Predictors of recovery from uncomplicated severe acute malnutrition among children aged 6–59 months in Kandahar, Afghanistan: a multicentre prospective cohort study

Abstract

Background

Severe Acute Malnutrition (SAM) is a significant contributor to child mortality and morbidity, particularly prevalent in low-resource settings such as Afghanistan. Studies have found various factors that influence the recovery from uncomplicated SAM. However, such associations are yet to be studied in the Afghan context, which has been affected by decades of conflicts and instabilities. This study aims to identify the predictors of recovery from uncomplicated SAM in children aged 6–59 months admitted to an outpatient therapeutic feeding program.

Methods

A multicentre prospective cohort study enrolled 713 children with uncomplicated SAM between October 2023 and March 2024. Baseline, follow-up, and end data were collected through structured questionnaires and anthropometric measurements, guided by the national guidelines. Recovery from uncomplicated SAM was defined as achieving a mid-upper arm circumference ≥ 125 mm or weight-for-height Z-score ≥ − 2SD, together with the absence of bilateral pitting oedema for two consecutive visits, within three months of treatment. Multivariable logistic regression analyses were conducted to identify predictors of recovery from uncomplicated SAM. Adjusted odds ratios (AOR) with 95% confidence intervals (CI) were reported. Statistical significance was set at p < 0.05.

Results

A total of (567, 79.5%) children recovered from uncomplicated SAM. Children with literate mothers (AOR 6.03, 95%CI 1.76–20.57, p = 0.004), those from small family size (AOR 4.05, 95%CI 2.56–6.41, p < 0.001), those with access to an improved water source (AOR 4.24, 95%CI 2.63–6.82, p < 0.001), good adherence to routine antibiotics (AOR 4.28, 95%CI 2.49–7.34, p < 0.001) and caregivers who received nutrition counselling (AOR 2.80, 95%CI 1.72–4.58, p < 0.001) were significantly more likely to recover from uncomplicated SAM.

Conclusions

In this study, 79.5% of children recovered from uncomplicated SAM. Factors such as maternal literacy, adherence to routine antibiotics, caregiver nutrition counselling, small family size, and access to improved water sources were linked to better recovery from uncomplicated SAM. These findings highlight the potential importance of educational, behavioural, household, and environmental factors in recovery from uncomplicated SAM and may help inform future research and program planning.

Background

Globally, nearly 12.2 million children are estimated to be affected by severe acute malnutrition (SAM). It remains one of the major causes of mortality and morbidity in under-five children worldwide. The burden of SAM is particularly high in South Asia and sub-Saharan Africa, where it continues to pose a significant public health challenge [1].

In Afghanistan, the number of under-five children suffering from acute malnutrition is one of the highest in the world [3.2 million], among which around 1.7 million children are suffering from SAM [2]. According to the 2024 Integrated Phase Classification report, Kandahar is one of the worst affected provinces in Afghanistan, with 103,386 under-five children suffering from SAM and the situation has been categorized as ‘‘critical’’ [3]. These numbers remain high due to the food insecurity that is driven by widespread poverty, economic fragility, climate change, land degradation, years of drought, and decades of conflict that have severely limited the spread of essential public services and access to nutritious food. Additionally, pervasive gender inequalities have limited women’s economic opportunities, which contributes to household food insecurity [4].

Identification and management of SAM starts with an accurate diagnosis. According to the World Health Organization (WHO) and national Integrated Management of Acute Malnutrition (IMAM) guidelines, the criteria for diagnosing SAM are as follows: a mid-upper arm circumference (MUAC) of < 115 mm, a weight-for-height/length Z-score (WHZ/WLZ) of < − 3SD, or bilateral pitting oedema [5, 6]. Children with SAM who have medical complications (such as 3 + bilateral pitting oedema, poor appetite, weight of <4 kg and general danger signs or serious associated disease) are managed in inpatient therapeutic feeding programs, while children without complications are managed in outpatient therapeutic feeding programs (OTFP) [6].

In the OTFP in Kandahar, children with uncomplicated SAM attend the site on a weekly basis to receive the ready-to-use therapeutic food (RUTF) and a course of routine medications including amoxicillin/cotrimoxazole, mebendazole/albendazole, and the measles vaccine [6]. RUTF is a cornerstone of outpatient management for uncomplicated SAM and has been shown to improve recovery outcomes and weight gain among affected children [7].

An essential part of the recovery from uncomplicated SAM is investigating factors that influence this outcome. Recovery from uncomplicated SAM can be defined as the fulfilment of a MUAC of ≥ 125 mm or a WHZ/WLZ of ≥ − 2SD and absence of bilateral pitting oedema for two consecutive visits, within three months of treatment [5–8].

