Validation of a child version of the Three-Factor Eating Questionnaire in a Canadian sample: a psychometric tool for the evaluation of eating behaviour

Psychometric evaluation of three-factor eating questionnaire -R18 in aging Finnish men with increased risk for type 2 diabetes

BACKGROUND

Deeper comprehension of eating-related behaviour (how and why people eat) can reveal new aspects to support health and prevent type 2 diabetes (T2D). However, such research is largely missing in aging men.

AIM

The aim was to investigate suitability of the Three-Factor Eating Questionnaire-R18 (TFEQ-R18) in Finnish aging men which is widely used to examine factors: cognitive restraint (CR), uncontrolled eating (UE), and emotional eating (EE).

METHODS

Study population consisted of 420 men aged 50-75, who completed the TFEQ-R18 at the baseline of the T2D-GENE lifestyle intervention study. Inclusion criteria were impaired fasting glucose (IFG) and body mass index ≥25 kg/m2. Confirmatory factor analysis was used to study psychometrics (reliability, validity, and model fit) and factor structure of TFEQ-R18.

RESULTS

The items loaded to the three factors (CR, UE, EE) as in previous studies, except two items at CR factor and one at UE factor, which were therefore omitted. UE was also discovered split into two sub factors (named as 'craving' and 'loss-of-control'), UE being a higher-order (h) factor. The resultant revised version was named as Three-Factor Eating Questionnaire Revised to 15-items with higher-order factor (TFEQ-R15h).

CONCLUSION

The original 18-item version of the TFEQ was not optimal in the population consisting of Finnish aging men with elevated T2D risk. A modified 15-item version of the TFEQ could be used to describe EB in this population instead.

Validity and Reliability of the Arabic Version of the Three-Factor Eating Questionnaire-R18

The present study translated and validated an Arabic version of the Three-Factor Eating Questionnaire-R18 (TFEQ-R18), which assesses eating behaviour traits among Arabic speakers in the three domains: cognitive restraint (CR), uncontrolled eating (UE), and emotional eating (EE). Three bilingual experts first translated the English questionnaire into Arabic, and the questionnaire's comprehension was tested among 298 Arabic speakers. Confirmatory factor analysis (CFA) tested the model fit and flagged four items in the questionnaire with low consistency. These items were modified, and the revised Arabic version of TFEQ-R18 was validated in a population of 513 individuals. Validity was evaluated using CFA and Pearson's correlation coefficients, internal consistency using Cronbach's α, and the reproducibility of the questionnaire was confirmed with Bland-Altman analysis and T-test. The construct validity of the Arabic TFEQ-R18 was evaluated by comparing the response score in the population based on gender, BMI and age group. CFA confirmed that the model fit is good and strongly agrees with the collected data. Except for item 15 in the CR domain, all questions showed moderate to high correlation within their respective domains. The Cronbach's α for UE, EE, and CR domains recorded was 0.778, 0.784 and 0.588, respectively. Item 15 had the lowest consistency in the Arabic-TFEQ-R18, and Cronbach's α increased to 0.608 with its removal. There was no significant difference between the first and second attempts of the Arabic TFEQ-R18, indicating good test-retest reliability. Moreover, UE and EE were positively correlated to the BMI of the participants (r = 0.159, p = 0·000; r = 0.158, p = 0·000, respectively). The study concludes that the Arabic TFEQ is a valid and reliable tool for studying the three psychometric domains of UE, EE and CR among Arabic speakers.

Validation of the Adult Eating Behaviour Questionnaire adapted for the French-speaking Canadian population

PURPOSE

The Adult Eating Behaviour Questionnaire (AEBQ) is a newly developed questionnaire adapted from the widely used Child Eating Behaviour Questionnaire. This questionnaire assesses four food approach scales, namely hunger, food responsiveness, emotional overeating (EOE) and enjoyment of food, and four food avoidance scales, namely satiety responsiveness (SR), emotional undereating (EUE), food fussiness and slowness in eating (SE). This study aimed to validate a French version of the AEBQ in controlled conditions among French-speaking adults from Quebec, Canada.

METHODS

The AEBQ was pre-tested through structured interviews with 30 individuals. Participants of the validation study (n = 197, aged 19-65 years) had their height and weight measured and completed the AEBQ, Three-Factor Eating Questionnaire (TFEQ) and Intuitive Eating Scale-2 to assess factorial structure, internal consistency and construct validity. Test-retest reliability over 2 weeks was assessed among 144 participants.

RESULTS

Confirmatory factor analysis indicated an excellent model fit (NNFI = 0.98, CFI = 0.98, RMSEA = 0.03, χ²/df = 1.17) and provided support for the use of the original 8-factor questionnaire. Internal consistency was adequate for most scales (Cronbach's alpha = 0.66-0.94) and moderate to excellent test-retest reliability was observed for all scales (ICC = 0.70-90). Women showed higher levels of EOE and SR, and individuals with overweight and obesity showed higher levels of EOE and lower levels of EUE and SE. Construct validity was also supported by expected correlations with disinhibition and susceptibility to hunger from the TFEQ and intuitive eating.

