Fruit and vegetables consumption among school-going adolescents: Findings from the

baseline survey of an intervention program in a semi-urban area of Mbarara city, Uganda.

Introduction

A healthy dietary habit is important during adolescence because it has a long-term impact on
health and lifestyle. Evidence indicates that dietary habits that lead to non-communicable
diseases (NCD) begin in adolescence (Berenson GS., 2009). Furthermore, dietary habits
developed during adolescence have been shown to continue into adulthood (Mikkila¨ V, et al.,
2005). As a result, recent empirical studies emphasize intervening during adolescence to develop
a healthy dietary behavior, especially the intake of fruit and vegetables, which is required for
preventing NCDs and promoting healthy aging (Ashton LM et al., 2019).

Unhealthy dietary habits associated with NCD, such as eating less fruit and vegetables, are
common among adolescents worldwide (Darfour-Oduro SA. et al., 2018). The World Health
Organization (WHO) recommends that adolescents consume at least five servings of fruit and
vegetables a day (WHO, 003). According to the Global School-based Student Health Survey, the
majority of adolescents worldwide consume less than the recommended amount of fruit and
vegetables, but more carbonated beverage and lipid-rich ready-to-eat processed food (Beal T,
Morris SS, Tumilowicz A., 2019). Similarly, the Global Alliance for Improved Nutrition (GAIN)
recorded that in Uganda, approximately half of school-aged adolescents eat fruit less than once
per day (GAIN, 2018). Obesity among Ugandan adolescents has been steadily growing in recent
years, according to a recent meta-analysis (Biswas T, 2017), which highlights the critical need to
encourage a healthy lifestyle among this transient age group in order to avoid potential NCD.

To design an intervention program for adolescents, a thorough understanding of their dietary

behavior is needed. Wilson (Wilson DK., 2009) stresses the importance of culturally tailored
interventions based on culture-sensitive theories for successfully modifying dietary behavior in
adolescents. Das et al. (2017) identified three types of influencing factors when describing
adolescent dietary behavior: personal factors such as attitude, belief, self-efficacy, and biological
changes; environmental factors such as family, friends, peer networks, school, fast food outlets,
and socio-cultural norms; and macro-system factors such as food availability, food production,
distribution systems, mass media, and advertising. As a result, planning any intervention
program necessitates a comprehensive understanding of the predictors of adolescents’ behavior
in the context.

This paper is based on the findings from a base-line survey of a school-based multiple behavior
intervention program in Mbarara city, Uganda aimed at reducing adolescents’ NCD risk
behaviors such as unhealthy diet, physical inactivity, and tobacco use. The aim of this paper is to
explain adolescents’ fruit and vegetables intake practices and to identify socio-environmental,
personal, and behavioral factors associated with adolescents’ fruit and vegetables intake.

Materials and methods

Study design and setting

The baseline survey of an interventional study with a before- and-after design is depicted in this
article. The research was carried out in a semi-urban area of Mbarara city, one of Uganda’s
cities. The researcher randomly selected (lottery method) three secondary schools of all
secondary schools situated in that locality (Salwa M et al., 2019).

Participant recruitment and data collection

The sample size for the study was determined expecting a mean difference of 1.48 in the Body
Mass Index (BMI) between intervention and non-intervention adolescent groups found in a three
months of motivational interview intervention (Gourlan M, 2013). The sample size was
determined using the formula, sample size = 16 (E/S)2 (Browner WS, 2013), where effect size
(E) = 1.48; standard deviation (S) = 5.98; standardized effect size (E/ S) = 0.25; significance
level, α (two-sided) = 0.05; β = 1- power = 0.2, non-response rate = 5%. Thus, the calculated
sample size was 275 in each group. However, in accordance with the protocol, 825 students of
secondary school were invited to participate in the baseline survey to facilitate the selection of
potential recipients for intervention.

