Harokopio University
Αξιολόγηση της επαναληψιμότητας διατροφικών ερωτηματολογίων συχνότητας κατανάλωσης τροφίμων, και των στατιστικών μεθόδων αποτίμησης διατροφικών προτύπων
Abstract
dc:descriptionIntroduction: The accuracy of a measurement is a cornestone in research and especially innutritional research where diet-disease association are explored. Various tools have been developedin order to record the dietary information, with the food frequency questionnaire (FFQ) to be themost commonly used. Due to several limitations during the implementation of the FFQ, it’s essentialto test its repeatability and validity before using them. Under the concept of the holistic nutritionassessment, both a-priori and a-posteriori statistical methods have been proposed, and used toevaluate diet-disease associations. However, their validity of the extracted patterns has rarely beenstudied. Aim: (a) To test the validity of a-priori and a-posteriori defined dietary patterns and to (b)to test whether the number of questions used for the development of a FFQ and the number ofresponse categories, influences it’s repeatability and validity. Material and subject: 500 participants (37±15 years, 38% males) were enrolled to test the first goal of the study (repeatability) by completing a 76-item FFQ, while another 500 participants (46±16 years, 40% males) were enrolled to test the validity of the FFQ by also completing a three day dietary record (3DD, as the reference tool). A sample of 250 participants (37±13 years, 34% males) were enrolled to test whether the number of questions and response categories influence the repeatability and validity of the tool, by also completing a 36-item FFQ with 6-scale responses, a 11-item questionnaire with 6-scale responses and a 11-item questionnaire with binary scale responses. Furthermore, the repeatability and the validity of dietary patterns, using the MedDietScore as the a-priori method, and principal component analysis (PCA) and cluster analysis (CA) as the a-posteriori techniques were tested. Spearman-Brown correlation coefficient, Kendall’s τ agreement coefficient, the Bland & Altman method and the Wilcoxon signed rank test were used to answer the research hypothesis. Results: The degree of repeatability of a FFQ is a decreasing function of the length of the questionnaire and the number of response categories. Moreover, dietary patterns’ identification revealed that the a-priori approach led to robust patterns, which actually reflect participants’ true dietary habits, in terms of its validity; whereas, the a-posteriori approach (either using PCA or CA) led to repeatable patterns only when aggregated dietary information was used. Conclusion: To achieve repeatability of a FFQ a moderate number of questions and response categories is suggested; however, the later may have an impact on the nutritional information retrieved. Concerning dietary patterns’ analysis the a-priori approach gave more robust results and is suggested. The a-posteriori approach led to repeatable and valid dietary patterns only when additional data handling was made (i.e., level of aggregation, axis rotation, reduced number of clusters).
Degree
thesis:*- Grantor dc:publisher
- Harokopio University
- Year dc:date
- 2012
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Μπουντζιούκα, Βασιλική
Subjects
dc:subject × 20- Διατροφή
- Κατανάλωση της τροφής
- Πολυμεταβλητή ανάλυση
- Διατροφική αξιολόγηση
- Ερωτηματολόγια
- Επαναληψιμότητα
- Εγκυρότητα
- Πολυμεταβλητή στατιστική ανάλυση
- Nutrition
- Food consumption
- Multivariate analysis
- Nutrition assessment
- Questionnaires
- Repeatability
- Validity
- Multivariate statistical analysis
- Ιατρική και Επιστήμες Υγείας
- Επιστήμες Υγείας
- Medical and Health Sciences
- Health Sciences
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/33350