| Sumario: | Background
The diet quality questionnaire (DQQ) is a standardized low-burden tool for collecting data on minimum dietary diversity for women (MDD-W) and other population-level diet quality indicators related to risk of noncommunicable disease (NCD). Although 24-h recalls (24hRs) are often used for evaluating validity of DQQ, they may underestimate consumption of specific food groups. Therefore, comparison with observed weighed food records (OWFR), can provide a more accurate assessment of DQQ criterion validity.
Objective
The aim of this study is to evaluate criterion validity of DQQ for estimating population-level diet quality using OWFR and 24hR as reference methods.
Methods
Cross-sectional data were collected among 281 Rwandan adults (Musanze district), using OWFR, DQQ, and 24hR on the same day. Diet quality indicators derived from each method were compared using parametric and nonparametric methods and the method of triads, which calculates pairwise validity coefficients to evaluate accuracy (low: <0.30; moderate: 0.30–0.70; high >0.70).
Results
Mean percent agreement in food group consumption data was high: 93% (DQQ-OWFR; DQQ-24hR). Compared with OWFR, DQQ overestimated MDD-W-prevalence [DQQ: 46.0% compared with OWFR: 40.4%; +6 percentage points (pp), P > 0.05], whereas using 24hR, the MDD-W-prevalence was 29.8% (16.2 pp < DQQ, P < 0.05, and 10.6 pp < OWFR, P < 0.05). Compared with OWFR, mean scores of food group diversity score (FGDS) and NCD-protect were 0.2 (ns) and 0.2 (P = 0.01) points higher by DQQ, respectively, and 0.4 (P < 0.001) points higher by DQQ compared with 24hR. NCD-risk median scores were 0 across methods. For DQQ, validity coefficients were 0.70 (FGDS), 0.67 (NCD-protect), and 0.66 (NCD-risk), compared with 0.93, 0.89, and 0.59 for OWFR, respectively, and 0.84, 0.83, and 0.98 for 24hR, respectively.
Conclusions
The DQQ showed high agreement with OWFR and 24hR for collecting population-level food group consumption data, and slight overestimations of diet quality indicator scores compared with observed intakes. DQQ is a valid and practical method for collecting population-level food group consumption data and estimating diet quality.
|
|---|