Socio-Demographic and Diet-Related Factors Associated with Insufficient Fruit and Vegetable Consumption among Adolescent Girls in Rural Communities of Southern Nepal

Sufficient fruit and vegetable (FV) consumption has been associated with reduced risks of chronic diseases and adverse health conditions. However, the determinants of insufficient of FV intake among adolescent girls in Nepal have not been determined. This study was undertaken to identify associations between socio-demographic and diet-related factors with insufficient fruit and vegetable consumption among adolescent girls living in rural communities. This community-based, cross-sectional study was conducted on 407 adolescent girls from rural communities in the Bateshwar rural municipality of Dhanusha district, Southern Nepal between 12 October, 2018 and 14 December, 2018. The study subjects responded to FV consumption and dietary factor-related questionnaires, and anthropometric measurements were taken. Data were analyzed using the univariate logistic regression followed by multivariable logistic regression analyses. Unadjusted and adjusted odds ratios with 95% confidence intervals (CI) are reported. From the 407 study subjects, 359 (88.2%) reported insufficient FV consumption. The factors significantly associated with insufficient FV consumption were education to under the 10th grade, household income in the first tercile, lack of awareness of the importance of FV consumption, the non-availability of FVs at the household level, the low level of dietary diversity, and undernutrition (BMI (body mass index) (<18.5)). The study shows almost 90% of adolescent girls consumed inadequate amounts of FV and that socio-demographic and dietary factors should be taken into account while designing preventive strategies to increase fruit and vegetable consumption to recommended levels.


Introduction
An expert committee of the World Health Organization (WHO) and the Food and Agriculture Organization (FAO) has recommended a dietary intake of at least 400 g of fruits and vegetables (FVs) per day (roughly equivalent to five servings per day) [1]. The presence of antioxidants and rural municipality has a population of 21,530 [34] and the total population of Dhanusha district is 754,777, which includes 84,860 adolescent (10 to 19 years of age) girls [35,36]. The study area is predominantly inhabited by rural communities, and thus, most residents perform agricultural work, although during recent years, foreign company employment has become more popular among skilled and unskilled manual workers [36].

Sample Size and Sampling Procedure
A multistage random sampling procedure was used to select study subjects. First, Bateshwar rural municipality was selected from among 6 rural municipalities in the Dhanusha district using a random number table. Second, two of 5 wards in Bateshwar with approximately equal populations and household numbers were also selected using a random number table. Third, a complete list of households and individuals in both wards was obtained from the office of the Baleshwar rural municipality. Fourth, households were systematically selected and an adolescent girl (10-19 years) was selected from appropriate households for interview. When there were two or more adolescent girls in one household, all were considered. Married adolescent girls and those with a physical or mental health issue were excluded.
OpenEpi software was used to calculate the sample size [37]. This calculation was based on a report that the percentage of Nepalese adolescents that consume the recommended five or more servings of FVs per day was approximately 4% [38]. Assuming a 2.5% marginal error at a 95% confidence level and considering a design effect of 1.5 to account for the intra-cluster effect, 354 adolescent girls were required. Based on a presumed non-response rate of 20%, 425 adolescent girls were invited to participate in the survey. However, 18 did not respond to the invitation, and thus, 407 were included in the study (a response rate of 95.7%).

Data Collection and Anthromometric Measurements
Data were collected during face-to-face interviews by trained researchers using a structured questionnaire adapted from the Nepalese Adolescent Nutrition Survey (2014), the Food Frequency Questionnaire (FFQ), and the dietary diversity questionnaire [38][39][40]; all of which were of already piloted study instruments. Interviews were followed by measuring anthropomorphic variables using calibrated instruments. The survey questionnaires consisted of three parts: (i) Personal profiles (socio-demographic characteristics); (ii) fruit and vegetable consumption, dietary behavior/diet-related factors, and anthropometric measurements; and (iii) the Food Frequency Questionnaire (FFQ). The FFQ was used to assess the frequency and amount of FV consumption and the dietary diversity questionnaire was used to assess dietary diversity [39,40]. Anthropometric measurements, such as weight and height, were obtained using standardized and calibrated study instruments. Body weights were measured to the nearest 100 g using calibrated portable scales with subjects wearing light clothing but without footwear. Heights were measured to the nearest centimeter using a calibrated measuring rod in a full standing position without footwear. Body mass indices (BMIs) were calculated by dividing weight in kilograms by height in meters squared. The questionnaires were translated into the local language (Nepali) and then back-translated into English to ensure translations were accurate.

