The finding from this study revealed that 12.6% (95%CI; 10.3%, 15.2%) of the respondents were thin. This prevalence was lower compared to study in Damota sore (19.5%) [13], Northern Tigray (58.3%) [16], Babile district (21.6%) [17], Aseko district (14.8%) [18], low land areas of southern Ethiopia (27.5%) [19], Tanzania (14%) [20], and West Bengal, India (49%) [21]. Whereas the prevalence was higher compared to the study in Telangana state, India (10.79%) [22].
The current study also found 21.1% stunting prevalence. The finding was comparable with the study in Aseko district (20.2%) [18] and it was higher compared with a study in Somalia (9.7%) [23], a low land area of Southern Ethiopia (8.8%) [19], and Tanzania (18%) [20], but it was lower compared with study in Northwest Ethiopia (47.4%) [24], Northern Tigray, and West Bengal, India (54%) [21].
The reasons for the observed undernutrition among adolescent girls participating in this study might be due to their low dietary diversity, low meal frequency, poor economic status, and large family size.
This study shows that the odds of thinness was three times higher among adolescent girls living in households with ≥ 7 family members compared to those living in households with < 7 family members. The finding was supported by a study conducted in low land areas of southern Ethiopia [19]. The odds of stunting was also higher among participants from a large family. This finding was supported by a study in Adwa, North Ethiopia [25]. This might be since having a large family imposes a burden on household food consumption and it affects the availability of an adequate amount of quality food [25].
Adolescent girls who usually eat two or fewer meals in a day had more than three times higher odds of thinness compared to those who had three or more meals. This finding was in line with studies in Adwa Northern Ethiopia, Damot Sore District, and southwest Ethiopia [13, 25, 26]. The possible explanation might be because inadequate dietary intake can contribute to thinness. Adolescence is a period characterized by the fastest growth and increased nutritional requirement. Therefore, inadequate dietary intake during this stage of life can easily lead to thinness [1].
Menstruation status of the adolescent girl was statistically associated with thinness. The odds of thinness was four times higher among participants who did not start menses. The result was in line with the study in Aksum Town, Northern Ethiopia, Aseko district, Eastern Arsi Zone, and Eastern Tigray [18, 27, 28]. This is due to the fact that poor nutritional status (thinness) will make adolescent girls see their menarche late. Also, early commencement in menstruation status is an indicator of good nutritional status [29].
Based on the finding of this study wealth status of the household was among the factors that had a significant association with thinness. The odds of thinness was two and three times higher among participants from poor and medium wealth status households, respectively. The finding was consistent with study findings in Aksum Town, and Nigeria [27, 30]. This might be because the rich wealth status of the household indicates better economic status which in turn enables them to consume a more nutritious diet that can contribute to a better nutritional status.
Individual dietary diversity was among the variable that had a significant association with the nutritional status of adolescents. The odds of thinness was two times higher among participants with inadequate dietary diversity compared to those who had adequate dietary diversity. The finding was consistent with a study in Awash Town, Tanzania, and Northwest Ethiopia [5, 20, 31]. This finding was also true for stunting, where adolescents with inadequate dietary diversity were more likely to be stunted and it is in line with studies in Awash Town, and Tanzania [5, 20]. This is because adolescents with adequate dietary diversity scores will get enough energy and other important nutrients that can support their growth spurt.
The age of the adolescent girls had a statistically significant association with thinness. The odds of thinness among participants aged 10–14 years was one and half times higher compared to those in the age category of 15–19 years. This finding was in line with a study in Awash Town, Afar Region, Ethiopia, a low land area of Southern Ethiopia and Northwest Ethiopia [5, 31, 32]. The possible explanation might be because those adolescents in the early age category are at the greatest gain in height (linear growth) compared to late adolescents. So inadequate nutrient intake during this period will make early adolescents more susceptible to chronic malnutrition.
Decision-making power regarding nutrition services was found to be statistically associated with the nutritional status of adolescent girls. The odds of stunting was higher among participants whose parents make solely decision either by the father or mother regarding accessing nutrition services compared to participants whose parents make a joint decision. The finding was in line with a study in the low land area of Southern, Ethiopia [19]. This might be because maternal involvement in decision-making regarding nutrition services plays an important role in the improvement of nutritional status [33].
In this study, household food security was among the factors that had a statistically significant association with stunting. The odds of stunting was two times higher among adolescent girls who lived in food-insecure households compared to those who lived in food-secure ones. The finding of this study is consistent with studies in the Northwestern part of Ethiopia [24, 31] and a study in Jimma Zone, southwest Ethiopia [34]. This is because household food insecurity is one of the underlying causes of undernutrition. It can negatively affect the dietary intake to meet the increased nutrient requirement of adolescent girls.
Limitations of the study
The nature of the design used in this study does not allow causal inferences. So, this study cannot tell the cause-and-effect relationship. The second limitation is that our results may be subject to recall bias.