Health-related quality of life,Mediterranean diet,physical activity and socioeconomic factors of Greek adolescents during COVID-19: A cross sectional study

2.1 Sampling and participants

In this cross-sectional research study 459 adolescents (47.1% boys; 52.9% girls) participated. They were aged between 12–16 years, attending schools throughout the Attica region. The research took place during May–June 2021. 1 A proportionate random stratified sampling method according to students’ population of secondary education of the school year 2018/ 2019 in Attica (Hellenic Statistical Authority), was employed and questionnaires were distributed in 7 public high schools in the Attica area. 2 The participation in the study was voluntary and took place during school hours. Prior to this, informed consent forms signed from the corresponding parent or guardian were collected. All participants received explanation of the purpose of the study and were reassured for the confidentiality of their answers (according to GDPR). Beforehand they were given written and oral instructions for filling in the questionnaire. The main investigator (S.P.) was present in the classroom in order to provide guidance during the completion of the questionnaires, which lasted approximately 30 minutes. The research was approved by the West Attica University Research Ethics Committee (18092- 03/03/2021) and the Institute of Educational Policy of Greece (59235/2021).

2.2 Measures

A structured questionnaire was developed for the study. We evaluated self-reported anthropometric data and used the International Obesity Task Force (IOTF) cut off points to categorize the adolescents into four groups, underweight normal, overweight and obese. Perceived HRQoL was assessed by the KIDSCREEN-27 questionnaire for children and adolescents. The adolescents also completed the KIDMED test and the Godin and Shephard Leisure Time Physical Activity Questionnaire (GSLTPAQ). To evaluate the socioeconomic status (SES) the Family Affluence Scale (FAS) III was used.

All the mentioned instruments are reliable and validated for the Greek population or different countries and are widely applied in research studies [26–29]. Also, a collaboration form for the use of the questionnaires was signed for the KIDSCREEN and the FAS III between The KIDSCREEN Group Europe and the University of Mental Health, Neurosciences and Precision Medicine, Research Institute Costas Stefanis. Finally, KIDMED and GSLTPAQ questionnaires were translated into the Greek language using the forward- backward translation method. The questionnaire included general background questions about adolescent demographic and anthropometric parameters, sedentary and physical activities, eating behavior, life style characteristics and sleep patterns. The questionnaires were first piloted using a group of 50 pupils in order to identify any ambiguous questions.

The HRQoL was assessed using the KIDSCREEN-27 questioner for children and adolescents [30] measuring 5 HRQoL dimensions: physical well-being (5 questions), psychological well-being (7 questions), autonomy and parent relations (7 questions), social support and peers (4 questions) and school environment (4 questions). These five dimensions provide insights for children and adolescents in relation to physical activities, fitness, health, satisfaction and positive feelings, the interaction with their parents and peers, the autonomy they have based on their perception, the satisfaction they gain with their financial resources, their social relations and finally, their feelings about school and their learning process. Higher score indicates better self-perceived HRQoL.

AMD was assessed using the KIDMED test [31] which classifies individuals into three categories on the KIDMED index -low, medium and optimal adherence to MD-, based on their YES or NO answers to 16 questions. Twelve of the questions with positive connotation to the Mediterranean Diet (MD) are assigned with +1and four of them with negative connotation are assigned with –1. In case the individual’s total score is higher than 8, the responder has an optimal MD. In case it is between 4 and 7, improvements need to be made to reach the MD patterns. In case the score is less than 3 the diet’s quality is very low.

The level of Physical Activity (PA) was assessed using the Godin and Shephard Leisure Time Physical Activity Questionnaire [32], based on self-reported weekly PA frequencies. The leisure activity score is based on the sum of the weekly frequencies of strenuous, moderate and mild activities multiplied by 9, 5 and 3 respectively. The adolescents were classified into three categories: active with a score of more than 24 units, moderately active with a score between 14 and 23 units and insufficiently active with a score less than 14 units.

Finally, the FAS III [28], implemented by the World Health Organization’s (WHO) Health Behaviour in the School-aged Children (HBSC) 2017/ 2018 study, was used to estimate adolescents socioeconomic assets. The FAS III includes 6 questions about material assets in the household such as owning a family car/ van/ truck (with responses no, one, two or more), having an own bedroom (no, yes), number of bathrooms (none, one, two, more than two), having a dishwasher (no, yes), number of computers in the household (none, one, two, more than two) and finally, the frequency of family travel abroad for holiday/vacation during the past year (not at all, once, twice, more than twice). According to their affluence score, adolescents were classified into three categories, low affluence (20%), medium (60%) and high (20%) [33].

The anthropometric questions involved data about gender, age, height and weight. The BMI (kg/ m²) was calculated according to the participants self-reported data about their weight and height. Greek adolescents were categorized accordingly as underweight normal weight, overweight and obese following the IOTF sex- and age- specific BMI cut- off points [34, 35].

