The Association of Family Characteristics with Dietary Diversity among Adolescent Girls in Denpasar City, Bali, Indonesia

sum of food groups consumed over the 24-hour period 5 . Food diversity is a key element of good dietary quality, with a diverse diet typically consisting of at least five food groups to obtain nutrient adequacy. External and internal factors influence the diversity of food consumption. The internal factors include age, gender, nutritional knowledge, beliefs,

suggested that poor dietary habits and sedentary lifestyles were found among Indonesian adolescents 3,6,7,8 .
Adolescent girls in Denpasar City tend to consume less diverse food.However, there is no published data on the relationship between family factors and the diversity of food consumption among adolescent girls.Therefore, this study aimed to determine the relationship between family characteristics and food diversity in adolescent girls in Denpasar City.It also identified dietary diversity and associated factors among late adolescent girls in Denpasar City, Bali Province, Indonesia.

METHODS
A community-based cross-sectional study was conducted in Denpasar City, Bali Province, Indonesia, involving 516 late adolescent girls selected through multistage sampling.The first stage involved the selection of 12 villages in 4 sub-districts, while the second stage was to select respondents from each village by simple random sampling.Respondents who fulfilled the inclusion criteria, such as those aged 18-24, residents of Denpasar City, non-pregnant, and without a mental disability, were given informed consent before the interview.Furthermore, respondents were listed with the help of local administrative and village health workers, and only one was selected from each household.
The dependent variable in this study was the dietary diversity score (DDS), calculated based on ten food groups of the 24-hour individual diet recall questionnaire developed by the Food and Agriculture Organization (FAO).The score was coded "1" and "0" for inadequate and adequate dietary diversity, respectively 5 , with a dichotomization of inadequate DDS (<5 groups) and adequate (≥5 groups).Ten food groups were used, including grains, white roots and, tubers, Vit.A-rich Vegetables and fruits, green leafy vegetables, other vegetables, other fruits, meat, poultry and fish, eggs, milk, pulses (beans, peas, and lentils), nuts, and milk products.On the other hand, the independent variables include age, educational and occupational status, family type and size, parents' educational and occupational status, household monthly income, number of family members at risk, receiving social government programs, and food security.A standard semi-structured questionnaire was prepared in Bahasa Indonesia to collect the characteristics data.The data collectors were public health graduates (S.KM) who were provided with a list of randomly selected respondents' identity numbers.Before the interview, data collectors introduced themselves, explained the study's purpose, and obtained written consent from the respondents.
Descriptive statistics were presented using frequency and percentage, while associated and determinant factors were analyzed by logistic regression.Adjusted odds ratios with 95% CI and variables having a p-value less than 0.05 were considered significant.This study was subjected to a thorough review process according to ethical principles.

RESULTS AND DISCUSSION
The mean age of the respondents was 20.6 years (SD±1.7),277 (53.7%) were aged ≥20 years old, 501 (91.7%) graduated from secondary school or above, and 364 (70.55) were employees, such as entrepreneurs, private, and government employees.In terms of family characteristics, 405 (78.5%) mothers and 473 (91.7%) fathers graduated from secondary school or above, 278 (53.9%) mothers were employees and 462 (89,5%) fathers were unemployed.A total of 417 (80.8%) households' income was below the Denpasar City minimum wage (IDR 2,800,000), and 348 (67.4%) were living in a nuclear family with 320 (62.0%) family members ≥ 5 individuals.Furthermore, 304 (58.9%) families did not face any specific risks related to their family members, 461 (89.3%) received government social programs, and 449 (87.0%) were food-secured, as shown in Table 1.The mean DDS was 6.7 (SD±1.7),with a maximum of 8.4, and 11.4% of the respondents had inadequate dietary diversity, as shown in Table 1.The respondents in this study had a more diverse diet than previous investigations conducted in Jakarta, Indonesia 3,7 , Ethiopia 8 , and Uganda 9 .The variation can be attributed to the different number of food groups included in the score, the reference period to calculate DDS, and the study setting.Meanwhile, the mean DDS was higher than the result of the previous studies conducted in Ethiopia, with mean DDS of 4.73±1.186and 5.38±1.75 8,10.
Table 2 showed the dietary diversity characteristic based on food types.More than 50% of respondents consumed grains, meat, milk, poultry, fish, green leafy vegetables, milk products, other vegetables, eggs, vitamin A-rich vegetables and fruits, and other fruits.On the other hand, food groups of pulses, namely beans, peas, lentils, and nuts, were consumed in small proportions at 28.9% and 22.9%, respectively.The consumption of all these food groups was below 12% among respondents who consumed inadequate dietary diversity; specifically, nuts were almost not consumed.Almost all respondents consumed grains (rice), similar to the results of other studies in Indonesia 11 , Uganda 9 , India 12 , and Bangladesh 13 , where grains were primarily produced in the whole country and had a high availability in the market.Anemia can be caused by grains and starchy foods, and high phytate content may inhibit iron absorption.The proportion of consumption pulses and nuts was lower compared to other studies 3,11,9 , such as in Jakarta, where 39.1% and 39.3% of adolescent girls with anemia and obesity, respectively, consumed legumes, nuts and seeds 3 .

