The study was conducted at Wolaita Sodo University Comprehensive Specialized Hospital (WSUCSH) Medical Center, located 329 km South of Addis Ababa, Wolaita Zone, Wolaita Ethiopia. The former “Otona Hospital” and the current WSUCSH were established in 1920 E.C as small clinics by Sudan Interior Mission (SIM). The hospital delivers different medical services for outpatient, emergency, and inpatients for about 450–500 cases per day. The total service coverage of the hospital is about three million people in catchment areas.

A facility based prospective cohort study was employed to determine predictors of treatment outcomes in under five children with SAM receiving care at WSUCSH Medical Center in the time period of follow-up from mid-May to mid-August 2022.

All under five years of age children patients admitted with the diagnosis of SAM and receiving care (treatment) in the study period were included and followed.

Children whose parents or care takers not willing to participate in the study were excluded from the study.

SAM: WFH < –3 Z score or MUAC < 11.5 cm for 6–59 months of age children.

Based on the number of patient flow in this hospital’s nutritional rehabilitation unit; we intend to study all of the admitted under five SAM children within the study period. Therefore; the implemented sampling technique was a total enumerative sampling method and so, all eligible children admitted in the study period were studied consecutively. In the study period (from mid-May to mid-August 2022) the total number of admitted eligible SAM children was 143 and all of them were enrolled and followed.

The data collection tool was a structured format adapted from reviewing different literature and previous similar studies done on this study title [15]. The prepared format contains five main different categories which include socio-demographic characteristics, history and examination, vital sign and laboratory tests, nutritional diagnosis and other medical diagnoses, and Final treatment outcome.

Before the actual data collection, a pretest was done to ensure the appropriateness of the data collection instrument and necessary modifications were made. Data was collected by three assigned and pre-trained health care professionals (HCP) i.e., by two nurses working at pediatric ward SAM treatment unit and one pharmacist as a supervisor. Data was collected in three main techniques; through direct patient (parent) interviews including socio-demographic and other patient related factors, by assessing the patient’s medical chart, and by direct measurement of patient conditions. We followed pre-designed conceptual framework to develop tools for data collection (Figure 1). The supervisor and the principal investigator supervised the data collection process daily and assured the precision of data collection.

Before analysis started, the completeness and consistency of the collected data were checked. Then, cleaning, editing, and coding of the data were also performed. Data was entered into SPSS version 25.0 for data analysis. The life-table analysis, Kaplan-Meier survival analysis, and log-rank test were used to estimate proportion surviving within each interval, to determine rate of nutritional recovery over time, and to compare time to recovery respectively. Bivariate and then multivariate logistic regression analyses were also used to identify predictors of treatment outcomes. Statistical significance was determined by setting p-values < 0.05 and 95% confidence intervals (CIs).

Prior to the actual data collection process, hands on training was delivered for both data collectors and the mentors concerning the data gathering tool and data gathering process for one day and before the beginning of data gathering, the consistency in the recording of variables in the follow-up forms was checked by taking few patients and amendments were done on the data collection tool after cheek up. The supervisor and principal investigator were assigned for a strict follow-up and supervision throughout the entire data collection period.

Before conducting the study, the study protocol was approved by the Research Ethics and Approval Committee of Wolaita Sodo University College of Health Sciences and Medicine with a project No. CHSM/ERC/06/14. Official letters of co-operation were written to WSUCSH to facilitate the study with letter No.4/467/20/1. Both written and verbal consent were obtained from the parents and study subjects confidentiality was assured as the information they gave will not be used for any purpose other than the study of treatment outcome. The privacy of the participants was maintained as data were anonymized or maintained with confidentiality. We confirm that the study complies with the Declaration of Helsinki.

Out of 157 children admitted for SAM treatment to the hospital during the study period (from mid-May to mid-August 2022), 14 were excluded due to their age of greater than 59 months and the rest children met the inclusion criteria and enrolled. Of the total of 143 children, 79 (55.2%) were males, and most 83 (58%) of them were aged between 6 to 24 months; while 15 (10.5%) of children were younger than six months of age. The mean and median age of the whole cohort was 21 (± 16) and 14 months (6–24) respectively. More than two thirds of the children were from rural areas (n = 100, 69.9%). The majority, 118 (82.5%) of the children were on breast feeding during admission while only 67 (46.9%) of the total children was on exclusive breast feed for the first six months of their age. The mean age for those children’s mothers was 35 (± 6) and the average number of children for the families was 3.4 (± 2) concerning the educational status of mothers; 54 (37.8%) of them attended primary education.

The mean weight of the children at admission was 7.46 (± 2.8) kg while at the time of discharge was 7.65 (± 2.5) kg. The majority of above five months of age children (60.8%) had MUAC below 11.5 cm and about 79.7% of the whole cohort had WFH/WFL less than 70% [–3 standard deviation (SD)] at the time of admission (Table 1).