Research has found that children receiving amoxicillin, deworming or vitamin A can increase the chance of recovery [9]. Conversely, children presenting with comorbidities such as diarrhoea or cough can have a lower likelihood of recovery [10]. Additionally, not having fever, and tuberculosis at admission, having an employed father, being from a high-income family, having educated parents, living in houses with an improved latrine, and being from a smaller sized family can positively influence recovery from uncomplicated SAM [10–12]. However, such associations are yet to be studied in Afghanistan. To our knowledge, this is the first prospective cohort study examining predictors of recovery from uncomplicated SAM in the Afghan context, which has been affected by decades of conflicts and instabilities, worsening food insecurity, limiting access to healthcare services, and disrupting consistent availability of essential therapeutic interventions [13, 14].

It is therefore imperative to generate local data that can inform national health policies, with the aim of improving recovery from uncomplicated SAM in children. The main objective of this study is to examine the potential influence of various factors, including those related to children, parents, household, health and adherence, on the recovery from uncomplicated SAM among children in Kandahar, Afghanistan.

Methods

Objectives

• Primary

  • To assess predictors of recovery from uncomplicated SAM among children aged 6–59 months and admitted to the OTFP.
• Secondary

  • To describe treatment outcomes of uncomplicated SAM among children aged 6–59 months and admitted to the OTFP.

Study design

A multicentre prospective cohort study.

Study setting

Kandahar, located in south, is the second largest city of Afghanistan [15]. There are 13 public health centres functioning in this city. These health centres implement the basic package of health services program which has seven components, including newborn and maternal care, immunization and child health, nutrition (covers treating uncomplicated SAM), treatment and control of communicable diseases, physical rehabilitation and disability services, mental health, and essential drugs provision. All 13 health centres were included in a sampling frame. The centres were listed, and four comprehensive health centres (CHC) were randomly selected as study locations using the lottery method. The selected CHCs included Nazo Ana, Amir Jan, Adam Darmal and Mirza Mohammad, all of them provide outpatient therapeutic feeding services and are engaged in treating children (6–59 months) with uncomplicated SAM.

Study population

Inclusion criteria

Children aged 6–59 months, diagnosed with uncomplicated SAM, had informed consent signed by caregiver and admitted to the OTFP of the selected health centres in Kandahar city between October 2023 and March 2024 were eligible for enrolment in the study.

Exclusion criteria

• Study participants with incomplete data.

• Children who had already been part of the study (repeated admissions).

• Children with a disability for which an anthropometric measurement was unsuitable.

Sample size calculation

Sample size was estimated using the prediction model sample size approach available through the Shiny application developed by the Institute of Tropical Medicine [16]. The calculation considered 17 candidate predictor variables for multivariable logistic regression analysis and an anticipated recovery proportion of 75%, which is the minimum standard for recovery from uncomplicated SAM according to the national guidelines [6]. A 15% allowance was added for potential non-response and incomplete data, which resulted in a target sample size of 800 participants.

Sampling strategy

The number of possible study participants was proportionally allocated between the four health centres based on the average of their SAM admission load from the last three months. In each health centre, the consecutive sampling technique was used, in which subjects meeting the inclusion criteria were enrolled until the required sample size was achieved (Table 1).

Table 1.

Study participants’ allocation to the study sites

Health centre	Number of participants allocated
Nazo Ana comprehensive health centre	281
Amir Jan comprehensive health centre	240
Adam Darmal comprehensive health centre	151
Mirza Mohammad comprehensive health centre	128

During the study, a total of 800 children were approached. However, 51 were excluded because of non-response/no consent from the caregivers and a further 36 were excluded due to incomplete data. Finally, 713 children were included in the final analysis (Fig. 1).

Fig. 1.

Flow chart of sampling among children with uncomplicated SAM. Abbreviations: OTFP: Outpatient therapeutic feeding program, SAM: Severe acute malnutrition

Data collection, tools, methods, quality control

Data were collected using a structured questionnaire, which was mostly adapted from literature [6, 10, 12, 17, 18]. The questionnaire included quantitative data only, prepared in English, then translated into the local language (Pashto) and back to English, to ensure the meaning was not lost in translation. The consistency of the questionnaire was reviewed by a local language expert and the principal investigator.

Prior to the start of the study, the data collectors (nurses by training with experience in nutrition) were trained by the principal investigator on anthropometric measurements (MUAC, weight, height, bilateral pitting oedema), the questionnaire and ethical considerations.

A pilot study was conducted on 40 samples to identify potential issues in the methodology. Based on its results, the questionnaire was adapted to the cultural appropriateness (change in wordings) of the population. The pilot study data were not included in the final analysis.

