1 INTRODUCTION

Healthy dietary habits in children and adolescents are formed by the family, school, and community.1 One of the aspects of a healthy diet is the consumption of three meals.2 Many studies have shown that a healthy diet improves the health condition of children and adolescents.3-5 For example, breakfast consumption may enhance the cognitive function, test the grades, and increase school attendance.5 In most countries, children and adolescents consist of a large part of the population. According to the World Health Organization (WHO), in 2003, adolescents were 19% of the entire world's population aged 10–19 years, with 84% of them living in developing countries.6 According to the 2011 Population and Housing Census, 16.34% of Iranian population were 10 adolescents aged 10–19 years.7

Adolescence is a crucial period of life that can affect adulthood, based on the establishment of nutritional habits. In addition, breakfast is considered as one of the main portions of a daily meal, and its consumption is an important healthy lifestyle indicator. However, the period between dinner and breakfast usually is the most extended time without absorption of nutrients and energy; therefore, skipping breakfast can result in metabolic changes and may have a negative effect on cognitive performance.5 Eating breakfast can prevent chronic diseases and may have some beneficial effects on the quality of diet. In contrast, skipping breakfast increases the risk of metabolic syndrome risk and may lead to cardiovascular diseases.8 Breakfast plays a vital role in maintaining the health of children and adolescents.9

Skipping breakfast is reported from several countries, mainly in children and adolescents.10 Some studies showed a correlation between skipping breakfast and occurrence of obesity.11, 12 Furthermore, other studies showed that demographic characteristics including the parental education and household income levels can influence breakfast skipping.13, 14

A systemic review conducted by Szajewska et al. (n = 57,481) showed that eating breakfast is closely connected to a reduced risk of obesity in children and adolescents in Europe.14 Therefore, to improve the health of children and adolescents, it is important to identify the determinants of skipping meal, especially breakfast. Identification of these factors can help to make better intervention programs to promote eating daily breakfast. Several studies showed that environmental conditions can affect the dietary habits of children and adolescents. A previous study reported that skipping breakfast is more common in adolescents in low SES families.15-19 Only a few studies have focused on the connection between skipping meal and socioeconomic position; moreover, these studies were conducted in high-income countries, with limited experience in developing countries. Therefore, the aim of this study was to assess the socioeconomic inequality in meal skipping patterns among children and adolescents.

2 MATERIALS & METHODS 2.1 Study design

The data were accessible through the fifth round of a school-based program entitled Childhood and Adolescence Surveillance and Prevention of Adult Non-Communicable Disease (CASPIAN-V) study in Iran. In this cross-sectional nationwide study, information was obtained from families living in 30 provinces of Iran. Further details on CASPIAN-V studies are available.20

The study protocol was approved by the ethical committee of Alborz University of Medical Sciences (Project number: 194049).

2.2 Study population

In this school-based survey, students were selected via multistage stratified cluster sampling from 30 provinces of Iran in 2015. Stratification was performed in each province according to the residence area (urban/rural) and school grade (elementary/high school). Eligible schools were stratified according to the information bank of the Ministry of Education.

The sampling designed as proportional to size sampling with equal sex ratio and the ratio in urban and rural areas were proportionate to the population of urban and rural students. In each province, equal clusters were used to reach the necessary sample size. In each province, 48 clusters and at each cluster 10 sample units (students and their parents) and a total of 14,880 students and an equal number of their parents were selected from 30 provinces.

2.3 Demographic information and family socioeconomic position

The questionnaire of the World Health Organization-Global School-based Student Health Survey (GSHS) was used to collect the demographic and socioeconomic characteristics. The questionnaire had already been translated into Persian, and validated; it included age, sex, living place (urban/rural), family size, maternal and paternal education (illiterate, a high school diploma or less and academic education), and maternal and paternal occupation (employed or unemployed), family composition (single parent or two parents), family assets, and dietary behaviors. Socioeconomic status (SES) was determined using principle component analysis (PCA) method. Variables including paternal education, occupation, possessing a car, school type (public/private), type of home (private/rented), and having a personal computer at home were summarized under one main component: SES(31). SES was categorized into five quintiles, from the first quintile, as the “lowest SES” group, to the fifth quintile, as the “highest SES” group.

2.4 Meal skipping behaviors

According to the GSHS questionnaire, the consumption pattern of all three main meals was assessed by a single-item question. Students were asked to report “usually how many days in a week eat breakfast, dinner, and launch on weekdays and weekends?” Meal frequency was defined as skippers (eating meals 0–4 days/week) and nonskippers (eating meals 5–7 days/week).

2.5 Statistical analysis

Continuous and categorical variables were presented as mean with standard deviation and frequency with percentage, respectively. Frequency of meal skipping patterns was reported as percentage and the 95% confidence interval (CI). Frequency of skipping meal based on family and socio-demographic characteristics was assessed using the Chi-square test. Univariate and multivariate (adjusted) logistic regression analysis was used to assess the association of family and socio-demographic characteristics with meal skipping. All family and socio-demographic characteristics were included in the adjusted model using Enter method. Results of logistic regression analysis were reported as odds ratio and 95% CI. Statistical analysis was performed using SPSS version 16.0.

