Background

Adolescents are vulnerable to malnutrition because of rapid growth and development and their high macro and micronutrient demand [1]. Adolescence is also a window of opportunity for establishing lifelong dietary habits that support nutritional well-being of generations [2]. The burden of malnutrition in developing countries among adolescents is high [3]. In Ethiopia, a significant number of school adolescents are affected by malnutrition [4]. Researches in Addis Ababa showed that 15. 15.9% of school age children are underweight [5]. In Ethiopia, stunting prevalence among school adolescents range from 5–29% [4, 6,7,8]. Risk factors for stunting include mothers’ occupation [7, 8], mothers’ education [7], adolescents age [6, 7] and grade [9], dietary diversity score [6], meal frequency [9], and food insecurity status [9]. In addition, in Ethiopia, most adolescents are dependent on their parents. Therefore, parents’ education and family income have an influence on adolescent nutritional status [10, 11]. Culturally boys had more free time to participate in outdoor physical activities than adolescent girls; which influence their body weight [12].

Studies indicate that thinness prevalence among school adolescent ranges from 4.5–29% [6,7,8, 13, 14]. In low-income countries, adolescents thinness is influenced by their parents education level [7], age of the adolescent [8, 14], family income [14], dietary diversity score [6], frequency of meal [6] and residency [8]. In addition, exposure of school adolescents to unhealthy food consumption [15] around school environment may have effect on their nutritional status. Adolescents’ food choices is based on peer pressure other than nutritional composition and adolescents may perceive either excess weight or thinness as a sign of well-being and attractiveness [16, 17]. The effect is eating disorders during adolescence [18].

Research recommend a need for consistently measured and standardized indicators for dietary diversity [19]. Dietary diversity is one dimension of diet quality. There are different type of dietary quality assessment methods [20, 21]. Global Diet Quality Score (GDQS) is one of the recent methods to capture both under- and overnutrition [22]. The GDQS was designed for non-pregnant, non-lactating women of reproductive age in low- and middle-income countries (LMICs) [23,24,25]. The method was tested for mothers but has not been tested for adolescents [20, 22, 26, 27]. Focusing on adolescents’ populations will benefit present and future generations in terms of health and development.

Dietary diversity practice has an association with underweight for school adolescents [28]. Poor dietary diversity and food variety are directly associated with inadequate micronutrient intake [29]. A review in low- and middle-income countries show that daily consumption of nutritious foods is low for school adolescents. On average, 46% adolescents consume daily meat, 44% fruits, and 37% vegetables [19].

Although there are improvements in nutritional problem in Ethiopia [30], the government is still off-track to achieve nutritional targets at 2025 [31]. This may help to develop nutrition plans for health and development program for this age group among national priorities. Therefore, school adolescent nutritional status assessment is very important. The aim of this study is to assess stunting and thinness in school-attending adolescents in Addis Ababa.

Methods

Design and setting

A cross-sectional study was conducted among school-attending adolescents aged 10–14 years. The study was conducted in Addis Ababa, the capital city of Ethiopia. Administratively the city is divided in to 10 sub-cities. Addis Ababa is the largest and one of the fastest growing cities in the continent. A total of 223 primary government school exists in the city. The survey was conducted from May to June 2019 in 20 primary schools. Schools in subcities such as Addis Ketema, Akaky Kaliti, Arada, Bole, Gullele, Kirkos, Kolfe Keraino, Lideta, Nifas Silk-Lafto, and Yeka were included.

Participants

Sample size was calculated assuming stunting prevalence of 7.2% [32], level of significance 2%, design effect of 1.5, 20 school clusters and 10% non-response rate. A total of 1078 sample size was calculated to get adequate power to answer our objective. While we have taken all 1200 students collected information to the project to acquire adequate precise information from the available data. A multi-stage stratified random sampling procedure was used to select schools and students. First, randomly two government schools were selected from each sub-city. A total of 20 schools were included in the study. From each selected school, randomly 15 students enrolled whose age were 10–14 years old from each of the grades 5–8. A total of 60 students per school were enrolled to the study. An updated student roster was used as a sampling frame to select students. A total of 1,200 school-attending adolescents were included in the study.

