Factors associated with food choices among Greek

Journal of Public Health | Vol. 30, No. 3, pp. 266 –273 | doi:10.1093/pubmed/fdn039 | Advance Access Publication 22 May 2008

Factors associated with food choices among Greek primary school students: a cluster analysis in the ELPYDES study Grigoris Risvas1,2, Demosthenes B. Panagiotakos1, Stavroula Chrysanthopoulou1, Konstantina Karasouli2, Antonia-Leda Matalas1, Antonis Zampelas2 1

Department of Nutrition – Dietetics, Harokopio University, El. Venizelou 70, 17671 Kallithea, Athens, Greece Unit of Human Nutrition, Department of Food Science and Technology, Agricultural University of Athens, Iera Odos 75, 11855 Athens, Greece Address correspondence to Antonis Zampelas, E-mail: [email protected] 2

A B S T R AC T Background Food choice in Greece follows a westernized model. We tried to identify the characteristics of clusters regarding food choice and

Methods Cross-sectional study in 2439 fifth and sixth grade students from the Attica and Thessaloniki regions. Three self-administered questionnaires were distributed assessing food consumption, nutrition knowledge and factors associated with dietary change. Data were analysed using principal components analysis (PCA) and K-means cluster analysis. Results A total of 28.4% (n ¼ 592) of the students were identified as demonstrating ‘unbalanced nutrition’ whereas 44.8% (n ¼ 1018) and 22.8% (n ¼ 319) demonstrated ‘balanced’ and ‘low food intake’, respectively. With regards to nutrition knowledge, the clusters were as follows: medium (n ¼ 319, 14.5%), good (n ¼ 1788, 80.9%) and bad knowledge (n ¼ 101, 4.57%) cluster. After analysing the results of PCA, three clusters were formed: A ‘negative effect’ (n ¼ 561, 28.8%), a ‘health oriented’ (n ¼ 777, 39.9%) and a ‘reinforced’ to eat fruits and vegetables (n ¼ 506, 31.3%) group. Conclusions The present study managed to identify clusters that correspond to food intake, nutrition knowledge and other factors associated with dietary behaviour and to describe their characteristics. Keywords health promotion, statistical methods

Background Food choice in Greece seems to follow a westernized model far from the Mediterranean Diet.1 Data from surveys in Greece2,3 showed that added sugar provided .10% and dietary fat accounted for 40% of daily energy intake, while 50% and 30% of a large sample of school-aged adolescents consumed soft drinks and sweets/confectionery, respectively, on a daily basis. Consecutively, a 202% rise in the prevalence of obesity has been observed in a sample of Cretan boys aged 12 years, from 1982 to 2002. During this 20-year period, there is also a significant increase in cardiovascular disease risk factors, namely total and LDL-cholesterol.4 Thus, there is a need to change the aforementioned nutrition habits, in order to promote health in Greek youth. Interventions aiming to change health behaviour represent efforts to modify certain mediating parameters.

266

For this reason, methods employed in health promotion programmes involve an initial assessment, in order to clarify the behaviours that need to be targeted.5 When students are the target, school is a unique setting to implement health programmes, as it provides access to the entire student population.6 Although more public health programmes within schools are needed before best practice examples may be created, the distinct needs of certain groups of students cannot be addressed unless assessed systematically.7

Grigoris Risvas, PhD Candidate Demosthenes B. Panagiotakos, Principal Epidemiologist Stavroula Chrysanthopoulou, Biostatistician Konstantina Karasouli, Nutritionist – Research Fellow Antonia-Leda Matalas, Associate Professor of Nutrition Anthropology Antonis Zampelas, Principal Investigator – Associate Professor of Human Nutrition

# The Author 2008, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved

Downloaded from http://jpubhealth.oxfordjournals.org/ by guest on June 8, 2013

behaviour in a large sample of Greek primary school students, in order to acknowledge some mediating parameters that need to be addressed when planning interventions to promote healthy nutrition.