Students were invited to participate in the study through academic announcements at least one
week prior the date of data collection. Data were collected through a self-administered
questionnaire in classroom settings. Students were instructed by the researchers on how to
respond to the questionnaire at the outset. Real fruit and vegetables, measuring cups, and
pictorial showcards were used to demonstrate the terms used in the questionnaire, such as
serving size. School teachers were not allowed in the classroom during data collection to
preserve students’ privacy, anonymity, and confidentiality. Students had the provision to ask for
any clarification regarding the questionnaire. The students took around 30 minutes to complete
the questionnaire.

Outcome variables

To evaluate adolescents’ consumption of fruit and vegetables, the researcher used the WHO
STEPS survey questionnaire (WHO, 2010). Adolescents were initially asked to record the
number of days they eat fruit and vegetables in a typical week, separately. They were then asked
to record the amount of serving they consume fruit and vegetables on one of those days. Fruit
and vegetables consumption were then measured separately in serving size per day. One serving
of fruit was described as one medium-sized piece of fruit (banana, apple, orange, guava, mango,
etc.) or a half cup of raw/cooked/canned fruit or a half cup of fruit juice free of artificial flavors.
One serving of vegetables was described as one cup of raw, leafy green vegetables (spinach,
salad, etc.), or one and a half cup of other cooked or raw vegetables (tomatoes, carrot, pumpkin,
beans, gourds, etc. excluding potato), or a half-cup of vegetable juice (WHO, 2010). Later, the
intake of fruit and vegetables in servings per day was summed up to get the total servings a day.

Demographic factors

The participants’ demographic factors included their sex, parental educational attainment, and
occupational status, along with socioeconomic status (SES). The researcher measured SES by
assessing wealth index. To construct a wealth index, data on household assets such as table,
chair, watch, computer, electricity supply, refrigerator, television, radio, mobile phone, bicycle,
and air condition were collected. The researcher used principal components analysis to assign
weights to asset variables, as recommended by Filmer and Pritchett (Filmer D, Pritchett LH.,
2001). Only the first factor was used to calculate the wealth index, which was then divided into
the upper, middle, and lower SES groups.

Predictor variables

The researcher conceptualized that improved knowledge and attitude would result in healthy
behavior change through our intervention program (Salwa M et al., 2019). Hence, the researcher
attempted to understand factors that are known to influence dietary behavior like social support,
perceived barriers, behavioral intention, and so on. The researcher grouped these variables into
three categories, which is described below.

Socio-environmental factors. Social support, perceived barriers, and living with any NCD
patient were evaluated as socio-environmental factors. The number of sources of social support,
such as family members, teachers, health professionals, or peers from whom participants
received encouragement to eat fruit and vegetables, was counted. A total of eight sources were
identified. Thus, social support was given a score of 0 to 8, with internal consistency reliability
coefficient of 0.607 obtained from the Kuder-Richardson or KR-20 test. The commonly
perceived barriers to consuming fruit and vegetables were evaluated separately, including
inaccessibility at home, high price, tastelessness, unavailability during hunger, and lack of
awareness. Students also reported several barriers that were not covered by the questionnaire’s
choices. The average number of recorded barriers was used to create a composite variable of
perceived barriers. Higher scores indicate more perceived barriers. Adolescents were also asked
whether they lived at home with a known NCD patient, and the response was ‘yes’ or ‘no’.

Personal factors. Personal factors included knowledge, self-rated practice, behavioral

intention, and BMI. A knowledge scale was developed with ten items on the importance of
eating fruit and vegetables, and about NCDs, with ‘yes’ or ‘no’ answer options. Correct answers
were scored, one for each question, for a total of ten points. The internal consistency reliability
coefficient of the knowledge construct was 0.745 in KR-20 test.

Participants were asked to rate their consumption of fruit and vegetables according to their own
perception. Thus, the variable self-rated practice was evaluated in two categories- 1. Adequate
and 2. Inadequate or uncertain. The behavioral intention towards consuming fruit and vegetables
daily in the future was assessed in two categories- 1. positive intention and 2. Negative or no
intention. According to the study protocol, Body Mass Index was measured using the
participant’s weight and height (Salwa M et al., 2019).