Variables measurement

Outcome’s definitions

Stunting and thinness were assessed according to WHO definition for adolescent. Height was measured in meters to the nearest centimetre. Weight was measured using portable standing digital scale. Measurements were transformed into height-for-age z-scores based on sex and age in year using WHO Anthro-Plus software [33]. Height-for-age z-scores were categorized according to WHO stunting cut-offs (Stunted: <− 2SD; Not stunted: ≥ −2 SD) [34]. Thinness was defined as adolescents with BMI-for-age with Z-score < − 2 SD from the median value of WHO’s 2006 reference data [34].

Predictors definitions

In the study the primary exposures of interest were school adolescents Global Diet Quality Score (GDQS). Adolescents were asked food groups consumed in the reference for the past one-week period. Points were assigned for consumption of food groups based three classifications: never or 1/wk, 2–4/wk and 5–7/wk or > 1/day. Points for each food group were then summed to give an overall score. Global Diet Quality Score were recorded based on the past 7 days for 25 GDQS food groups. GDQS was classified in to two GDQS sub-metrics, such as GDQS positive (GDQS+) sub-metric and the GDQS negative (GDQS–) sub-metric diet quality outcomes [22, 26].

Healthy food group

The GDQS + is the total score across the 16 healthy GDQS food groups, with a possible range of 0 to 32. Healthy food groups are; Dark green leafy vegetables, Cruciferous vegetables, Deep orange tubers, Deep orange vegetables, Deep orange fruits, other vegetables, Citrus fruits, other fruits, Legumes, Nuts and seeds, Poultry and game meat, Fish and shellfish, Whole grains, Liquid oils, Low-fat dairy, Eggs. The mean GDQS + expressed the adolescent population healthy diet quality [22].

Unhealthy food group

The GDQS − is the total score across the 7 unhealthy GDQS food groups and the 2 GDQS food groups that are unhealthy when consumed in excessive amounts, with a possible range of 0 to 17. Unhealthy food groups are; White roots and tubers, red meat, refined grains and baked goods, Processed meat, Sugar-sweetened beverages, Sweets and ice cream, High-fat dairy, Juice, and Purchased deep fried foods. The mean GDQS- expressed the adolescent population unhealthy diet quality [22].

Overall GDQS

The overall GDQS is a sum of the points across all 25 GDQS food groups. The GDQS has a possible range of 0 to 49. A GDQS ≥ 23 is associated with a low risk of nutrient inadequacy, scores ≥ 15 and < 23 indicate moderate risk, and scores < 15 indicate high risk [22].

Dietary Diversity Score (DDS) was calculated based on 10-item food indicators. These food items were; starchy staples, pulses, nuts and seeds, dairy, meat, poultry and fish, eggs, dark green leafy vegetables, other vitamin-A rich fruits and vegetables, other vegetables, and other fruits. For each food item a score of “1” was given for groups consumed and “0” for not consumed over the last one week consumed. The total DDS range from 0 to 10. Those who scores ≥ 5 were good Minimum Dietary Diversity for Women (MDD-W) and < 5 were recorded insufficient MDD-W [35].

Data management and statistical analysis

Data were collected using a pretested standard questionnaire by trained data collectors. WHO Anthro-plus was used to convert height and weight measurements to Z-scores for 5–19 years [36]. Questionnaires were programmed using ODK for electronic data collection and translated into local languages (Amharic in Ethiopia). Data collectors and supervisors were trained on research ethics, how to take consents and anthropometric measurements. Regular supervision was conducted during data collection in the field. Height and Weight were measured for each school adolescent. Height was measured to the nearest 0.1 cm and weight to the nearest 0.1 kg using a calibrated SECA weighing scale. Each subject weighed with light clothing and no footwear. All the measurements were taken twice; when necessary, any discrepancies resolved by a third measurement. The mean values were used for data analysis.