Concerning co-morbidities on admission; almost all (94%) of the admitted children had at least one medical condition or complication with SAM and the most common co-morbidities accompanied with SAM at admission were: anemia (74.1%), diarrhea (47.6%), and edema (39.9%). About 60.1% of the cohort was diagnosed with non-edematous SAM. Tuberculosis co-infection 2.8%, pneumonia 30.1%, malaria 2.1%, and vomiting 28% were the other medical problems identified on admission.

Almost all of the admitted children, 131 (91.6%) had been given F-75 formula milk as a main nutritional support. In addition, 49.0% and 10.5% of the children received ready-to-use therapeutic food (RUTF) and F-100 formula milk during their admission stay respectively. Concerning the routine medication use; except in one case, all 142 children were provided the routine anti-biotics which were mainly ampicillin and gentamycin among others. Also, from 68 SAM children with diarrhea; 30% of them were given rehydration solution for malnutrition (ReSoMal) as a fluid and electrolyte replacement.

After 1,802 child days of follow-up; 87 (60.8%) recovered or cured from malnutrition state, 47 (32.9%) transferred out for other medical reasons or outpatient treatment program, 6 (4.2%) defaulted and 3 (2.1%) cases were died during the data collection period. The overall incidence rate of recovery was 4.8 cases per 100 child days of follow-up. Based on the treatment outcome: 60.8% of the children were recognized as ‘recovered’ while the rest (39.2%) were called ‘censored’ for the purpose logistic regression analysis model. The overall average weight gain for the children who recovered from wasting was 7.6 g/kg/day. The mean (± SD) length of hospitalization for the cohort was determined and it was 12.6 (± 5.8) days and specifically for recovered children was 13.3 (± 6.7) days.

In the first week of admission, only 8 (9%) of the children recovered while 49 (56.3%) recovered within the second week of admission. The total percentage of recovery within 3 weeks of admission was estimated to be 95.4%. The overall mean recovery time was also estimated and it was 15 days (95% CI, 14.3–17.2).

The overall median recovery time of the children was 14 days (95% CI; 12.9–15.1). The Kaplan-Meier survival analysis indicated that there was a statistically significant difference in the median time to recovery on the variables of sex, residence, and admission MUAC level. There was a difference in median time to recovery with (log-rank = 5.615, p = 0.018) between males (13 days, 95% CI; 11.71–14.28) and females (16 days, 95% CI; 14.53–17.46). Also, those SAM children from urban areas recovered faster than those from rural areas (log-rank = 9.349, p = 0.002) (Table 2 and Figure 2).

A Cox proportional hazard model was tested to see the effect of predictor variables on the survival time and so; twenty-two independent variables were analyzed in the bivariate Cox proportional hazard analysis along with the dependent variable. However, only one variable (residence) was found to be independent predictor of survival time. i.e., SAM children from rural areas were with 0.266 times poor prognosis of survival as compared to SAM children from urban [adjusted hazard ratio (AHR) = 0.26, 95% CI: 0.13–0.55, p < 0.001] (Table 2 and Figure 2).

Independent variables underwent both bivariate and multivariable logistic regression; three of them were found to have significant association with the outcome variable. Also, presence of edema on admission and hypothermia on admission did not show significant association on bivariable analysis but these predictors show significant association with the outcome variable on the final logit model. Children who were hypothermic on admission were 83.1% less likely to recover as compared to those without hypothermia [adjusted odds ratio (AOR = 0.17, 95% CI: 0.03–0.94; p = 0.042)]. Likewise, children with diarrhea on admission were 72% less likely to recover from SAM as compared to those without diarrhea (AOR = 0.28, 95% CI: 0.12–0.66; p = 0.004). Also, children with edematous SAM were found to be less likely to recover from SAM than non-edematous ones (AOR = 0.21, 95% CI 0.07–0.64; p = 0.006). In addition, children who provided RUTF had more chance of recovery (AOR = 2.01, 95% CI: 0.92–4.96; p = 0.098) (Table 3). Other predictor variables like: exclusive breast feed for the first six months of age, time length of breast feed, and presence of cough on admission showed associations with the outcome variable but; that was not statistically significant.

Limitation of the study could be due to use of the total enumerative sampling technique, the representativeness of the sample may be limited and also, even if our study design requires large sample size by its nature, we were limited to small sample size according to our resource limitations, this may impact our study finding generalizability power. Further multicenter and large sample size-based studies should be conducted for detailed investigation of predictors and management strategies to solve severe malnutrition problems in developing countries. Management strategies should address recent advances in role of AI in nutrition management or implementing Internet of Things (IoT) systems for remote monitoring and other technologies [27, 28].

The study highlighted suboptimal recovery rates and average daily weight gain among children treated for SAM. Diarrhea, hypothermia, and edema on admission were associated with lower nutritional recovery rates. These findings underscore the need for targeted interventions to address these factors and improve treatment outcomes in children with SAM.