Limits...
The Reproducibility and Relative Validity of a Mexican Diet Quality Index (ICDMx) for the Assessment of the Habitual Diet of Adults

View Article: PubMed Central - PubMed

ABSTRACT

The study of diet quality in a population provides information for the development of programs to improve nutritional status through better directed actions. The aim of this study was to assess the reproducibility and relative validity of a Mexican Diet Quality Index (ICDMx) for the assessment of the habitual diet of adults. The ICDMx was designed to assess the characteristics of a healthy diet using a validated semi-quantitative food frequency questionnaire (FFQ-Mx). Reproducibility was determined by comparing 2 ICDMx based on FFQs (one-year interval). Relative validity was assessed by comparing the ICDMx (2nd FFQ) with that estimated based on the intake averages from dietary records (nine days). The questionnaires were answered by 97 adults (mean age in years = 27.5, SD = 12.6). Pearson (r) and intraclass correlations (ICC) were calculated; Bland-Altman plots, Cohen’s κ coefficients and blood lipid determinations complemented the analysis. Additional analysis compared ICDMx scores with nutrients derived from dietary records, using a Pearson correlation. These nutrient intakes were transformed logarithmically to improve normality (log10) and adjusted according to energy, prior to analyses. The ICDMx obtained ICC reproducibility values ranged from 0.33 to 0.87 (23/24 items with significant correlations; mean = 0.63), while relative validity ranged from 0.26 to 0.79 (mean = 0.45). Bland-Altman plots showed a high level of agreement between methods. ICDMx scores were inversely correlated (p < 0.05) with total blood cholesterol (r = −0.33) and triglycerides (r = −0.22). ICDMx (as calculated from FFQs and DRs) obtained positive correlations with fiber, magnesium, potassium, retinol, thiamin, riboflavin, pyridoxine, and folate. The ICDMx obtained acceptable levels of reproducibility and relative validity in this population. It can be useful for population nutritional surveillance and to assess the changes resulting from the implementation of nutritional interventions.

No MeSH data available.


Bland-Altman plots illustrating the relationship between ICDMx score estimated from FFQ2 and the average ICDMx from nine-day DR (n = 97).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5037506&req=5

nutrients-08-00516-f002: Bland-Altman plots illustrating the relationship between ICDMx score estimated from FFQ2 and the average ICDMx from nine-day DR (n = 97).

Mentions: Figure 2 illustrates the Bland-Altman plot for the total ICDMx score. This figure illustrated a high level of agreement between methods. The total score for ICDMx2 tended to overestimate the total score of ICDMxDR by an average of 1.8 points (/100). The ICDMx2 score ranged from 26.5 points above and 23 below the total ICDMxDR score. The slope of this bias was not significantly different from zero (p = 0.65) in the linear regression analysis with the differences between methods as dependent variable and the means of the methods as independent variable. This means that the bias was consistent across scores ICDMx. In terms of the total value of the ICDMx and of each of its components, scores fell mostly within the limits of agreement (±1.96 SD).


The Reproducibility and Relative Validity of a Mexican Diet Quality Index (ICDMx) for the Assessment of the Habitual Diet of Adults
Bland-Altman plots illustrating the relationship between ICDMx score estimated from FFQ2 and the average ICDMx from nine-day DR (n = 97).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5037506&req=5

nutrients-08-00516-f002: Bland-Altman plots illustrating the relationship between ICDMx score estimated from FFQ2 and the average ICDMx from nine-day DR (n = 97).
Mentions: Figure 2 illustrates the Bland-Altman plot for the total ICDMx score. This figure illustrated a high level of agreement between methods. The total score for ICDMx2 tended to overestimate the total score of ICDMxDR by an average of 1.8 points (/100). The ICDMx2 score ranged from 26.5 points above and 23 below the total ICDMxDR score. The slope of this bias was not significantly different from zero (p = 0.65) in the linear regression analysis with the differences between methods as dependent variable and the means of the methods as independent variable. This means that the bias was consistent across scores ICDMx. In terms of the total value of the ICDMx and of each of its components, scores fell mostly within the limits of agreement (±1.96 SD).

View Article: PubMed Central - PubMed

ABSTRACT

The study of diet quality in a population provides information for the development of programs to improve nutritional status through better directed actions. The aim of this study was to assess the reproducibility and relative validity of a Mexican Diet Quality Index (ICDMx) for the assessment of the habitual diet of adults. The ICDMx was designed to assess the characteristics of a healthy diet using a validated semi-quantitative food frequency questionnaire (FFQ-Mx). Reproducibility was determined by comparing 2 ICDMx based on FFQs (one-year interval). Relative validity was assessed by comparing the ICDMx (2nd FFQ) with that estimated based on the intake averages from dietary records (nine days). The questionnaires were answered by 97 adults (mean age in years = 27.5, SD = 12.6). Pearson (r) and intraclass correlations (ICC) were calculated; Bland-Altman plots, Cohen’s κ coefficients and blood lipid determinations complemented the analysis. Additional analysis compared ICDMx scores with nutrients derived from dietary records, using a Pearson correlation. These nutrient intakes were transformed logarithmically to improve normality (log10) and adjusted according to energy, prior to analyses. The ICDMx obtained ICC reproducibility values ranged from 0.33 to 0.87 (23/24 items with significant correlations; mean = 0.63), while relative validity ranged from 0.26 to 0.79 (mean = 0.45). Bland-Altman plots showed a high level of agreement between methods. ICDMx scores were inversely correlated (p < 0.05) with total blood cholesterol (r = −0.33) and triglycerides (r = −0.22). ICDMx (as calculated from FFQs and DRs) obtained positive correlations with fiber, magnesium, potassium, retinol, thiamin, riboflavin, pyridoxine, and folate. The ICDMx obtained acceptable levels of reproducibility and relative validity in this population. It can be useful for population nutritional surveillance and to assess the changes resulting from the implementation of nutritional interventions.

No MeSH data available.