C LU S T ER S R EG A R D IN G FO O D C H O IC E I N G R EE K S TU D EN TS

Methods The ELPYDES study is a cross-sectional study in a large sample of 4752 students in Greece. Details of sampling methods for the ELPYDES study have been published previously.10 In the present paper, the results regarding 2439 fifth and sixth grade students, from the aforementioned schools, will be reported. Three self-administered questionnaires were administered to a group of students (fifth and sixth grade) evaluating: nutritional knowledge, self-efficacy and social support for dietary change and consumption of 15 food groups. The Nutrition Knowledge Questionnaire included demographic information on age, school and grade, sex, nationality of parents and adults living with the student or caretakers and continued with 14 nutrition knowledge multiple choice questions. It was based on sections of a respective questionnaire used in the CATCH study11 and validated for assessing nutrition knowledge in the current population. It has been translated in Greek and back-translated in English by two different groups of experts, same as all the questionnaires used in the study. All questionnaires have been modified for regional dietary habits by another group of experts in our institution. The self-efficacy and social support for dietary change Questionnaire was based on a respective questionnaire used by Parcel et al.,12 which consisted of a scale developed to evaluate self-efficacy of elementary school students regarding nutrition issues (Child Dietary Self-Efficacy Scale—CDSS).

Further details and validation of this questionnaire have been reported previously.10 A simple Food Frequency Questionnaire was also used, which included 15 food groups corresponding to the food groups in the Mediterranean Diet pyramid plus carbonated soft drinks. The reliability of the questionnaire has been measured using the Cronbach alpha statistic; the result is satisfactory (a ¼ 0.64), while its repeatability was tested and verified at an earlier stage in a sub-sample of 50 fifth and sixth grade elementary students. Further details on this questionnaire have been reported previously.10 The study was approved by Harokopio University Human Research Committee.

Statistical analysis

For the description of the results, frequencies (and percentages) for categorical variables and means + standard deviations or medians (and range) for continuous or ordinal variables were used. Normality tests have been performed using the Kolmogorov –Smirnov criterion. Tests for differences in mean values of continuous or ordinal variables between two groups have been performed with Student’s t-test (for the normally distributed) or the Mann –Whitney test (for the discrete or skewed variables). Tests for mean differences of continuous or ordinal variables between more than two groups have been performed with analysis of variance (ANOVA) or Kruskal – Wallis test, respectively. Tests for differences in percentages between two or more groups have been performed with Z-statistic. In order to evaluate potential correlations between continuous variables and to reduce the dimensions of the initial information, a multivariate technique has been performed, the principal component analysis (PCA). It is known that in order for the results from a PCA to have some value, strong correlations between the variables need to exist; the correlation matrix of the variables used in the present analysis showed that there were several correlation coefficients j r j . 0.2, indicating that the variables were correlated together and, therefore, a PCA could be effective for assessing various patterns. PCA was performed in the variables that evaluate the perceived ability of students to choose certain foods (PCA1) and the variables that evaluate their perception on the support they can have when trying to make these choices (PCA2). Finally, a combined PCA (PCA3) on the remaining components after the application of PCA1 and PCA2 was performed. In order to group subjects based on various characteristics, clustered analysis was also performed, particularly the k-means technique. As a measure of distance, the Euclidian distance was used for continuous variables, whereas for


Nutrition education interventions in Greece have been traditionally focused on improving knowledge of the students through classroom teaching and homework. In one published paradigm, researchers succeeded in improving knowledge, body mass index (BMI) and some of the biochemical parameters in primary school children by focusing more on behaviour change.8 However, the intervention had greater beneficial effects in certain subgroups and fewer or no effects in others.9 In order to address the issue of longterm success of such programmes, certain clusters of the student population sharing common characteristics and needs must be identified. Under the context of the ELPYDES study10 designed to assess the needs of students for subsequent nutrition education programmes in Greek primary schools, we tried to identify characteristics of clusters regarding food choice and behaviour in a sample of Greek schoolchildren, in order to acknowledge the mediating parameters that need to be addressed when planning interventions.

267

268

J O U RN A L O F P U B L I C H E A LTH

categorical (2-value) variables, the simple-matching coefficient was preferred. For the selection of the final number of clusters, the aforementioned methods were used repeatedly, by changing the predetermined number of clusters each time, and the results were analysed empirically. The decision on the final number of clusters used in our analysis depended on the nature of the data. All statistical tests have been performed at a level of statistical significance of 0.01. The analysis has been performed using the STATA 8 software (College Station, TX, USA) and the pca procedure was used to perform the first analysis. The orthogonal rotation (rotate with varimax option) was used to derive optimal non-correlated components (i.e. patterns). The information was rotated in order to increase the representation of each food or food group to a component.

