The aim of the study was to determine treatment outcomes of SAM and identify the major predictors of treatment outcome. From the total of 423 under-5 children included in the study, 81.3% of children were recovered, 11.1% were died and 7.6% of children were default their treatment. Compared to the sphere project value, the overall recovery and default rate were in acceptable ranges (> 75% and < 15% respectively). However, the death rate was somewhat higher than acceptable death figure (< 10%) even though it was not alarming (> 15%) (21). Compared to other studies in different countries, the recovery rate in this study was higher than studies done in Indonesia, Zambia, Bahir Dar city, Ayder hospital, Gedeo Zone-health, Wolaita Zone, Debre Markos and Finote Selam hospitals Dilchora Dawa and North Shoa Zone (10, 11, 20, 22–27)
Differences in sample size, socio-demography and health care setup might be possible reasons for difference of recovery rate in Indonesia, Zambia and Ayder referral hospital. However, the low recovery rate observed in Felege-Hiot referral hospital Gedeo Zone, Wolaita Zone, Debre Markos and Finote Selam hospital, Dilchora hospital and North Shoa Zone might be due to the health care setup(these studies include HC, primary hospitals and general hospitals with no TFU).
However, the recovery rate for the present study was lower than studies done India, Wolisso St. Luke catholic hospital and Ghana(28–30). The higher recovery rate in St. Luke catholic hospital might be due to greater sample size (855). Nevertheless, differences in socio-economic status, quality of care provided in each hospital, the health seeking behavior and accessibility of different medications and therapeutic foods to treat SAM might be reasons for higher recovery rates in India and Ghana.
The present study also reported a death rate higher than reported from the study done in India, Gambia, Bahir Dar City, North Shoa, Debre Markos and Finote Selam hospitals, and Wolaita zone(8, 11, 20, 26, 27, 31). The possible reason for lower death rate in India and Gambia might be only case specific mortality will be recorded and there might be differences in socio-economic status, treatment and caring practice. The possible variations for other Ethiopian studies might be in the present study setting, debilitated children referred from different parts of the country and the case becomes complicated in this way. As a result, the death rate might be higher.
Regarding predictors of nutritional recovery time, from all socio-demographic characteristics, age was the only significant factor for treatment outcomes of SAM. For every 1-month increase in child's age, nutritional recovery rate was increased by 14.6% (AOR = 1.146; 95%CI: 1.052–1.249). The scientific explanation for this might be due to discontinuation of breastfeeding and inappropriate complementary feeding practices as children’s age increases. The present study was consistent with the study done in Debre Markos and Finote Selam hospitals, and Northern India (20, 28). However, this study is contrary to studies done in Malawi, Wolisso St. Luke catholic hospital, Gamo-Gofa Zone and Shebedido woreda OTP center (30, 32–34). This difference may be due to differences in research design, and health care setting.
Related to immunization status, those under-5 children who were fully vaccinated for their age were about 4 time more likely to recover than children who were not fully vaccinated according to their age. Scientifically, the body of un-vaccinated children could not fight major childhood diseases like malaria, pneumonia, diarrhea, and measles. The condition of immune- suppression becomes worse when the child is under starvation. As result, the chance of recovery become lower and lower (35, 36). The present study is consistent with a study finding in Bahir Dar (19). Nevertheless, studies done in North Shoa as well as Enderta Woreda, Tigray, did not show any association between vaccination status and nutritional recovery time(26, 33). The reason for variation could be deference in health care setup and sample size.
Among all comorbidities, comorbid diseases like HIV/AIDS, pneumonia, diarrhea and stunting, were the only significant predictors of treatment outcomes of SAM. Depending on children's exposure for different comorbid diseases, children who were HIV positive were about 83% less likely to recover than their counter-parts. It is obvious that HIV/AIDS can compromise the child's immune status and expose children for different opportunistic infection. This may result in devastating outcomes of SAM like prolonged hospitalization and death. The present study is in line with studies done in West Ethiopia, and Finote-selam and Debre-Markos hospitals(20, 37) and contrary to studies done in northwest Ethiopia (14, 38).
Relative to children who did not have pneumonia, children with pneumonia were 66% likely to recover. This could be explained in terms of the synergistic relationship between pneumonia and malnutrition. Children with respiratory infections like pneumonia may present with tachypnea, retractions, and other signs of respiratory distress, but these are undetectable signs in children with SAM. Therefore, health care providers cannot early detect and treat early and result in an evil outcome.
The present study is in line with a retrospective cohort study done in Zambia and Debre Berhan referral hospital, Enat general hospital and Mehal Meda primary hospital(25, 26, 39). However, pneumonia was not a significant predictor of nutritional recovery time in a retrospective cohort study done in Southern Ethiopia, Wolaita Zone and Bahir Dar city(11, 23, 40). The reason for difference might be in those hospitals pneumonia might be detected and treated early compared to the present study setting since it is referral for malnutrition.
Similarly, children who were stunted were 60% less likely to recover compared to children who were stunted. The explanation for this association might be, management of acute malnutrition is similar regardless of whether there is stunting, although obviously the most stunted children will have the highest risk of failing to respond to therapy and die in hospital. Around 25% of under-5 children in the present study were stunted, however none of studies done in Bahir Dar city, Debre Markos and Finote Selam, Debre Berhan referral, Enat general and Mehal Meda primary hospitals(11, 20, 26) included stunting as co-factor.
Likewise, children who experienced diarrhea during hospital stay were about 72% less likely to recover than those who did not experience it. Unless it is prevented and detected early diarrhea could result in many evil complications including metabolic complications like acidosis and alkalosis. As a result the child may end up with death(1). The present study is in line with studies done in Zambia and Ethiopia (24–26, 41). However, diarrhea was not a significant determinant of treatment outcome in studies done in Zambia teaching hospital and in Woldia hospital, Ethiopia (42, 43). The difference with Zambia teaching Hospital might be explained with difference in sample size, but the difference with Woldia Hospital might be, since the present study setting is referral, diarrhea might not be early identified and treated and result in dehydration and death.