| Summary: | Background: Accuracy of a measurement is a cornerstone in research in order to make robust
conclusions about the research hypothesis. Objective: To examine whether the number of items
(questions) and the number of responses of consumption included in nutritional assessment tools
influence their repeatability. Methods: During 2009, 400 participants (250 from Greece, 37±13
yrs, 34% males and 150 participants from Spain, 39±17 yrs, 41% males) completed a diet index
with 11-items and binary (yes/no) responses, a diet-index with 11-items and 6-scale responses, a
36-item and a 76-item food frequency questionnaire (FFQ) with 6-scale responses. Participants
completed these tools, twice, within 15-days period. Spearman-Brown (rsb), Kendall’s tau
coefficients and the Bland-Altman method were applied to answer the research hypothesis.
Results: The highest repeatability coefficient was observed for the 11-items with binary
responses index (rsb=0.948, p<0.001), followed by the 11-items with 6-scale responses index
(rsb=0.943, p<0.001), the 36-item (rsb=0.936, p<0.001) and the 76-item FFQs (rsb=0.878,
p<0.001). Statistical comparisons revealed no significant differences between repeatability
coefficients of the first three tools (p>0.23); whereas the aforementioned tools had significantly
higher repeatability coefficients as compared with the 76-item FFQ (p=0.002). Sub-group
analyses by gender, education, smoking and clinical status, confirmed the aforementioned
results. Conclusion: Repeatability has been revealed for all food frequency assessment tools
used, irrespective of the number of items or the number of responses included.
|
|---|