Multinomial logistic regression was conducted to investigate factors associated with adolescent nutritional status. Adolescents’ nutritional status outcome was classified as thinness, normal and overweight. Normal weight as the reference category compared to thinness and overweight. Binary logistic regression analysis was conducted to see the association between predictor variable and stunting status of school adolescents. For stunted adolescent were considered as height-for-age Z-score less than − 2 standard deviation of the new WHO 2007 reference population.

First, univariate analysis was conducted to identify associations between individual variables and school-attending adolescent nutritional status (Stunting and thinness). Variables in the univariate analysis with P-Value less than 0.2 were included in the final model to. control confounders. In the final model for both stunting and thinness sex, age, mother occupation, father’s education, Global diet Quality Score, Dietary Diversity score were included in the model. Clustering effect was observed using school as a cluster effect. P-value < 0.05 was considered as a cut-off point for statistical significance. Statistical analyses were conducted using STATA 14 software.

Results

Socio-demographic characteristics

A total of 1200 school-attending adolescents included in the study with a median age of 13 years with interquartile range of 12 and 14 years. The mean age of participant was 12 years ± 1.16 years standard deviations. A quarter of students 316 (26.38%) had only one sibling. For the majority of school adolescent, the education level of fathers and mothers was at the primary level (Table 1).

Table 1 Socio demographic characteristics of urban school-attending adolescents in Addis Ababa ,2019 Ethiopia,
Full size table

Dietary characteristics and nutritional status

School-attending adolescents frequently (5–7 times a week) consumed wholegrains (95.25%) and liquid oils (86.58%). Poultry, fish, dark green leafy vegetable, cruciferous vegetables, deep orange fruits, deep orange tubers were rarely consumed food groups. Legumes food group were consumed 2–4 times per week by half of the students. Juice was one of the frequently consumed food groups and 94.17% of school-attending adolescents took 5–7 times per week. The second most frequently consumed unhealthy food group was sugar-sweetened beverages, 82.08% of students consumed 5–7 times per week. Except for legumes, other vegetables, low fat dairy, and liquid oils, most of the students consumed healthy food groups one time per week. Only other vegetables, whole grains, and liquid oils were consumed at least 5 days in a week. The students rarely ate red meat and high fat dairy food groups. But most frequently ate sugar sweetened beverages, sweets and ice cream, and juice (Table 2).

The overall mean score GDQS food groups consumed by school-attending adolescents one week period was 19.99 ± 2.81 Standard Deviation (SD). The one-week period mean healthy (GDQS+) sub metrics score across 16 healthy food groups was 8.81 with ± 2.86 SD. The mean score for unhealthy (GDQS-) food groups consumed within one week was 11.17 with ± 1.59 SD.

Table 2 GDQS food groups consumed by school-attending adolescents and their Nutritional status in Addis Ababa
Full size table

The overall stunting prevalence was 8.17% with 95% CI: 6.74–9.85%. From the total stunted cases 61(62.24%) are males and 37(37.76%) female school adolescents.

Factors associated with stunting among school adolescents

Sex of school-attending adolescents were significantly associated with stunting status of school adolescents. Males were 1.95 times more likely to be stunted compared to female adolescents with 95% CI: 1.11,3.39 in the final model controlling other factors (Table 3).

Table 3 Factors to stunting, Urban school-attending adolescents in Addis Ababa, 2019 Ethiopia
Full size table

Factors associated to thinness among school adolescents

Multivariable multinomial logistic regression indicates that gender and grade were statistically significant with thinness status of school adolescents. The associations between consumption of citrus fruits, low fat dairy and white roots and tubers school adolescents’ thinness status were statistically significant. Except white roots and tubers all food groups (Citrus fruits, Low fat dairy) were under healthy diet food group category. Thinness compared with consumption of Low-fat dairy from 2 to 4 times per week compared to non-consumption, the relative risk for thinness to normal would be expected to increase by a factor of 1.66. While consumption of citrus fruits and white roots and tubers 2–4 times per week compared to non-consumption or once per week the risk for thinness decreased by a factor of 0.52 and 0.28 respectively compared to normal school adolescents’ nutritional status (Table 4).

