Time-to-recovery from severe acute malnutrition in children 6–59 months of age enrolled in the outpatient treatment program in Shebedino, Southern Ethiopia: a prospective cohort study

The study was conducted from June to August 2015 in Shebedino district of Sidama zone, Southern Ethiopia. The district is located in the Great Rift Valley area, about 300 kms South of Addis Ababa, the capital of Ethiopia. Shebedino is administratively subdivided into 35 kebeles (32 rural and 3 urban). A kebele is the smallest administrative unit in Ethiopia comprising approximately 1000 households. In 2015, Shebedino had an estimated population of 294,214; of these 14% were infants and children 6–59 months of age.

Shebedino is affected by recurrent and chronic food insecurity. In the district, the average farmland ownership by a household is around 0.5 ha. Crop cultivation and livestock rearing are the major livelihood activities in the rural areas. Maize and Enset (false banana) are the major staple foods.

The district has one primary hospital, nine health centers and thirty two health posts, making the potential health service coverage 98%. According to the health care system of Ethiopia, every kebele is expected to have a health post whereby at least two health extension workers (HEWs) are deployed to provide a package of preventive and essential curative services including the management of uncomplicated SAM in children. HEWs identify SAM cases from their catchment area through multiple modalities including periodical growth monitoring and promotion, enhanced outreach strategy (EOS)/community health day (CHD) campaigns, and static service provided at the health post.

A prospective cohort study was conducted among children aged 6–59 months with uncomplicated SAM enrolled at the OTP sites of the district following a CHD campaign conducted in late May 2015. The cases were followed for the maximum eight weeks through weekly visits starting from June 01, 2015. However, children who recovered earlier were only followed until recovery. Screening of the children and administration of the treatment were made by the frontline health workers according to the national protocol without any direct involvement of the research team.

All children 6–59 months of age who were newly diagnosed with uncomplicated SAM during the CHD campaign and got enrolled in the OTP program were eligible for the study. According to the national protocol, uncomplicated SAM cases are diagnosed as children with good appetite and no major medical complication having MUAC of less than 110 mm and/or first or second degree bilateral pitting oedema [4].

According to the national protocol patients fulfilling the admission criteria are enrolled and given a weekly Plumpy’Nut ration – trade name of a peanut-based ready-to-use therapeutic food (RUTF). Each week, their weight is taken until they achieve a target weight stated in the protocol. On each visit the children are expected to receive a medical assessment and caregivers should be given nutrition education [4]. As the study employed an observational design, the research team was not involved in any aspect of the treatment of the children.

From the total 32 rural health posts found in the district, 25 were selected purposively based on the availability of new SAM cases recruited for OTP during the CHD screening. The total sample size 219 was distributed to health posts proportionally to their newly recruited SAM cases and ultimately the study subjects were selected using quota sampling technique (Fig. 1).

At the end of the CHD camping 219 malnourished children were recruited for the study. Nevertheless, at the first follow-up 3 children were excluded as they were receiving the treatment from health posts found outside Shebedino district. The remaining 216 children were followed for a maximum duration of eight weeks and hence included in the analysis.

Data were gathered by eleven trained enumerators and supervisors using a structured and pretested questionnaire. Baseline data were collected at enrolment and follow-up measurements were made on weekly bases for a maximum of eight weeks.

Socio-demographic and economic variables were gathered at baseline using standard questions extracted from the DHS questionnaire [17]. Dietary Diversity (DD) of the children was assessed at baseline and consecutive weekly follow-up visits by asking the caregivers whether the child had taken from the standard seven food groups recommended by the WHO in the preceding day of the study without setting a minimum intake restriction [18]. The seven food groups were: (i) grains, roots and tubers; (ii) legumes and nuts; (iii) milk and milk products excluding breast milk; (iv) flesh foods; (v) eggs; (vi) vitamin A-rich fruits and vegetables; and (vii) others fruits and vegetables [18]. Household food security was measured at baseline using the Household Food Insecurity Access Scale (HFIAS) by asking about the occurrence and frequency of occurrence of nine food insecurity related events in the preceding four weeks of the survey. Ultimately the food security situation was classified into four ordinal categories: secure, and mild, moderate and severe insecurity [19]. Recent illness history of the child was assessed by asking the caregiver whether the child had fever, cough and diarrhoea in the preceding two weeks of the interview. The questionnaire used for collecting the data is provided as a supporting file with this manuscript (Additional file 1).

Anthropometric measurements – height, weight and MUAC – of the children were taken at baseline and on successive weekly visits using calibrated equipments following standardized procedures. Height and weight were measured without shoes and wearing light clothes using portable stadiometer and Salter spring scales. Height and weight were measured to the nearest 0.1 cm and 100 g, respectively. MUAC was measured at the middle point of the left arm to the nearest 0.1 cm using MUAC tape. Bilateral pitting oedema was assessed by applying normal thumb pressure for 3 s to the both feet.

The dependent variable of the study is time-to-recover from SAM (i.e. the event of interest is recovery and that the response variable is rate of recovery). The independent variables considered are: age and sex of the child, maternal and paternal educational status, level of household food insecurity, household wealth index, distance from the OTP sites, perceived severity of SAM by the caregivers, perceived benefit of SAM treatment, type of malnutrition (Marasmus or Kwashiorkor), dietary diversity and clinical symptoms (diarrhoea, cough and fever). As described in the following conceptual framework, the independent variables were grouped into distal and proximal factors (Fig. 2).