CONCLUSION

This study indicates that the French AEBQ is a valid and reliable tool to measure eating behaviours in the adult population of Quebec.

LEVEL OF EVIDENCE

Level III: Evidence obtained from well-designed cohort or case-control analytic studies. The data are cross-sectional, but all measurement were undertaken in controlled laboratory conditions and the study provided new information.

Translation and validation of the Child Three-Factor Eating Questionnaire (CTFEQr17) in French-speaking Canadian children and adolescents

OBJECTIVE

To translate and validate the Child Three-Factor Eating Questionnaire (CTFEQr17), assessing cognitive restraint (CR), uncontrolled eating (UE) and emotional eating (EE), among French-speaking Canadian young individuals.

DESIGN

Phase 1 comprised a translation and the evaluation of the comprehension of the questionnaire. Phase 2 comprised a confirmatory factor analysis (CFA), the evaluation of internal consistency (Cronbach's α), test-retest reliability (intra-class correlation coefficients (ICC)) and construct validity, including correlations among the CTFEQr17 and Eating Attitudes Test (EAT-26), anthropometrics, dietary intake and diet quality.

SETTING

Primary and secondary schools, Québec City, Canada.

PARTICIPANTS

Phases 1 and 2 included 20 (40 % boys, mean age 11·5 (sd 2·4) years) and 145 (48 % boys, mean age 11·0 (sd 1·9) years) participants, respectively.

RESULTS

Phase 1 resulted in the questionnaire to be used in Phase 2. In Phase 2, the CFA revealed that the seventeen item, three-factor model (CTFEQr17) provided an excellent fit. Internal consistency was good (Cronbach's α: 0·81-0·90). Test-retest reliability was moderate to good (ICC = 0·59, (95 % CI 0·48, 0·70), ICC = 0·78, (95 % CI 0·70, 0·84), ICC = 0·50, (95 % CI 0·38, 0·62) for CR, UE and EE, respectively). CR correlated with EAT-26 score (r = 0·43, P < 0·0001). UE and EE correlated negatively with BMI Z-scores (r = -0;·26, P = 0·003; r = -0;·19, P = 0·03, respectively). CR correlated with the proportion of energy intake from protein and diet quality (r = 0·18, P = 0·04; r = 0·20, P = 0·02, respectively).

CONCLUSION

The CTFEQr17 is suitable to use among French-speaking Canadian young individuals.

The Role of Eating Behaviours in Genetic Susceptibility to Obesity

PURPOSE OF REVIEW

Eating behaviours are hypothesised to be the behavioural expression of genetic risk of obesity. In this review, we summarise findings from behavioural genetic research on the association between genetic risk for obesity and validated psychometrics measures of eating behaviours in children and adults (published in the past 10 years).

RECENT FINDINGS

Twin studies have produced some evidence for a shared genetic aetiology underlying body mass index and eating behaviours. Studies using measured genetic susceptibility to obesity have suggested that increased genetic liability for obesity is associated with variation in obesogenic eating behaviours such as emotional and uncontrolled eating. More research on this topic is needed. Especially longitudinal studies using genetically sensitive designs to investigate the direction of genetic pathways between genetic liability of eating behaviours to weight and vice versa, as well as the potential subsequent link to eating disorders.

Children's and Mothers' Perspectives of Problematic Eating Behaviours in Young Children and Adolescents: An Exploratory Study

The aim of this study was to (a) compare children's perspectives of problematic eating behaviours with those of mothers and (b) check if there are differences in the level of these problematic eating behaviours between girls and boys in different age groups (young children: 8-11 years old vs. adolescents: 12-16 years old). The study involved 203 children (50.74% girls) and 203 mothers. The average age of children was 11.06 years (SD = 2.31), and the average BMI was 18.27 kg/m2 (SD = 2.29). Two questionnaires were used to assess children's perspectives of problematic eating behaviours: The Three-Factor Eating Questionnaire (TFEQ-R13) and the Dutch Eating Behaviour Questionnaire for Children (DEBQ-C). One questionnaire was used to evaluate mothers' perspectives: The Child Eating Behaviour Questionnaire (CEBQ). The main results in the study indicate the following: (a) the children's perspective based on the DEBQ-C is the most effective at predicting their BMI (this model of problematic eating behaviours explains 29% of the variance in the child's BMI); and (b) for almost all problematic eating behaviours, older girls have the highest levels. From the current study, it can be concluded that the type of questionnaire (TFEQ-R13 vs. DEBQ-C vs. CEBQ) and the perspective (child vs. mother) differentiate the results obtained regarding the assessment of children's problematic eating behaviours and their relation to BMI.