According to WHO/UNICEF/IGN, the median urinary iodine concentration in school children is a good indicator of the iodine status in the general population . In this study the median UIC was 96.12 μg /l. Even though this median UIC level was slightly below normal, more than half, 66 (50.8%) of the children had UIC < 100 μg/l. However; we cannot use this value to categorize iodine status of the study population, since UIC value of an individual child cannot be used to categorize iodine status in any population [2, 31, 32].
The median UIC finding of this study (96.12 μg /l) was better than reports from south Ethiopia, Hawassa town (1.9 μg /l)  and Burie and Womberma district of west Gojam in Ethiopia (5 μg/l) . The differences could be related to efforts made by Ethiopian government and other partner’s with regard to salt iodization, distribution, awareness creation of public through trained HEWs which could have increased the demand for iodized salt utilization compared to the periods where other studies were conducted.
Prevalence of goiter is considered as a public health problem if it is equal to or greater than 5%  and endemic if it exceeds 10% . In this study setting, the prevalence of goiter was found to be 54.6%, confirming that it is an endemic and severe public health problem. The finding was in line with the reported prevalence of Burie and Womberma district in Ethiopia (54.5%) , but higher than the reports from northwest Ethiopia, Metekel zone (39.5%)  and Dabat district (29.1%) . It is also higher than the reported prevalence from one of the districts of India (7.75%)  and Nigeria (40.2%) .
The high goiter prevalence in current study setting could be attributed partly to the mountainous and hilly nature of the area and high rainfall in most seasons of the year which might have resulted in leaching away iodine from top soil for long time. The other reason could be related to the time of study. Those studies conducted in other parts of Ethiopia and cited here [11, 26] were done after the implementation of USI program in the country in 2011. Therefore; the program implementation and enforcement might have contributed to the reduction of goiter prevalence in some accessible areas of the country.
As indicated above, there was discrepancy between median UIC (96.12 μg /l, nearly normal) and goiter prevalence (54.6%, high). The observed difference in this study may reflect the fact that goiter is an indicator of chronic ID. That is, some of the cases currently with normal iodine intake may be the old iodine deficient one who already developed it before but it takes long time for the thyroid enlargement to resolve and they are still living with goiter.
When we assess the iodine content of salt sample at household level, almost two-third (63.5%) contains inadequate iodine. This result shows not the absence of iodized salt in the area rather the insufficiency of iodine in the salt. This could be due to loss of iodine under local conditions in the supply- chain management from time of iodization (production level) to time of distribution at consumer level. A study conducted in Ethiopia by Shawel D. et al. revealed 57% reduction of iodine concentration in the salt sample from the production site to the consumers (household level) . However, the aim of this study was not to evaluate the loss of iodine and we invite further research in this area.
Goitrogenic food like cassava, cabbage, kale and sorghum were consumed with different frequency in current study area (Table 2), but only cassava consumption was significantly associated with UIC level in children. The possible explanation for this could be the presence of cyanogenic glycosides in these food groups other than cassava could be low and within the tolerable ranges. Those children who consume cassava were more likely to have low level of UIC than the non-consumer counter parts (P < 0.001). This was in line with similar studies conducted somewhere in Ethiopia [6, 13].
Goiter was more prevalent in female than male students [AOR = 9.20; 95% CI: 5.30, 15.97] and about 2 times higher in children aged 9–12 years as compared to age group 7–9 years [AOR = 2.23; 95% CI: 1.32, 3.77]. Both results were in line with previous reports from Ethiopia (9, 25]. As age increase in school children, the iodine demand increase in both sexes due to physical and physiological changes in the body. Females are more vulnerable than males due to early puberty, inhibitory effect of estrogen on iodine uptake and also its goitrogenic effect by increasing thyroid follicular proliferation .
Children of those families who gain higher income (> 500 Eth. Birr) per month were less likely develop goiter as compared to the lower income groups (< 500 Eth. Birr) [AOR = 0.02; 95% CI: 0.01, 0.04]. The odds of developing goiter among children from households with medium to high wealth status were lower as compared to children from poor (low income) households. This finding was supported by the recent studies done in Assosa town, west Ethiopia and Dabat district, northwest Ethiopia [6, 11]. The lower prevalence of goiter in households with higher wealth index could be related to household’s food purchasing power and food security status. Rich family has better access of both iodized salt and variety food. This in turn improves the child’s dietary diversity and prevents him from goiter.
Children with higher age group (10–12 years) were about 3.6 times more likely to have insufficient UIC level than the lower age groups (7–9 years), [AOR = 3.57; 95% CI: 1.33, 9.62]. As age increases the iodine requirement also increased in this age group to support rapid growth (pubertal growth spurt). Moreover; this age is the period of early adolescence where iodine requirement increased due to physiological and hormonal changes. To compensate this requirement the body utilizes more iodine as supported by several studies [9, 17, 23].
Even though, UIC level was significantly associated with students sex in bivariate analysis, it was not significant during multivariable logistic regression analysis. In previous studies UIC level was significantly associated with sex, Ethiopia  and Nigeria . The discrepancy could be attributed to the small sub-sample size of UIC (n = 130).
Children living in highland area were less likely to have sufficient UIC level than their counter parts in lowland. This could be attributed to the fact that highland area is more mountainous and characterized by high rainfall than the low one. This leads to erosion and leaching away of iodine from top soil. Consequently; crops grown in such area can’t provide sufficient iodine when consumed leading to low level of UIC [2, 13, 20].
There was significant association between cassava consumption and the levels of UIC. Children who consume cassava were more likely to have low level UIC relative to the non-consumer ones [AOR = 44.82, 95% CI, (11.00, 182.61). Here, the wider confidence interval could be a reflection of smaller sub-sample of UIC. Similar result was observed in recent study conducted by Gabriel M, et al., 2014 in Nairobi, Kenya .
It is well known that consumption of cassava (including the frequency and duration of consumption) has been shown as a risk factor for goiter development among children . However; in this study there was no association between cassava consumption and goiter. This could be related to the time of introduction of cassava. It is well known that goiter is the long term consequence of iodine deficiency but cassava was introduced to Dawro zone more recently and propagated to lowland kebeles around Omo River bordering Wolaita zone.
Strength and limitation of the study
Both school and community-based recruitment strategies were used with multiple indicators (UIC, goiter grade, salt iodine content and interview) to measure the prevalence of IDD in the study population. Despite the strength, this study is not free of limitations. One of the limitations was the method used to test the salt samples. We tested salt samples using iodine rapid test kit which shows only the presence or absence of iodine in the salt (qualitative). If Iodometric titration test was done, it would have estimated the iodine content of the salt samples quantitatively. But, we believe this does not affect our finding and conclusions related to the objective of the study, because our primary aim was not to evaluate the iodine content of the salt.
The other limitation was both urine and salt samples were collected from sub-samples, due to resource shortage. However; the sample size of the sub-samples is still fairly sufficient to estimate the prevalence of IDD in the study population. The last limitation to be acknowledged could be the nature of the study design (cross-sectional). In this type of study, it is hard to explain the causal relationship between IDD and its predictors, rather the finding alerts one to conduct further research using more controlled methods to identify causal relationship and to appreciate whether the salt is not iodized at site of production or due to post-production losses at different levels of supply-chain.