3 RESULTS

Overall, 14,286 students completed the study (99% participation rate). Table 1 shows the family and socioeconomic characteristics of the study participants. Of 14,286 children and adolescents included in this study, 50.6% were boys, and 71.4% lived in urban areas. The majority of both fathers and mothers earned a high school diploma or less than diploma. Regarding occupation, 90.6% of the fathers were employed, while this number for mothers stood at 12.7%. Also, 93.5% (13,353) of the children and adolescents lived in a two-parent family.

TABLE 1. Family and socioeconomic characteristics of the study participants (n = 14,286) Variable Total Missing n (%) n (%) Age (year) 12.8 (3.1)a 12 (0.1) Sex Boy 7228 (50.6) 12 (0.1) Girl 7046 (49.3) Living area Urban 10,194 (71.4) 12 (0.1) Rural 4080 (28.6) Family size ≤Four 6742 (47.2) 208 (1.5) >Four 7336 (51.4) Maternal educational level Academic education 1524 (10.7) 114 (0.8) Diploma/under diploma 10,148 (71.0) Illiterate 2500 (17.5) Paternal educational level Academic education 1893 (13.3) 496 (3.5) Diploma/under diploma 10,163 (71.1) Illiterate 1734 (12.1) Maternal occupation status Employed 1814 (12.7) 47 (0.3) Unemployed 12,425 (87.0) Paternal occupation status Employed 12,943 (90.6) 96 (0.7) Unemployed 1247 (8.7) Family composition Two parents 13,352 (93.5) 98 (0.7) Single parent 836 (5.9) Family socioeconomic status Low 4562 (31.9) 648 (4.5) Mid 4521 (31.6) High 4555 (31.9) Breakfast skipping No 12,167 (85.2) 154 (1.1) Yes 1965 (13.8) Lunch skipping No 13,173 (92.2) 150 (1.0) Yes 963 (6.7) Dinner skipping No 13,098 (91.7) 120 (0.8) Yes 1068 (7.5) a Are presented as mean (SD).

The frequencies of skipping breakfast, lunch, and dinner were 13.8% (95% CI: 13.3–14.5), 6.8% (95% CI: 6.4–7.2), and 7.5% (95% CI: 7.1–7.9), respectively.

Skipping breakfast behavior was more prevalent among girls than boys (14.7% vs. 13.1%, p-value = 0.006), but no significant difference was observed between them in terms of lunch and dinner (both p-value > 0.05). Compared to students with unemployed mothers, a higher percentage of students with employed mothers reported not eating breakfast (16.4% vs. 13.5%), lunch (8.7% vs. 6.5%), and dinner (9.4% vs. 7.3%) (all p-value < 0.05). Family composition and family SES were significant elements in breakfast skipping; breakfast skipping was more prevalent among the students living in a single-parent family than those living in a two-parent family (20.0% vs. 13.5%, p-value < 0.001; Table 2).

TABLE 2. Frequency of skipping meals according to family and socio-demographic characteristics in Iranian children and adolescents: the CASPIAN–V study Variable Breakfast skipping p value Lunch skipping p value Dinner skipping p value n (%) n (%) n (%) Sex Boy 939 (13.1) 0.006 476 (6.6) 0.437 513 (7.2) 0.080 Girl 1026 (14.7) 487 (7.0) 555 (7.9) Living area Urban 1426 (14.1) 0.189 690 (6.8) 0.838 784 (7.7) 0.141 Rural 539 (13.3) 273 (6.7) 284 (7.0) Family size ≤Four 901 (13.5) 0.251 439 (6.6) 0.369 497 (7.4) 0.586 >Four 1033 (14.2) 506 (7.0) 557 (7.7) Maternal education level Academic 221 (14.7) 0.194 92 (6.1) 0.084 145 (9.5) 0.005 ≤Diploma 1359 (13.5) 671 (6.7) 722 (7.2) Illiterate 364 (14.7) 191 (7.7) 190 (7.7) Paternal education level Academic 243 (13.0) 0.282 154 (8.2) 0.008 155 (8.2) 0.393 ≤Diploma 1409 (14.0) 651 (6.5) 740 (7.3) Illiterate 254 (14.8) 133 (7.7) 127 (7.4) Maternal occupation status Employed 294 (16.4) 0.001 157 (8.7) 0.001 170 (9.4) <0.001 Unemployed 1663 (13.5) 798 (6.5) 895 (7.3) Paternal occupation status Employed 1760 (13.7) 0.018 889 (6.9) 0.080 974 (7.6) 0.004 Unemployed 200 (16.2) 69 (5.6) 66 (5.4) Family composition Two parents 1786 (13.5) <0.001 891 (6.7) 0.689 967 (7.3) 0.023 Single parent 165 (20.0) 58 (7.1) 78 (9.4) Family socioeconomic status Low 755 (16.7) <0.001 321 (7.1) 0.502 375 (8.3) 0.004 Mid 564 (12.6) 290 (6.5) 292 (6.5) High 568 (12.6) 311 (6.9) 341 (7.5) Note: Values shown are mean (standard deviation) and p-value. Bold values denote statistical significance at the p < 0.05 level.