Data were entered, cleaned, and analyzed using SPSS for windows, version 20. Data were described using frequencies, percentages and proper measure of central tendency and dispersion.

During enrollment and follow-ups, dietary diversity scores (DDSs) were determined weekly by summing up the number of unique food groups the child received in the preceding day of the assessment. Ultimately a grand DDS was computed by averaging all the available weekly scores by the number of observations. A grand score of 4 or more was considered as optimal DDS [18].

The treatment outcomes were classified as recovered, non-responder and defaulter in line with the national protocol for the management of SAM [4] and the effectiveness of the program is judged by the Global SPHERE standards [20]. Recovery was defined based on the criteria used to diagnose SAM upon enrollment. For children admitted to OTP based on low MUAC, MUAC greater than 110 mm at two consecutive weeks and/or achieving target weight gain within the maximum stay of 8 weeks in the OTP were used to define recovery. For children admitted based on edema, recovery was resolution of edema at two consecutive weeks. Conversely, children who fail to achieve the aforementioned recovery criteria within the maximum eight weeks treatment were considered as non-responders. Children who missed appointments for two consecutive weeks while being confirmed that they are alive were considered as defaulters.

The time-to-recover from SAM was determined by calculating the differences (in day) from the start of treatment until the child were declared recovered. The average time-to-recover in days was estimated using Kaplan-Meier survival analysis.

Household wealth index was computed using Principal Component Analysis (PCA) as an indicator of household wealth status. A total of fifteen variables related to ownership of selected household assets, size of agricultural land, quantity of livestock, materials used for housing construction, and ownership of improved water and sanitation facilities were considered. Ultimately the generated score was divided into quintiles: poorest, poorer, middle, richer, and richest.

The research protocol was reviewed and approved by the institutional review board (IRB) of College of Medicine and Health Science, Hawassa University. Data were collected after securing informed verbal consent from the caregivers of the children. Verbal consent, instead of written consent, was preferred because most of the study respondents were not literate. The same was approved by the IRB that reviewed the protocol of the study. Confidentiality was maintained while handling participants’ information. Nutrition education was given to the entire caregivers.

Among 216 study subjects enrolled in the study, the boys-to-girls ratio was 1.08 and at enrolment 36.1% were younger than 24 months of age. The mean (±SD) age of the caregivers was 30.1 (±7.0) years and 87.5% were married. More than two-thirds (72.2%) didn’t attend any formal education and about three-fourths (76.4%) were housewives. Nearly two-thirds (65.7%) of the children were sampled from midland areas (1750 to 2300 m above sea level) (Table 1).

Household food insecurity assessment at baseline indicated that all of the households had experienced food insecurity with different degrees of severity in the preceding four weeks of the survey. Nearly half of the respondents (45.4%) had to walk for more than an hour to receive the OTP service from the nearby health post.

At baseline 68.1% of the cases were Marasmic (MUAC< 110 mm) while the remaining 31.9% had Kwashiorkor (presence of bilateral pitting oedema irrespective of anthropometric status). Study participants enrolled in the study with an average weight (±SD) of 8.5 (±2.6) kgs. On admission the vast majority (91.2%) of the children had suboptimal DDS. But nearly half (47.7%) of them were still breastfeeding. Regarding the occurrence of common childhood ailments, 12.5, 42.4 and 45.1% of the children, had cough, diarrhea and fever in the reference two weeks, respectively.

During the first follow-up visit carried out a week after OTP enrolment, the RUTF utilization pattern was assessed. It was found that RUTF sharing (35.2%) and selling (20.8%) practices were not rare. Nearly quarters (24.5%) of the respondents were aware that RUTF is both food and medicine to children with severe malnutrition (Table 2).

From the total study subjects, 79.6% (95% CI: 74.2–85.0%) successfully recovered from SAM within the first eight weeks of treatment. Conversely, nearly one-fifth (20.4%) were censored. Reasons for censoring were: failure to respond to the treatment (11.1%), defaulting from the treatment (3.7%) and transferred out (5.6%).

The median time-to-recovery as determined by the Kaplan-Meier survival analysis, was 5 weeks (95% CI: 4.67–5.33) or 36.0 days (95% CI: 34.3–37.7). The overall mean (±SD) daily weight gain rate was 5.4 (2.6) gm/kg/day for the recovered children.

Predictors of recovery were identified using Cox-proportional hazard model fitted separately to the distal and proximal factors in line with the conceptual framework of the study. In the distal multivariable model, maternal education status, agro-ecological zone of the kebele, household food insecurity status, and distance from the OTP site turned out to be significant predictors of recovery from SAM. Children having caregivers with no formal education had 46% reduced chance of recovery than their counterparts. Children from the highlands showed 43% lower probability of recovery as compared to those from the midlands. Those from severely food insecure household were 53% less likely to recover than cases from mildly food insecure households. Children who reside more than an hour walking distance from the OTP site had 31% reduced chance of recovery than their counterparts (Table 3).

In the proximate model six variables emerged statistically significant. SAM cases who were admitted on the basis of low MUAC were less likely to recover than those admitted based on presence of edema. Children who had diarrhoea at baseline or during follow-up had 37% reduced probability of recovery than their counterparts. The chance of recovery was almost reduced by half among children whose RUTF was shared with other household members. Furthermore, among children whose caregivers were aware that SAM can be fatal, the chance of recovery was nearly two times higher (Table 3).