A total of 2439 students from fifth and sixth grade with a mean age of 11.3 + 0.77 years and a participation rate of 95% were included in the present analysis.

groups Food

Student clusters

groups Group 1 (n ¼ 592)

Group 2 (n ¼ 1018) Group 3 (n ¼ 476)

‘Unbalanced

‘Balanced nutrition’ ‘Low food intake’ Consumptiona + SD Consumptiona + SD

nutrition’ a

Consumption + SD Fruits

5.38 + 1.39

5.38 + 1.23

5.02 + 1.52

Vegetables 4.74 + 1.57

4.58 + 1.53

4.22 + 1.75

Red meat

3.87 + 1.35

2.92 + 1.15

3.32 + 1.43

Chicken

3.76 + 1.38

2.97 + 1.10

3.36 + 1.36

Fish

3.32 + 1.45

3.05 + 1.18

3.12 + 1.45

Dairy

5.83 + 1.17

5.45 + 1.29

5.18 + 1.71

Starch

5.79 + 1.10

5.01 + 1.49

5.24 + 1.52

Potatoes

4.66 + 1.41

3.58 + 1.28

4.45 + 1.46

Olive oil

5.67 + 1.17

4.47 + 1.72

3.91 + 1.92

Other fat

4.84 + 1.32

2.12 + 1.33

1.73 + 1.00

Legumes

4.00 + 1.56

3.41 + 1.34

3.54 + 1.67

Nuts

3.47 + 1.80

2.04 + 1.28

3.85 + 1.97

Eggs

4.17 + 1.59

3.88 + 1.74

3.11 + 1.35

Sweets

4.65 + 1.67

4.49 + 1.56

2.70 + 1.43

Beverages

5.22 + 1.42

4.07 + 1.78

2.41 + 1.35

Principal component analysis

The results of the PCA have been published previously.9 However, it should be noted that PCA3, which was performed on the aforementioned components, extracted six components that explained 66% of the total variation in factors associated with dietary change: (a) readiness to make the healthiest choice when competitive foods are considered, (b) impact of parents, friends and advertisements on students’ choices and also impact of taste and smell and cooking method on the consumption of vegetables, (c) readiness to choose fresh foods instead of ready-to-eat, prepackaged choices, as a main meal and awareness of health value of fruits and vegetables, (d) tendency to consume fresh fruits and vegetables, as a result of the impact parents have in students’ choices, (e) overall effect of social support in food choice and (f) impact of parents eating behaviour and food preparation methods, including chopping and pilling of fruits and advertisement on consumption of sweets. Cluster analysis

Cluster analysis was based on three different characteristics of the subjects, namely consumption of certain food groups (Table 1), nutrition knowledge (Table 2) and the components of PCA3 regarding factors associated with dietary change (Table 3). In the three tables derived by each analysis, descriptive statistics of each cluster are given.

a

Values correspond to a scale where 1: rarely/never and 7: many times/day.

When considering food consumption, our sample can be clustered into three groups (Table 1). The first group included 592 (28.4%) students that formed the ‘unbalanced nutrition’ cluster; specifically they consumed more fat and sugar compared with other clusters. Interestingly, fruit and vegetable intake of the ‘unbalanced nutrition’ group was similar to that of the ‘balanced nutrition’ group. The second group included 1018 (48.8%) students, who followed a ‘balanced nutrition’ and the third group of 476 (22.8%) students, boys and girls who generally consumed less food compared with others, thus creating the ‘low food intake’ group. Table 2 shows that, based on the percentage of correct answers given by the students, three clusters may be created. The first group included 1788 (80.9%) students with good knowledge on all nutrition issues, since in this group the highest percentage of correct answers was observed, among all questions on nutrition knowledge. The second group included 319 (14.5%) students with medium nutrition knowledge, especially as far as fruits and vegetables were concerned and the third included 101 (4.57%) students who had the lowest percentage of correct answers in almost all questions on nutrition knowledge.


Results

Table 1 Clustering of students according to consumption of 15 food


269

Table 2 Clustering of students according to their nutrition knowledge Nutrition knowledge

Student clusters Group 1 (n ¼ 1788) ‘Good

Group 2 (n ¼ 319) ‘Medium

Group 3 (n ¼ 101) ‘Bad

knowledge’ Frequenciesa (%)



Breakfast every morning

1756 (98.2)

304 (95.3)

94 (93.1)

Large amount of food does not cause weight gain

1115 (62.4)

208 (65.2)

40 (39.6)

Which food has no fat?