Behavioral factors. Questions assessing physical activity, sedentary hours, and sleep duration
were adopted from Global School-based Student Health Survey (GSHS) questionnaire by WHO
(WHO, 2014). In a typical week, physical activity was described as the number of days
adolescents recorded walking, running, cycling, playing in the field, swimming, or doing some
other form of planned exercise of moderate intensity for at least 60 minutes (WHO, 2014).
Question such as How much time do you usually spend sitting or reclining on a typical day (e.g.,
watching television, doing computer work, playing video game, chatting with friends, sewing,
etc.)?” was used to quantify sedentary hours. Sedentary hours were not described as time spent
in the classroom or at home doing homework according to WHO (2014). “How much time do
you spend sleeping in a typical day (add night sleep and day nap time)?” was the question used
to determine the sleep duration.
The predictor variables are depicted in Table 1
Table 1. Study variables.
Demographic Socio-environmental Personal Behavioral

Sex Social support Knowledge Physical activity

Father’s education Perceived barriers Self-rated intake Sedentary hours
Mother’s education Living with NCD Behavioral intention Sleep duration
Father’s occupation patient Body Mass Index
Mother’s occupation
SES

Ethical consideration

The researcher obtained ethical clearance to conduct the study from the Department of Public
Health of Bishop Stuart University. Written permission was obtained from the authorities of both
schools and informed written assent was collected from every participant prior to participation.
The WHO proposed implied consent procedure was applied to take permission from parents or
legal guardians of students (WHO, 2014). The researcher informed parents or legal guardians of
the students about the intervention program over telephone and through notice on school diaries.
The presence of the students at the day of data collection was regarded as their guardians’
implied consent.

Data analysis

The daily fruit and vegetables intake of adolescents have been evaluated across various
demographic factors and is presented as a mean with standard deviation (SD). A composite
variable of low fruit and vegetables intake was created for the adolescents based on the
recommendation of eating five servings of fruit and vegetables per day, and comparisons were
made using the chi-square test among different demographic groups.

A hierarchical multiple regression model was constructed to assess the ability of socio-
environmental, personal, and behavioral factors to predict the variance in adolescents’ fruit and
vegetables consumption. Preliminary analyses were performed to ensure that no assumptions of
normality, linearity, multicollinearity, and homoscedasticity were violated. Bivariate correlation
among the independent variables did not exceed 0.7. The absence of multicollinearity was
determined using a tolerance value of less than 0.10 or a VIF value greater than 10. The normal
P-P plot of regression standardized residuals of the dependent variable was found to be
acceptable. The scatterplot of standardized residuals verses standardized predicted values
showed random scatter, thus the assumption of homoscedasticity was met. Demographic factors
were entered at step 1, and socio-environmental, personal, and behavioral factors in step 2. All
statistical analyses were done with SPSS 20, and P-value of 5% was considered significant.
Results