Table 4 Specific food item factors to thinness among Urban school-attending adolescents in Addis Ababa ,2019 Ethiopia, (n = 1200)
Full size table

Discussion

The prevalence of stunting and thinness were 8.17% and 12.67% respectively among school-attending adolescents in Addis Ababa. In this finding there was no association between adolescents stunting and thinness with the overall GDQS. Males and older school-attending adolescents were significantly associated with stunting status. Consumption of citrus fruits and white roots tubers were positively associated to thinness. While consumption of low-fat dairy for 2–4 times per week was positively associated with thinness status of school adolescents.

The prevalence of stunting in this study was low compared to previous studies among in school adolescent [6,7,8]. This low stunting prevalence in our study may be due to this study conducted in urban settings [8] .In urban setting families have an opportunity to get more media exposure compared to rural setting [37]. In addition, reginal variation due to sociodemographic and climatic variation between regions in Ethiopia may be the cause of the difference compared to previous studies among school adolescents [11]. This implies the need of different setting contextual action for the implementation of nutritional intervention program for adolescent population. Boys are more stunted than girls in our finding. Our finding is supported by similar study conducted in Ethiopia [38]. In Ethiopia culturally boys had more free time to participate in outdoor activities compared to adolescent girls; adolescent boys physical activity may influence their body weight [39].

The prevalence of thinness in Our finding is lower than studies conducted in the northern 29% [8] and 14.9% [14] and 15% [7] southern part of Ethiopia. This may be due to different sociodemographic characteristics and adolescents from rural settings which are more likely to be involved in activities which need more energy expenditure [40] and different cultural characteristics in the rural setting [41]. While our finding was higher than compared to previous studies conducted in Ethiopia 8.8% [6], 4.9% [13], and 5.2% [4]. The high prevalence in the study setting needs further investigation.

The male school-attending adolescents were more likely to be stunted than female adolescents; which is consistent to previous studies [13, 42]. In Ethiopia, boys have more free time to participate in outdoor physical activities compared to adolescent girls; which influence their body weight [12]. School adolescents age increases the probability of stunted increased [11], while in this study there is no statistical significance between adolescents age to stunting. This may be due to available difference between boys and girls towards their growth spurt age [43].

The overall GDQS was not statically significant with school adolescents stunting and thinness. In this study the mean overall GDQS food group score was 19.99 while in non-lactating and non-pregnant women in India overall GDQS food group score was 24 [25]. The overall mean difference GDQS food group score value between in this school adolescent group and non-lactating and non-pregnant women may be due to age of the two populations. This indicate that the need for further food quality measurement for adolescents.

Healthy food groups like citrus fruits, deep orange tubers, low fat dairy are an important diet components for humans [44]. In this study indicated that school adolescents’ thinness has association with consumption of citrus fruits, Low fat dairy and white roots and tubers. Almost 80% adolescents did not eat for the past seven days either citrus fruits or low-fat dairy products. In the other way 58% and 94% of adolescents take unhealth food groups like white root tubers and Juice respectively. This is an indication of malpractice for the frequency of consumption of health food group. Low-fat dairy products are reduced or naturally low-fat dairy products (≤ 2% milk fat) [45]. The frequency and consumption of healthy diets like citrus fruits and low-fat dairy products increased the overall GDQS score. While Consumption of unhealth food group like White roots and tubers has decreased the overall GDQS score. This inverse finding may be due to adolescent population difference compared to adult population. In addition, it indicated that the need for well-designed strong design to implement for adolescent population.

Some of the limitations of this study include social desirability bias leads to positively distorting dietary habit of adolescents. The study focused only for government schools and not including private schools in which stunting and thinness are not considered issue [32]. Consumption of certain food groups were considered as a wealthier household status in the community. So, this misconception and attitude of the respondent may increase the Global dietary quality score for school adolescents. Secondly, recall bias may underestimate the type of food consumed leading to lower dietary scores.

Conclusion

This study revealed that the prevalence of thinness and stunting was high among urban school attending adolescents. The overall GDQS food score has no association with stunting and thinness status of school adolescents’ Nutritional interventions need to consider health food consumption. Consider the need for dietary quality, gender focused promotion interventions through effective behaviour change communications.