1566 (87.6)

236 (74)

21 (20.8)

Large amounts of any kind of food may be harmful

1566 (87.6)

236 (74)

21 (20.8)

Beverages have amounts of sugar

1591 (89)

269 (84.3)

69 (68.3)

1517 (84.8)

237 (74.3)

15 (14.9)

How many times/day should you eat fruits?

1517 (84.8)

237 (74.3)

15 (14.9)

Vegetables protect our organism from diseases

1740 (97.3)

302 (94.7)

53 (52.5)

Vegetables contain important substances

1375 (76.9)

27 (8.5)

26 (25.7)

How many times/day should you eat vegetables?

1375 (76.9)

27 (8.5)

26 (25.7)

Substitute vegetables with fruits

1721 (96.3)

294 (92.2)

85 (84.2)

If you eat lots of fruits there is no need of vegetables

1755 (98.2)

308 (96.6)

81 (80.2)

Which food has lesser salt?

1755 (98.2)

308 (96.6)

81 (80.2)

Obese children become obese adults

1446 (80.9)

1 (0.3)

76 (75.3)

a

Values correspond to frequencies (and percentages) of correct answers for each question posed.

Table 3 Clustering of students according to PCA3 Principal components derived from PCA3 (refer to text)

Readiness to make the healthier choice Impact of parents, friends, advertisements, taste and smell and

Student clusters Group 1 (n ¼ 561)

Group 2 (n ¼ 777)

Group 3 (n ¼ 506)

‘Negative effect’

‘Health oriented’

‘Positive effect’

13.57 + 1.89

17.35 + 1.25

20.87 + 1.58

5.41 + 1.71

4.05 + 1.64

5.5 + 2.12

4.05 + 1.84

5.03 + 1.55

4.08 + 1.54

1.06 + 1.6

1.21 + 1.48

1.06 + 1.4

cooking method Readiness to choose fresh foods and awareness of health value of fruits and vegetables Tendency to consume fresh fruits and vegetables, as a result of the impact parents have in the students’ choices Overall effect of social support in food choice

0.92 + 1.52

0.75 + 1.44

Impact of parents eating behaviour and preparation methods on

0.29 + 1.46

20.15 + 1.38

0.88 + 1.47 0.16 + 1.4

fruits’ consumption and advertisement on sweets consumption Values are means + SD which show the degree in which the relevant components refer to each cluster. The bigger the absolute value the more associated is the component with the cluster.

Table 3 shows the mean values for each of the six components of PCA3 as described above. Three clusters were formed that can be characterized as follows: a group of 561 (28.8%) students who receive a negative effect from their environment regarding fruits and vegetables consumption, a second group of 777 (39.9%) students, who base the choice

of any food on whether it is fresh or ‘healthy’ and a group of 506 (31.3%) students who are reinforced by their environment to eat fruits and vegetables. Each one of the aforementioned clusters was further examined in relation to the other variables of interest. Tables 4 and 5 present the descriptive statistics of the


A beverage is as nutritious as a fresh fruit juice

270


Table 4 Descriptive statistics for the three student clusters according to consumption of 15 food groups on variables regarding knowledge, factors associated with dietary change and BMI (only significant differences shown) Variables

Student clusters

P-value

Group 1 (n ¼ 632)

Group 2 (n ¼ 875)

Group 3 (n ¼ 579)

‘Bad diet’

‘Balanced nutrition’

‘Low-consumers’

Knowledge score (range, 0 –14)

11.38 + 1.81

11.58 + 1.82

11.33 + 1.88

0.019

Readiness to make the healthier choicea

17.17 + 3.09

18.09 + 3.04

16.68 + 3.27

,0.001

0.97 + 1.44

1.24 + 1.54

1.06 + 1.47

0.002

Tendency to consume fresh fruits and vegetables, as a result of the impact parents have in the students’ choicesa All values are means + SD. a

Values show the degree in which the relevant components refer to each cluster. The bigger the absolute value the more associated is the component with

Table 5 Descriptive statistics for the three student clusters according to PCA3 on variables regarding dietary behaviour, stages of dietary change, nutrition knowledge and BMI (only significant differences shown) Variables

Student clusters

P-value

Group 1 (n ¼ 561)

Group 2 (n ¼ 777) ‘Health

Group 3 (n ¼ 506)

‘Negative effect’

oriented’

‘Positive effect’