The questionnaire had three sections as follows:

1. Baseline data included sociodemographic characteristics, baseline anthropometric measurements, routine medications, measles vaccination, and the clinical status of each child at the time of admission to OTFP site.

2. Follow-up data on clinical status of the child, adherence to RUTF and routine medications and nutrition counselling session was collected during second visit.

3. End data included treatment outcomes (recovery, no response to treatment, death, loss to follow up (LTFU)) and anthropometric measurements of each child at the time of exit from OTFP.

Children enrolled to the OTFP attended the site for weekly follow-up visits. During each follow-up visit, routine anthropometric measurements, including MUAC, weight and height/length measurements, were performed. Children were assessed for recovery from uncomplicated SAM throughout the treatment period. Children who for two back to back follow up visits had a MUAC of ≥ 125 mm or a WHZ/WLZ of ≥ − 2SD and absence of bilateral pitting oedema were classified as recovered at the second consecutive visit and were therefore discharged from the OTFP.

Children were declared recovered based on the same anthropometric criterion through which they were admitted to treatment. In cases where a child met multiple admission criteria simultaneously, then MUAC was used as the primary criterion for determining recovery.

The caregivers of the children were interviewed in Pashto for data collection. The anthropometric measurements were carried out by the trained nurses with the use of standard measurement procedures and equipment in line with the Afghan national IMAM guidelines [6]. Formal inter-observer and intra-observer reliability assessments were not performed.

To ensure the precision of the mid-upper arm circumference (MUAC) tapes, a validation process was conducted by comparing their measurements against the measurement of a plastic pipe (10, 20 cm in circumference). The MUAC was measured on the left arm, precisely at the midpoint between the acromion (shoulder tip) and olecranon (elbow tip), and recorded to the nearest 1 mm [6].

Prior to weighing the child, the scale was calibrated using certified weights (2, 5, 10 kg) and reset to zero for each child. Paediatric scale (Electronic) was employed to measure the children’s weight with an accuracy of 0.1 kg. The child was weighed wearing the absolute minimum (light underwear only) in a separate room to allow privacy [6].

For accurate standing height measurement, the standard height measuring board was used. The shoes of the child were taken off. They were then positioned at the centre of the measuring board and their arms were beside them. The ankles and knees of child were pressed alongside the board with their head held straightforward. The child’s head, shoulders, buttocks, and heels were positioned against the measuring board, with the feet placed closely together. To precisely measure the recumbent length, the child was placed lying flat on the length board with the head positioned against the fixed headboard and the body aligned centrally. The legs were gently straightened and the movable footboard was brought into contact with the heels. Both standing height and recumbent length were recorded to the nearest 0.1 cm [6].

The weight-for-height/length Z-score (WHZ/WLZ) was calculated using the WHO child growth standards which are based on the child’s weight, height/length, and sex [6]. The following procedures were followed:

1. In the event that the child was under the age of 2 years or had a height of less than 87 cm and their age was unknown, then recumbent length was measured while the child was lying down and the WLZ was taken from the standard WHO weight-for-length look-up table [6].

2. In case the child was at the age of 2 years or older, or if the child was at least 87 cm tall and their age was unknown, then standing height was measured. If the child was unable to stand, then recumbent length was measured and 0.7 cm was subtracted to obtain the equivalent standing height. The WHZ was obtained using the standard WHO weight-for-height reference table [6].

Nutritional oedema was identified through the presence of bilateral pitting oedema in both feet, induced by the application of gently pressing both feet at the same time with thumbs for three seconds and saying “one thousand ONE, one thousand TWO, one thousand THREE”. In case indentation remained visible on both feet, the child was classified as having (+) oedema. However, if the indentation remained visible on both feet and the lower limbs/hands then it was classified as (++) oedema. Finally, if the indentation remained visible on both feet, the lower limbs/hands and around eyes then it was classified as (+++), which would be a complicated case and excluded from the study [6]. Causes of non-nutritional oedema were excluded by a medical doctor.

Adherence to RUTF/medications was assessed during follow-up visits by requesting caregivers to return empty RUTF sachets and counting the returned sachets. In addition, caregivers were asked whether the child had consumed/taken the prescribed RUTF/medications as recommended and whether any RUTF had been shared with other household members.

During study period, the principal investigator conducted regular supervisory visits to study sites. The collected data were verified by randomly crosschecking with data sources (e.g. caregiver, records). The collected data were cleaned using SPSS to identify outliers, and duplicates.

The operational definitions of key study terms are summarized in Table 2.

Table 2.