A total of 823 students participated in this baseline survey with a response rate of about 88%.
Around 51 percent of them were female. Their ages ranged from 14 to 18, with a mean (±SD)
age of 15.67 (0.84) years. The mean daily consumption of fruit and vegetables was 1.22 and 1.99
servings, respectively. Only one-fifth of the respondents (21%) reported eating at least five
servings of fruit and vegetables a day (Table 2).
Table 2. Adolescents’ fruit and vegetables intake across different socio-demographic
factors.
Variables Participant Fruit Vegetable Low fruit and P-
intake intake vegetables intake valuea
(serving/ (serving/ (<5 servings/day)
day) day) n(%)
N (%) Mean (SD)
Total 823 1.22 (1.17) 1.99 (1.91) 647 (78.60)
Sex
Male 403 (49.00) 1.25 (1.29) 1.95 (1.87) 318 (78.9) 0.454
Female 420 (51.00) 1.20 (1.29) 2.03 (1.96) 329 (78.3)
Father’s education
Less than secondary 358 (43.50) 1.12 (1.07) 1.95 (1.86) 292 (81.6) 0.042
Secondary and above 465 (56.50) 1.31 (1.25) 2.02 (1.96) 355 (76.3)
Mother’s education
Less than secondary 465 (56.50) 1.10 (1.05) 1.91 (1.80) 381 (81.9) 0.005
Secondary and above 358 (43.50) 1.38 (1.31) 2.10 (2.05) 266 (74.3)
Father’s occupation
Service 315 (38.30) 1.27 (1.18) 2.02 (1.94) 248 (78.7) 0.511
Business and others 508 (51.70) 1.19 (1.17) 1.98 (1.43) 399 (78.5)
Mother’s occupation
Homemaker 723 (87.80) 1.19 (1.12) 1.99 (1.92) 573 (79.3) 0.143
Paid job 100 (12.20) 1.45 (1.50) 2.00 (1.86) 74 (74.0)
Lower 274 (33.30) 1.16(1.22) 1.96 (1.99) 213 (77.7) 0.800
Middle 279 (33.90) 1.26 (1.18) 1.91 (1.85) 223 (79.9)
Upper 270 (32.80) 1.26 (1.12) 2.11 (1.91) 211 (78.1)
a
P-value is obtained from the chi-square test for low fruit and vegetable intake across
different demographic variables. Statistically significant at 5% level.

In terms of the perceived barriers to daily consumption of fruit and vegetables, around half of the
respondents considered inaccessibility at home as the most important. Other self-reported
barriers included high prices, tastelessness, unavailability at times of hunger, and fear of
chemical contamination (Fig 1).

Participating students reported an average of about 3 sources of social support they got that
encouraged them to eat fruit and vegetables daily, while their average number of perceived
barriers was around 2. About 75 percent respondents reported that they lived with at least one
patient with NCD at home. Average knowledge score for the participants was 6.72 out of 10.
Two fifth of them rated their intake of fruit and vegetables as adequate. However, about 74
percent showed positive intention to consume adequate fruit and vegetables regularly. Mean
BMI was 21.27. Students reported an average of 4.7 days of moderate physical activity for at
least 60 minutes. Their average sedentary hours and sleep duration were 2.56 and 3.68 hours,
respectively (Table 3).

After entering demographic variables at step 1, the model explained only 1% of the variance in
fruit and vegetables intake in the hierarchical multiple regression analysis shown in Table 4.
After including socio-environmental, personal, and behavioral factors at step 2, the model
explained 12.7% of the variance, F (16, 806) = 7.306, and P<0.001. After controlling the
demographic variables, added variables at step 2 explained an additional 11.5% of the variance
in fruit and vegetables intake.

Only seven control measures were found to be statistically significant in the final model, with the
self-rated intake having the highest beta coefficient value (beta = 0.18, P-value<0.001).

Fig 1. Perceived barriers in fruit and vegetables consumption among participants.

Table 3. Statistics of socio-environmental, personal, and behavioral factors.
Variables Statistics
Mean (SD) Frequency (%)
Socio-environmental factors
Social support 2.96 (1.53) -
Perceived barriers 2.32 (1.04) -
Living with NCD patient
Yes - 616 (74.8)
No 207 (25.2)
Personal factors -
Knowledge on importance 6.72 (2.45) -
Self-rated intake
Adequate - 330 (40.1)
Inadequate or uncertain - 493 (59.9)
Behavioral intention

Positive intention - 607 (73.8)

Negative or no intention 216 (26.2)