Definitions of Variables
The main study outcome variable was sufficient intake of FVs and was defined as sufficient if a subject consumed five portions (400 g) or more per day. Based on the available literature, one cup (250 mL) of raw green leafy vegetables or a half cup (125 mL) of cooked or chopped raw vegetables was considered equivalent to one serving of vegetables. Similarly, one portion of fruits was defined as one whole medium-sized fruit (e.g., apple, banana or orange) or half a cup of chopped, cooked, canned fruit, or fruit juice [39]. The study used show cards to collect information for the FFQ on FV intakes.
The independent study variables were: (i) Socio-demographic variables (age, caste/ethnicity, religion, household structure, education, main household occupation, household income); (ii) dietary-related factors; (iii) dietary diversity; and (iv) nutritional status. The subjects were dichotomized as early (10-14 years) or late adolescents (15-19 years). The castes/ethnicities were classified as: (i) Upper caste group-Brahmin, Chhetri, and other relatively-advantaged Terai caste groups (Yadav, Shah, Koiri, Thakur); (ii) Adibasi/Janajati-Janajati or indigenous groups; and (iii) Dalit (relatively disadvantaged) [41]. Religions were classified as Hindu or others (Muslim/ Buddhists, Christians). Household structures were classified as (i) nuclear or (ii) joint or extended. Education was classified as: (i) <10th grade (less than 10 years of schooling) or (ii) ≥10 grade. The main household occupations (parenteral occupations) were categorized as: (i) Agricultural work on farms owned by the household; (ii) services/business work in private companies for the government or their own business; (iii) foreign employment when working in a foreign country; and (iv) others, which included skilled, unskilled, or daily work. Household incomes per month were recorded in Nepali rupees and categorized using terciles: (i) 1st tercile (income <11,243 (Nepalese Rupees (NRs)/month)); (ii) 2nd tercile (11,343 NRs/month); and (iii) 3rd tercile (>22,343 NRs/month). The subject-reported FV consumptions were in accordance with World Health Organization (WHO) recommendations for FV consumption [42]. Initially, subjects were asked about FV requirements per day and responses were categorized as yes (correct response) or no (incorrect response). The availabilities of FVs at the household level were classified as: (i) Yes (if grown on the farm or afforded them); or (ii) no. Dietary habits were classified as vegetarian or non-vegetarian. Meal frequencies were classified as <4 or ≥4 per day.
The consumption of any amount junk food or processed food (such as chips, Kurkure, lays, noodles, salty cookies, biscuits, cakes etc.) at least once a week was classified as yes or no [38]. Addictions to alcohol, cigarettes, gutka, pan masala, betel-quid, khaini or surti, zarda, snuff, and gul were also recorded as yes or no as previously reported [41]. The dietary diversity was classified as low, medium, or high based on food items consumed in 16 food groups as recommended by the FAO. The consumption of food in ≤ 3 food groups was considered as low, 4-5 food groups as medium, and ≥6 food groups as high dietary diversity [40,43]. In order to assess nutritional statuses, Body Mass Indices (BMIs) were calculated using measured heights and weights as described above. The study participants were classified as underweight (BMI < 18.5), normal (BMI between 18.5 and 24.9), overweight (BMI between 25.0 and 29.9), or obese (BMI ≥ 30) [44], but overweight and obese were merged into one category (overweight/obese BMI ≥ 25).

Ethical Considerations
The study protocol was approved beforehand by the Ethical Committee of Janaki Medical College, Tribhuvan University, Nepal (approval number: 16-2075-076). The study purpose and procedure were explained to each participant before data collection was started or before anthropomorphic measurements were taken. All study subjects provided written informed consent. In addition, written parenteral consent was received for each study subject who were below 18 years of age. Privacy and confidentiality were fully maintained throughout and personal identifiers were removed prior to data analysis.

Statistical Analysis
Univariable logistic regression was used to assess associations between FV intake and independent variables of interest, and all factors found to be significant (p < 0.05) were included in the multivariable logistic regression analysis with backward elimination to control for confounders. This study used Hosmer-Lemeshow goodness-of-fit to examine model fitness. The unadjusted and adjusted odds ratios with 95% confidence intervals (CI) are reported. The analysis was performed using the Statistical Package for Social Sciences version 22.0 (SPSS, IBM, Armonk, NY, USA).