Fruits

2.53 + 1.31

2.64 + 1.29

3.03 + 1.51

0.0001

Vegetables

3.21 + 1.52

3.28 + 1.49

3.81 + 1.74

0.0001

Fish

4.68 + 1.33

4.93 + 1.23

5.01 + 1.39

0.0001

Other fat

5.37 + 1.78

5.11 + 1.83

5.01 + 1.89

0.002

Legumes

4.24 + 1.55

4.41 + 1.47

4.52 + 1.52

0.005

Sweets

4.71 + 1.78

4.26 + 1.73

4.01 + 1.84

0.0001

Beverages

4.67 + 1.87

4.49 + 1.85

4.22 + 2

0.0004

Knowledge score (range: 0-14)

11.14 + 1.92

11.47 + 1.77

11.49 + 1.77

Readiness to make the healthier choice

13.57 + 1.89

17.35 + 1.25

20.87 + 1.58

0.0001

5.41 + 1.71

4.05 + 1.64

5.5 + 2.12

0.0001

4.05 + 1.84

5.03 + 1.55

4.08 + 1.54

0.0001

0.29 + 1.46

20.15 + 1.38

0.16 + 1.4

0.0001

Food groupsa

Impact of parents, friends, advertisements, taste and smell and cooking method

0.003

b

Readiness to choose fresh foods and awareness of value of fruits and vegetablesb Impact of parents eating behaviour and preparation methods on fruits and advertisement on sweets consumptionb All values are means + SD. a

Values correspond to a scale where 1: rarely/never and 7: many times/day.

b

Values show the degree in which the relevant components refer to each cluster. The bigger the absolute value the more associated is the component with

the cluster.

clusters regarding knowledge, dietary behaviour and factors associated with dietary change, as well as the P-values for the differences observed between clusters. In particular, there were no statistically significant differences in dietary behaviour and factors associated with dietary change

between the three levels of nutrition knowledge described by this cluster analysis of the students (data not shown). Table 4 shows statistically significant differences between the three student clusters with regard to the consumption of 15 food groups on readiness to make the healthier choice


the cluster.


Discussion Main findings of this study

Participants were classified into three sub-groups on the basis of their food choice (balanced nutrition, unbalanced nutrition and low food intake), another three groups based on their knowledge (good, medium and bad) and three more clusters based on reinforcement they get from their environment to follow a healthier diet or their perceptions on the health value of foods ( positive effect, negative effect and health oriented).

What is already known on this topic?

The purpose of cluster analysis is to place subjects into groups suggested by the data, not defined a priori, such that subjects in a given cluster tend to be similar to each other in some sense, and subjects in different clusters tend to be dissimilar. Cluster analysis offers advantages over the alternative quantitative approaches (i.e. PCA) as it aims to identify distinct, relatively homogeneous groups based upon selected attributes. Until today, cluster analysis has been used primarily to describe dietary patterns13 and it has never been used to describe factors associated with food choice, hence trying to investigate how these patterns are formed. Cluster analysis cannot be used to identify why individual students eat, what they eat; however, it identifies groups of students and can therefore be used to target interventions to group characteristics. The main factors associated with food choice, according to the Social Cognitive Theory,14 include behavioural capability, self-efficacy and reciprocal determinism. Rimal15 has explained the relationship between behavioural capability and self-efficacy very thoroughly using data from three crosssectional and two longitudinal studies in the US. He showed that knowledge– behaviour correlations were greater among those with high compared to low self-efficacy and decreased among those who decreased their self-efficacy. As a result, it seems that individuals’ ability to act in knowledge-consistent ways is largely a function of their perceived abilities. If perceived abilities constitute the internal component in reciprocal determinism, external factors, such as parents and television advertisement, affect children’s choice mutually. Nevertheless, recent studies have shown that watching television during family meals was associated with a more healthful diet than not eating regular family meals at all,16 showing that parents’ positive example can overwhelm any negative influence. Moreover, in a study looking at perceived important influences on food choice of adults in the EU the five most important factors were quality, price, taste, family preferences and trying to eat healthily.17 Thus, parent’s positive behaviour towards availability and consumption of fruits and vegetables is expected to boost their offspring’s intake of these food groups. Another hypothesis could be that advertising and encouraging the benefits of increasing fruit and vegetable consumption as a tasteful and healthy choice could result in an increased intake, for the population segment that values health as an important factor of the decision process regarding nutrition. What this study adds