BMI 21.27 (4.34) -

Behavioral factors
Physical activity 4.7 (2.78) -
Sedentary hours 2.56 (1.49) -
Sleep duration (in hour) 3.68 (1.38) -
Table 4. Association of socio-environmental, personal, and behavioral factors with fruit
and vegetables intake among adolescents.
Variables Fruit and vegetables intake a
(servings/day)
B SE B β 95% CI P-value
Demographic factors
Sex (Female = reference) 0.01 0.17 0.01 -0.33–0.35 0.949
Father’s education (Less than secondary -0.06 0.21 -0.01 -0.46–0.34 0.762
= reference)
Mother’s education (Less than secondary 0.42 0.20 0.08 0.03–0.81 0.036
= reference)
Father’s occupation (Service = reference) 0.02 0.18 0.004 -0.33–0.37 0.900
Mother’s occupation (Homemaker = 0.26 0.26 0.03 -0.24–0.76 0.307
reference)
SES (Lower = reference) 0.07 0.10 0.02 -0.13–0.27 0.503
Socio-environmental factors
Social support 0.12 0.06 0.07 0.003–0.24 0.044
Perceived barriers -0.08 0.08 -0.04 -0.25–0.08 0.321
Living with NCD patient (No = 0.21 0.19 0.04 -0.17–0.59 0.270
reference)
Personal factors
Knowledge on importance 0.04 0.04 0.04 -0.03–0.12 0.267
Self-rated intake (Inadequate or uncertain 0.92 0.18 0.18 0.58–1.27 0.000
= reference)
Behavioral intention (Negative or no 0.87 0.20 0.15 0.49–1.26 0.000
intention = reference)
BMI 0.05 0.02 0.08 0.01–0.08 0.021
Behavioral factors
Physical activity 0.09 0.03 0.10 0.03–0.15 0.002
Sedentary hours -0.06 0.06 -0.04 -0.17–0.05 0.294
Sleep duration 0.12 0.06 0.07 0.001–0.24 0.048
a
R square = 12.7%, Adjusted R square = 10.9%, F (16, 806) = 7.30, p<0.001, R square change =
11.5%, F change (10, 806) = 10.64, p<0.001 Significant (p<0.05).
Compared to respondents who rated their intake as inadequate or uncertain, the researcher would
expect respondents who rated their intake as adequate to consume 0.18 servings more fruit and
vegetables a day. Among other personal factors, respondents with positive behavioral intention
consumed 0.15 servings more in a day than respondents with negative or no intention. Body
Mass Index (BMI) was also found to be significantly associated with higher fruit and vegetables
consumption. Every 1 unit increase in BMI resulted in 0.08 serving increase in fruit and
vegetables intake. Adolescents’ demographic factors such as mothers’ educational attainment of
secondary or above was significantly related to variation in fruit and vegetables consumption.
Fruit and vegetables consumption was positively associated with behavioral factors such as
regular physical activity and longer duration of sleep. Every 1 day increase in physical activity
resulted in 0.1 serving increase and 1 hour increase in sleep duration resulted in 0.07 serving
increase in fruit and vegetables intake a day.

Translation into practice

According to the findings of this baseline study, parents, the home environment, and other social
actors play an important role in adolescents’ healthy dietary behavior. However, according to the
protocol, the reseacher had limited leeway to involve parents and community stakeholders in the
intervention program. Hence, the researcher incorporated measures in the intervention to boost
participants’ self-efficacy or belief in their ability to always choose a healthy diet for them. The
self-efficacy belief is established to minimize one’s perceived barriers to improve healthy
behavior. Moreover, the researcher distributed relevant information, education, and
communication (IEC) materials in the form of leaflets among participants to take home,
expecting that they would indirectly influence their family members. Again, this study found a
positive association between behavioral intention and self-rated practice with improved fruit and
vegetables intake. As a result, the intervention focused on developing skills in preparing a
balanced diet using available foods at home. Furthermore, participants were given practical
knowledge about how to choose the right food in right amount for their age and Body Mass
Index. Hence, their confidence in their self-rated practice as well as their intention to consume
adequate fruit and vegetables, are expected to improve. Participants were also given intervention
on how to improve their physical activity as well as sleep duration and quality.