Dietary Diversity's Associated Factors
This study analyzed 14 variables; only six were associated with inadequate dietary diversity.They include the mother's educational status, occupational status, family size, family member at risk, household's monthly income, and family that received a government social program (Table 3).These six variables were included in a binary logistic regression model, and the result showed that only the mother's educational status, family size, and household's monthly income were significantly associated with inadequate dietary diversity (Table 4).The odds of having inadequate dietary diversity among adolescents living in a family size of five and above were 6.98 times higher than those less than five (AOR=6.986;95% CI: 2.718-17.957;p-value=0.000).This result is consistent with the reports of studies in South Africa and Ethiopia 14 .A large number of family members will experience an economic shortage to meet the family needs.Consequently, the focus tends to be fulfilling daily needs rather than prioritizing diet quality.
The odds of having inadequate dietary diversity among adolescents living with mothers who did not complete secondary school were 1.90 times higher than those with higher education levels (AOR=1.904;95% CI: 1.029-3.525;p-value=0.04).In this study, the consumption behavior of a diverse diet was decreased by 57% among adolescents whose household income was under the Denpasar City minimum wage (AOR=0.422;95% CI: 0.227-0.784;p-value=0.006)(Table 4).It was also found that mothers' knowledge can improve the family food variety and selection, which affected the eating behavior of family members.Moreover, educated families generally have a better economic ability to afford a quality diet.This result is consistent with the reports of the studies conducted in Iran, Nigeria, northern Ethiopia, and Gurage-Ethiopia 15,16,17,18 .The estimated portion size of food consumed by respondents was not addressed in this study, which is one of its limitations.Additionally, the data on dietary diversity scores were collected through a 24-hour dietary recall method, which might be prone to recall and social desirability bias.

CONCLUSIONS
In conclusion, family characteristics, such as a family size of five or above, a lower mother's education, and a household's monthly income below the district minimum wage, were significantly associated with inadequate dietary diversity among late-adolescent girls.The significant figure of dietary diversity among late adolescent girls emphasized the need for policy attention.Furthermore, socialization on the importance of diverse food consumption to adolescent girls at the family level was a good chance to break the intergenerational malnutrition cycle.
It was declared to have passed ethical clearance from the Research Ethics Commission for Research and Development of the Faculty of Medicine, Udayana University, with number 1356/UN.14.2.2.VII.14/LT/2022,dated May 25, 2022.

Table 1 .
Frequency distribution of respondents and family characteristics ≥20 years old 277 (53.7)Educational status Graduated from junior high school or below 15 (2.9) Graduated from junior high school or below 501 (91.7)Occupation

Table 2 .
Frequency distribution of dietary diversity characteristics based on type of food

Table 3 .
Bivariable analysis of the factors associated with inadequate dietary diversity