The sample can be discerned in three groups as far as food consumption is concerned. There are not any other studies


and tendency to consume fresh fruits and vegetables, as a result of the impact parents have in the students’ choices. Specifically, students of the ‘balanced nutrition’ group seemed to score higher in the components that expressed readiness to make the healthier choice and tendency to consume fruits and vegetables following the advice of their parents. These students showed better nutrition knowledge too, although this difference was of borderline significance. Finally, Table 5 shows that several significant differences emerged among the three student-clusters when the environmental effect on food choice or ‘healthy’ connotation of foods was considered. First, the group which is reinforced by the environment to eat more fruits and vegetables, followed a more balanced diet overall, as students in this group, besides fruits and vegetables, also consumed more frequently fish and legumes and less frequently fat, sweets and sugar-sweetened beverages compared with their counterparts in the ‘negative effect’ group. No differences in starch, dairy and olive oil were acknowledged; however, the frequency stated herein is within the desirable limits. Nutrition knowledge was high for all groups; nevertheless, the third group gave somewhat better answers than the other two groups. As far as factors associated with food choice were concerned (Table 5), certain factors can be identified that played a more important role for the cluster that was reinforced to eat more fruits and vegetables, whereas others played a role for the other two groups. Specifically, advertisements played the biggest negative role with regards to fruit and vegetable consumption and the only counteracting measure could be parents’ example to their offspring. Readiness to make the healthier choice when competitive foods are considered was the most important factor for the students who received reinforcement from their environment to eat fruits and vegetables. Lastly, the freshness and health value of a food played an important role only for the cluster that based choice on these criteria.

271

272


Limitations of this study

An important limitation of our study is its cross-sectional design, which provides only information about association, and not prediction or causality. In addition, data from all regions of Greece were not collected, in order to discern factors associated with food choice in semi-urban and rural areas. Nevertheless, the clustering procedure cannot provide groups that can be called representative of the entire population; therefore, the information obtained refer merely to our sample. In addition, it is highly likely that food intake, nutrition knowledge, self-efficacy and social support are all mediated by common socioeconomic factors such as family income, parents’ educational level and employment status,

housing, and so on. It would be more revealing if the relationship between such socioeconomic factors and nutritional behaviour among the Greek student population could be elucidated, in order to plan and introduce public health interventions at population level. However, due to the lack of socioeconomic indices in Greece, data on socioeconomic status cannot be reported, since any attempt to categorize the socioeconomic parameters of the children’s families would be prone to subjective bias.

Conclusions The present study managed to identify clusters that correspond to food intake, nutrition knowledge and other factors associated with dietary behaviour. These groups seem to correspond to earlier observations regarding nutrition habits of Greek schoolchildren and verify the role of self-efficacy as a mediator of the knowledge– behaviour relation. Furthermore, parents seem to play a positive role on their offsprings’ nutrition behaviour, whereas advertisements seem to have a negative effect. However, further research is needed in order to confirm or refute our results in a representative sample of the population. Moreover, socioeconomic characteristics may be considered critical for studies of nutritional behaviour; hence, socioeconomic indices in Greece would be very useful in order to better understand how food choice is made in different parts of the population and target interventions to group characteristics. Finally, health still plays a major role on food choice for a considerable part of the population, but the way ‘healthy’ is defined still remains to be clarified.

Funding This work was supported by the Iraklitos Program ‘Research Scholarships in Harokopio University’ co-funded by the E.U. and the Hellenic Ministry of Education.

References 1 Arvaniti F, Panagiotakos DB, Pitsavos C et al. Dietary habits in a Greek sample of men and women: the ATTICA study. Cent Eur J Publ Health 2006;14:74– 7. 2 Roma-Giannikou E, Adamidis D, Gianniou M et al. Nutritional survey in Greek children: nutrient intake. Eur J Clin Nutr 1997;51:273 – 85. 3 Yannakoulia M, Karayiannis D, Terzidou M et al. Nutrition-related habits of Greek adolescents. Eur J Clin Nutr 2004;58:580 – 6.


