproducibility and relative validity and can be used to examine associations between DPs and health outcomes in older New Zealand adults. Key Message Valid DPs improve evidence for assessing DP and health outcome associations.