Personal Profiles of the Study Subjects
The socio-demographic characteristics are detailed in Table 1. Regarding the 407 study subjects, 359 (88.2%) reported insufficient FV consumption. The majority of the girls (61.4%) were in early adolescence, slightly more than half (56.0%) of them were from the upper caste group, and almost all (93.4%) were Hindus. The majority (69.8%) of the study participants completed <10th grade education, and 60.1% lived in a nuclear family. Nearly half (48.4%) of the households were in the service/business or agriculture occupation and slightly less than half (45.4%) had a household income in the 1st tercile. Subject educations, household types, main parenteral occupations, and household incomes were significantly associated with FV consumption (Table 1).  Table 2 shows the result of the logistic regression analysis with crude odds ratio which demonstrates the associations between awareness, dietary-related factors, and nutritional status and FV consumption. Adolescent girls unaware of FV consumption recommendations, with the non-availability of FVs at the household level, the low and medium level of dietary diversity, and an underweight nutritional status were found to be significantly associated with insufficient FV intake.

Discussion
The current study shows that the majority of adolescent girls living in rural communities in Southern Nepal consume less than the recommended amounts of fruit and vegetables. Only 11.8% of the study subjects were found to consume adequate amounts. This finding was higher than those reported in other Nepalese studies, including the nationwide STEPS survey, which reported almost all adolescent girls failed to meet FV intake recommendations [22,23,25]. This disagreement may be due to the subject age and settings differences. For example, these other Nepalese studies included adults and study settings varied. In addition, our younger study subjects may have had greater access to FVs, and fewer barriers to FV consumption. This could be explained in a way that younger age group subjects in our study setting might have to depend more on household foods than older age groups who might have preference on pre-packaged and fast foods outside leading to decreased FV consumption. On the other hand, our findings are similar to some other studies conducted on undergraduate medical students and on [18][19][20][21][22][23][24][25][26][27][28][29] year olds in Kathmandu [24,26], which suggest age and study setting may influence FV intake. Furthermore, our study participants were adolescent girls, and gender may also have influenced FV consumption as compared with studies conducted on males and females [19]. Moreover, other studies [45,46] reported that, boys did not like to eat FV as much as girls did.
Our observations concur with previous findings [18,[47][48][49] that FV consumption is positively associated with socioeconomic status of study participants and their families. This study found those being educated at under the 10th grade level and those with a household income in the first tercile were at greater risk of inadequate FV consumption, which was entirely expected as these individuals have less access and are less likely to be health conscious than those with higher incomes and levels of education [50]. In addition, those that were unaware of FV consumption recommendations and with limited access FVs at the household level were at higher risk of inadequate FV consumption, which is in line with the findings of previously published papers [18,51]. In fact, socio-economic position, affordability, individual preferences, perception, knowledge, awareness of FV intake recommendations, and home availability/accessibility are among the factors previously reported to importantly determine FV intake [18,[52][53][54][55][56].
Dietary diversity and undernutrition were two other diet-related factors found to be strongly associated the risk of inadequate FV consumption. It is well known that dietary diversity provides the opportunity to consume more types of food items [57,58]. For example, Faber et al. reported that dietary diversity significantly increased the consumption of vitamin A rich FVs [58]. In contrast to a previous study [51], this study found that undernourished subjects were more likely to consume inadequate amounts of FVs, though they probably also ate less than the recommended levels of other essential nutrients.
This study has some strengths that are worthy of mention. First, it is the first study undertaken to identify factors associated with insufficient fruit and vegetable consumption among adolescent girls. Second, the study had a very high response rate and made use of already piloted study instruments [38][39][40]. However, it also has several limitations. First, because of its cross-sectional design, this study could not establish cause and effect relationships. Second, the study had a relatively small sample size, which threatens the generalizability of our study findings to the national level. Third, the self-reporting method used might have introduced bias. Finally, this study did not include adolescent boys, which may have different findings. Further studies are needed to differentiate the gender influence on the factors affecting recommended FVs intake.

Implications of the Study
The findings of this study can serve as a reference to further research in Nepal to explore the reasons for not consuming the recommended FVs among adolescents. In addition, it could be useful to design nutrition intervention programs targeting adolescents. There is a lack of dietary guidelines at a national level to educate and provide counseling to adolescents on dietary requirements. This study indicates the need for dietary guidelines for promoting healthy eating practices.

Conclusions
In the present study, it was found that the majority of adolescent Nepalese girls from a rural background consumed inadequate amounts of FVs. Socioeconomic and diet-related factors, such as schooling to under the 10th grade level, household incomes in the first tercile, the lack of awareness of FV consumption recommendations, the non-availability of FVs at the household level, a low level of dietary diversity, and undernutrition were observed to be significantly associated with inadequate FV consumption. We suggest these factors be taken into account while designing preventive strategies to increase fruit and vegetable intakes. Further, longitudinal studies should be performed to determine risk differences between adolescent girls and boys in rural communities in Nepal. Funding: None. Self-financed.