that classify a Greek student population in groups according to their dietary habits; however, food choice has been reported in the past as following a westernized model far from the Mediterranean Diet.2,3 Therefore, it is not surprising that only half of our sample could be classified as being on a balanced diet. On the other hand, nutritional discrepancies may be observed in both extremes, thus creating a large group of low food intake, just below one quarter of our sample, where the majority of Greek students that reported being on a diet or dissatisfied with their weight could probably be classified,3 which accounted for 37.6% of the relevant population. As far as knowledge is concerned, the vast majority of the students could be classified as having good nutrition knowledge, indicating that knowledge does not need to be increased in order to succeed in nutrition behaviour change. Students who received the proper reinforcement to eat more fruits and vegetables had both the knowledge and intention to perform this behaviour. Advertisements exhibited the most pronounced negative role with regards to fruits and vegetables consumption, which is not surprising as they are considered to be an integral part of the ‘toxic environment’ and have been accused of playing a major role in the aetiology of obesity.18 On the other hand, parental modelling has been related to fruit and vegetable intake in the past,19 and it can be a very successful counteracting measure to negative influence on their offspring’s eating habits, as shown also in our study. Finally, for a considerable proportion of our sample, health is an important consideration when making choices about which foods to eat. Additionally, due to the tradition of Mediterranean Diet, which is promoted in schools through health education programmes and taught as curriculum, Greek students and their parents have a tendency to regard anything fresh and traditional as being the healthier choice.


273

4 Magkos F, Manios Y, Christakis G et al. Secular trends in cardiovascular risk factors among school-aged boys from Crete, Greece, 1982 – 2002. Eur J Clin Nutr 2005;59:1 – 7.

12 Parcel G, Edmundson E, Perry C et al. Measurement of self-efficacy for diet-related behaviors among elementary school children. J Sch Health 1995;65:23 –27.

5 Bartholomew LK, Parcel GS, Kork G. Intervention mapping: a process for developing theory- and evidence-based health education programs. Health Educ Behav 1998;25:545 – 63.

13 Pryer JA, Nichols R, Elliott P et al. Dietary patterns among a national random sample of British adults. J Epidemiol Community Health 2001;55:29 –37.

6 Harrell JS, McMurray RG, Bangdiwala SI et al. Effects of schoolbased intervention to reduce cardiovascular disease risk factors in elementary-school children: The Cardiovascular Health in Children (CHIC) Study. J Pediatr 1996;128:797– 805.

14 Perry CL, Baranowski T, Parcel GS. How individuals, environments, and health behavior interact: Social Learning Theory. In: Glansk K, Lewis FM, Rimer BK (eds), Health Behavior and Health Education: Theory, Research and Practice. San Francisco: Jossey-Bass, 1990, 161 – 86.

7 Flynn MA, McNeil DA, Maloff B et al. Reducing obesity and related chronic disease risk in children and youth: a synthesis of evidence with best practice recommendations. Obes Rev 2006;7:1 –5. 8 Manios Y, Kafatos A. Health and nutrition education in elementary schools: changes in health knowledge, nutrient intakes and physical activity over a six year period. Public Health Nutr 1999;2:445– 8.

15 Rimal RN. Closing the knowledge-behavior gap in health promotion: the mediating role of self-efficacy. Health Commun 2000;12:219– 37.

9 Kafatos A, Manios Y, Moschandreas J, Preventive Medicine and Nutrition Clinic University of Crete Research Team. Health and nutrition education in primary schools of Crete: follow-up changes in body mass index and overweight status. Eur J Clin Nutr 2005;59:1090 – 2.

17 Lennernas M, Fjellstrom C, Becker W et al. Influences on food choice perceived to be important by nationally-representative samples of adults in the European Union. Eur J Clin Nutr 1997;51:S8– S15.

10 Risvas G, Panagiotakos DB, Zampelas A. Factors affecting food choice in Greek primary-school students: ELPYDES study. Public Health Nutr 15 Oct 2007 [Epub ahead of print].

18 Lustig RH. The ‘skinny’ on childhood obesity: how our western environment starves kids’ brains. Pediatr Ann 2006;35:898 –902, 905 – 7.

11 Perry C, Stone E, Parcel G et al. School-based cardiovascular health promotion: The Child and Adolescent Trial for Cardiovascular Health (CATCH). J Sch Health 1990;60:406– 13.

19 Domel SB, Thompson WO, Davis HC et al. Psychosocial predictors of fruit and vegetable consumption among elementary school children. Health Educ Res 1996;11:299 – 308.


16 Feldman S, Eisenberg ME, Neumark-Sztainer D et al. Associations between watching TV during family meals and dietary intake among adolescents. J Nutr Educ Behav 2007;39:257– 63.