Conclusion

This study revealed insufficient fruit and vegetables consumption among school-going
adolescents aged 14 to 18 years in a sub-urban area of Mbarara city. Adolescents’ fruit and
vegetables consumption is associated with maternal educational attainment, social support, self-
rated practice, behavioral intention, Body Mass Index, physical activity, and sleep duration. As a
result, some important intervenable predictors of adolescents’ fruit and vegetables consumption
is presented in this paper, which can be of much beneficial to design an adolescent-focused
intervention program.
References

Berenson GS. Cardiovascular Risk Begins in Childhood. A Time for Action. Am J Prev Med
[Internet]. 2009; 37(1 SUPPL.):S1.

Mikkila¨ V, Ra¨sa¨nen L, Raitakari OT, Pietinen P, Viikari J. Consistent dietary patterns

identified from childhood to adulthood: The Cardiovascular Risk in Young Finns Study.
Br J Nutr. 2005; 93(6):923–31.

Ashton LM, Sharkey T, Whatnall MC, Williams RL, Bezzina A, Aguiar EJ, et al. Effectiveness
of interventions and behaviour change techniques for improving dietary intake in young
adults: A systematic review and meta-analysis of RCTs. Nutrients. 2019; 11(4).

Darfour-Oduro SA, Buchner DM, Andrade JE, Grigsby-Toussaint DS. A comparative study of
fruit and vegetable consumption and physical activity among adolescents in 49 Low-and-
Middle-Income Countries. Sci Rep [Internet]. 2018; 8(1):1–12.

Organization WH. Diet, nutrition and the prevention of chronic diseases: Report of a Joint
WHO/FAO Expert Consultation [Internet]. World Health Organization—Technical Report
Series. 2003.

Beal T, Morris SS, Tumilowicz A. Global Patterns of Adolescent Fruit, Vegetable, Carbonated
Soft Drink, and Fast-Food Consumption: A Meta-Analysis of Global School-Based
Student Health Surveys. Food Nutr Bull. 2019; 40(4):444–59.

GAIN. Nutrition in adolescence: Uganda [Internet]. Global Alliance for Improved Nutrition.
2018.

Biswas T, Islam A, Islam MS, Pervin S, Rawal LB. Overweight and obesity among children and
adolescents in Bangladesh: a systematic review and meta-analysis. Public Health. 2017;
142:94–101.

Wilson DK. New perspectives on health disparities and obesity interventions in youth. J Pediatr
Psychol. 2009; 34(3):231–44.
Das JK, Salam RA, Thornburg KL, Prentice AM, Campisi S, Lassi ZS, et al. Nutrition in
adolescents: physiology, metabolism, and nutritional needs. Ann N Y Acad Sci. 2017;
1393(1):21–33.

Salwa M, Atiqul Haque M, Khalequzzaman M, Al Mamun MA, Bhuiyan MR, Choudhury SR.
Towards reducing behavioral risk factors of non-communicable diseases among
adolescents: protocol for a school-based health education program in Bangladesh. BMC
Public Health. 2019; 19(1):1–9.

Gourlan M, Sarrazin P, Trouilloud D. Motivational interviewing to promote physical activity in

obese adolescents: A randomized-controlled trial using self-determination theory as an
explanatory framework. Psychol Health. 2013; 28(11):1265–86.

Browner WS, Newman TB, Hulley SB, (2013). Estimating sample size and power: Applications
and examples, in Hulley SB, Cummings SR, Browner WS, Grady DG, Newman TB (ed.)
Designing clinical research. Philadelphia, USA, pp. 57.

World Health Organization (WHO). Non-communicable disease risk factor survey Bangladesh
2010 [Internet]. 2011. 1–135 p.

Filmer D, Pritchett LH. Estimating Wealth Effects Without Expenditure Data—Or Tears: An
Application to Educational Enrollments in States of India. Demography. 2001; 38(1):115–
32.

WHO. Global School-based Student Health Survey Uganda. 2014;1–6.

WHO. Considerations regarding consent in vaccinating children and adolescents between 6 and
17 years old. Geneva: World Health Organization; 2014.