2.3 Statistical and econometric analysis

Due to the sampling procedure described above, the present data yields robust estimates in this study. Descriptive statistics were performed using frequencies (N), percentages, means and standard deviations. Cronbach’s alpha coefficient was used to measure internal consistency. The normality assumption of the sample seemed to be reasonable because of sample size and of the low values of skewness and kurtosis coefficients (< |1|) [36], except the distributions by weight categories where Kruskal Wallis tests were employed for estimating the statistically significant differences between several groups of adolescents. For the statistical and econometric analyses, first of all the correlation coefficient was determined with possible associations between total HRQoL score along with the aforementioned five dimensions and BMI, AMD, PA, family affluence and other lifestyle variables. Secondly, T-test, Anova and Kruskal Wallis tests were employed for estimating the statistically significant differences between several groups of adolescents. Thirdly, Chi-square and multiple linear regression analysis were performed to study the effect of demographic, socioeconomic parameters and AMD on adolescent’s HRQoL. More specifically, the following equation was estimated: HRQo L i = a i + b i X i + e i

Where HRQoL _i is the total score of KIDSCREEN-27 for each individual, a _i is the constant term, X _i is the vector of socioeconomic and anthropometric characteristics, b _i is the estimated coefficient for each independent variable and e _i is the error term for the estimated regression. 3 .

3.1 Sample characteristics

The following table (Table 1) presents the sample characteristics in total and by gender.

The sample consists of 241 girls (52.9%) and 215 boys (47.1%). The majority of girls indicate that they read around 2–3 hours per day while boys read less (41.4% read around 1hour per day) and used screens more (31.7% ∼ around 2–3h per day). Also, one in four boys go out for a walk almost every day but this applies to only 13.8% of the girls. Regarding their nutrition habits, around 11% of the adolescents consume breakfast with their families every day, while more than 54% consume lunch and around 51% consume dinner together. Out of the total sample 72.3% of the adolescents were of normal weight, 14.6% were overweight and 6.1% were obese. Boys were found to be overweight (17.3%) and obese (8.1%) in higher rates comparing to girls (12.2% and 4.2% respectively). Also, a low presence of 17.8% of the adolescents were found to have high AMD, 56.7% had medium adherence and 25.5% low AMD. A greater percentage of girls, compared to boys, had medium AMD.

The value of the Cronbach’s alpha for KIDSCREEN-27 was higher than 0.7 in each dimension showing acceptable reliability. In particular, it was found a = 0.75 for physical health, a = 0.71 for psychological well-being, a = 0.74 for parent relations and autonomy, a = 0.79 for social support and peers and a = 0.78 for school environment while the mean test scale was a = 0.79. Regarding the estimated scores of KIDSCREEN-27, the following Table 2 presents mean scores and standard deviations of the selected anthropometric characteristics, PA, the family affluence and the AMD for each KIDSCREEN dimension. Furthermore, T-test, Anova and Kruskal Wallis comparison tests are employed for two-group and more than two-groups.

Significant differences between gender were observed. Boys scored higher in almost every KIDSCREEN dimension except from school environment (p-value = 0.86) and social support and peers (p-value = 0.19). In particular, boys scored significantly higher in physical health (p-value <0.001), psychological well-being (p-value <0.001), parent relations and autonomy (p-value <0.01) and total KIDSCREEN score (p-value <0.001).

Adolescents who are physically more active have better scores in total KIDSCEREEN score (p-value <0.1) and in some dimensions compared to moderately and insufficiently active individuals. In particular, physically more active adolescents scored significantly higher in physical health, than moderately active and insufficiently active as can be expected (p-value <0.001). However, a statistically significant difference was not found in the rest of the KIDSCREEN dimensions between active, moderately or insufficiently active respondents (all estimated p-values are greater than 0.05).

Regarding socioeconomic characteristics, it was found that adolescents with higher SES seem to have higher scores in physical health, parent relations, autonomy, social support and peers’ dimensions. More specifically, adolescents with a high family affluence scored significantly higher in physical health than adolescents with a low FAS score (p-value <0.1). Regarding parent relations and autonomy, adolescents with a high affluence present a higher score than adolescents with medium and low affluence (p-value <0.001). Simultaneously, adolescents with medium affluence have also a higher score in social support and peers’ dimension than adolescents with a lower level of affluence (p-value <0.1).

Regarding their diet, adolescents with a higher AMD also have a higher score in all KIDSCEREEN dimensions apart from social support and peers’ dimension. In particular, adolescents with optimal adherence or medium adherence have a higher score than adolescents with a low adherence in physical health (p-value <0.001), in parent relations and autonomy (p-value <0.001) and school environment (p-value <0.001). Also, adolescents with optimal adherence have a higher score than adolescents with a medium or low adherence in psychological well-being (p-value <0.001). Weight categories seems to insignificantly affect the level of KIDSCREEN dimensions except physical activity in which obese adolescents emerge a lower level of score (p-value <0.05).