Also, skipping breakfast and dinner behaviors were more frequent among students in families with low SES levels compared to their counterparts in families with mild and high SES (both p-value < 0.001; Table 2).

As shown in Table 2, no significant difference was observed in meal skipping behaviors between students with respect to the living area (urban or rural), size of the family, and the parents' education level (all p-values > 0.05).

Tables 3–5 show the associations of family and socioeconomic characteristics with skipping breakfast, lunch, and dinner in univariate and multivariate (adjusted) logistic regression models.

TABLE 3. Associations of demographic and socioeconomic characteristics with skipping breakfast in Iranian children and adolescents: the CASPIAN-V study Variables Crude model Adjusted modela OR (95% CI) p value OR (95% CI) p value Age (year) 1.00 (0.99–1.02) 0.345 1.00 (0.98–1.02) 0.554 Sex Boy 0.87 (0.79–0.96) 0.006 0.91 (0.83–1.01) 0.088 Girl Reference - Reference - Living area Urban Reference - Reference - Rural 0.93 (0.83–1.03) 0.189 0.93 (0.83–1.04) 0.214 Family size ≤Four Reference - Reference - >Four 1.05 (0.96–1.16) 0.251 0.95 (0.85–1.07) 0.437 Maternal educational level Academic Reference - Reference - ≤Diploma 0.91 (0.78–1.06) 0.234 0.82 (0.66–1.01) 0.062 Illiterate 1.00 (0.83–1.20) 0.955 0.72 (0.55–0.94) 0.017 Paternal educational level Academic Reference - Reference - ≤Diploma 1.09 (0.94–1.26) 0.217 1.03 (0.86–1.24) 0.692 Illiterate 1.16 (0.96–1.40) 0.117 0.86 (066–1.11) 0.260 Maternal occupation status Employed Reference - Reference - Unemployed 0.80 (0.69–0.91) <0.001 0.81 (0.68–0.96) 0.020 Paternal occupation status Employed Reference - Reference - Unemployed 1.21 (1.03–1.42) 0.018 0.99 (0.81–1.21) 0.974 Family composition Two parents 0.62 (0.52–0.74) <0.001 0.53 (0.42–0.67) <0.001 Single parent Reference - Reference - Family socioeconomic status Low 1.39 (1.23–1.56) <0.001 1.79 (1.50–2.14) <0.001 Mid 1.00 (0.88–1.14) 0.919 1.12 (0.96–1.30) 0.132 High Reference - Reference - Note: Bold values denote statistical significance at the p < 0.05 level. a In the full model, all socioeconomic status variables are entered in the multiple logistic regression analyses. TABLE 4. Associations of demographic and socioeconomic characteristics with skipping lunch in Iranian children and adolescents: the CASPIAN-V study Variables Crude model Adjusted modela OR (95% CI) p value OR (95% CI) p value Age 1.01 (0.97–1.02) 0.405 1.00 (0.98–1.02) 0.454 Sex Boy 0.94 (0.83–1.08) 0.437 0.96 (0.83–1.10) 0.554 Girl Reference - Reference - Living area Urban Reference - Reference - Rural 0.98 (0.85–1.13) 0.838 1.07 (0.91–1.25) 0.381 Family size ≤Four Reference - Reference - >Four 1.06 (0.93–1.21) 0.369 1.00 (0.86–1.17) 0.914 Maternal educational level Academic Reference - Reference - ≤Diploma 1.10 (0.88–1.38) 0.373 1.69 (1.26–2.26) <0.001 Illiterate 1.29 (1.00–1.68) 0.047 1.83 (1.25–2.67) 0.002 Paternal educational level Academic Reference - Reference - ≤Diploma 0.77 (0.64–0.93) 0.007 0.68 (0.54–0.86) 0.001 Illiterate 0.93 (0.73–1.19) 0.608 0.80 (0.57–1.12) 0.205 Maternal occupation status Employed Reference - Reference - Unemployed 0.72 (0.60–0.86) <0.001 0.57 (0.46–0.72) <0.001 Paternal occupation status Employed Reference - Reference - Unemployed 0.79 (0.62–1.02) 0.081 0.65 (0.46–0.90) 0.010 Family composition Two parents 0.94 (0.71–1.24) 0.689 0.99 (0.67–1.44) 0.958 Single parent Reference - Reference - Family socioeconomic status Low 1.03 (0.88–1.22) 0.649 1.15 (0.89–1.49) 0.255 Mid 0.94 (0.79–1.11) 0.475 1.05 (0.86–1.28) 0.614 High Reference - Reference - Note: Bold values denote statistical significance at the p < 0.05 level. a In the full model, all socioeconomic status variables are entered in the multiple logistic regression analyses. TABLE 5. Associations of demographic and socioeconomic characteristics with skipping dinner in Iranian children and adolescents: the CASPIAN-V study Variables Crude model Adjusted modela OR (95% CI) p value OR (95% CI) p value Age (year) 1.05 (1.03–1.08) <0.001 1.06 (1.04–1.08)