Operational definitions

Term	Definition
Admission criteria for OTFP	Children aged 6–59 months were admitted to OTFP sites based on following criteria, a MUAC of < 115 mm, or a WHZ/WLZ of < − 3SD (Marasmus), or bilateral pitting oedema of (+) or (++) without a MUAC of < 115 mm or a WHZ/WLZ of < − 3SD (Kwashiorkor). The child should not have any medical complications [6, 19].
Medical complications	Children presenting with 3 + bilateral pitting oedema, poor appetite, weight <4 kg and general danger signs such as vomiting everything, convulsions, lethargy, unconsciousness or a serious associated disease [6].
OTFP	This program manages SAM cases without medical complications [6, 11].
New case	Child met admission criteria and started treatment in the OTFP for the first time [6].
Returned LTFU	Child was LTFU and returned within two months to continue treatment [6].
Referred in/Transferred in	Child stabilized in inpatient program and referred in to continue treatment in OTFP or child referred in from other health centres to continue the same treatment [6].
Improved sanitation facility	It includes septic tanks, pour-flush to pit latrines, pit latrines with slabs, piped sewer systems, composting toilets, flush toilets, and ventilated pit latrines [11].
Improved water source	It consists tube well or borehole, protected dug well, water piped into the dwelling or yard, protected spring, rainwater or public tap [11].
Handwashing practice	Washing hands before preparing food, before eating meal, before feeding child, after using latrine, and after cleaning house [19].
Literacy	The ability to read and write [20].
Diarrhoea	Child has ≥ 3 watery stool within 24 h [11].
Fever	The child’s body temperature is more than 37·5 °C [11].
Passed appetite test	The child needs to eat at least one-third of RUTF sachet (92 g) within 30 min [6].
Exclusive breastfeeding	Refers to breastfeeding of a child for the first six months of their life without being fed with any other solids and liquids [11].
Current breastfeeding	A child is breastfed in present at least once daily [11].
RUTF	These are pastes or soft food specially formulated in the right amount of nutrients (type 1 & 2) and calorie components for treating children over the age of six months with SAM. They require no preparation or integration with another food or water. Their calorie components and nutrients are similar to F-100, however have less water activity, meaning that bacteria cannot grow in them, so can be taken at home. RUTF come in 500 Kcal sachets weighing 92 g. Children with SAM receive various number of RUTF sachets depending on their body weight [6, 10].
Good adherence	Child eats/takes the prescribed dose of RUTF/medications completely [6].
Routine medication and vaccine	Each child receives a course of Amoxicillin/Cotrimoxazole, mebendazole/albendazole, and a dose of measles vaccine at the time of admission to OTFP site [6, 10].
Treatment outcome/ exit from OTFP	Classified as “recovery” and “non-recovery” from uncomplicated SAM [10].
Recovery	Child with MUAC ≥ 125 mm (if admitted by MUAC), or WHZ/WLZ of ≥ − 2SD (if admitted by WHZ/WLZ), or absence of bilateral pitting oedema (if admitted by oedema) for two consecutive visits, within three months of treatment initiation [6].
Non-recovery	Child with no response to treatment, died or LTFU [6].
No response to treatment	Child remains in treatment for three months and does not recover despite proper assessment and treatment, or, Child referred out due to worsening of condition [6, 8].
Death	Child dies while under treatment for uncomplicated SAM in OTFP site [6, 10].
LTFU	Child is absent for three consecutive visits from OTFP site [6, 8].
Length of stay	Total number of treatment days from admission to exit from OTFP [6].

Abbreviations: SAM Severe acute malnutrition, OTFP Outpatient therapeutic feeding program, MUAC Mid-upper arm circumference, LTFU Loss to follow up, WHZ Weight-for-height Z-score, WLZ Weight-for-length Z-score, RUTF Ready-to-use therapeutic food

Study variables

Outcome variable

• Recovery from uncomplicated SAM: Child who achieved a MUAC of ≥ 125 mm or a WHZ/WLZ of ≥ − 2SD and absence of bilateral pitting oedema for two consecutive visits, within three months of treatment initiation.

• Non-recovery from uncomplicated SAM: Child who did not respond to treatment, died or was LTFU.

Independent variables

• Child related factors (sex, age, weight, height, MUAC, WHZ/WLZ, bilateral pitting oedema, admission criteria, diarrhoea, vomiting, cough, fever).

• Parental factors (employment status, education level, birth interval).

• Household factors (monthly income, family size, access to improved sanitation facility, access to improved water source, handwashing practice).

• Healthcare related factors (walking distance from OTFP site, measles vaccination, breastfeeding, nutrition counselling, waiting time, total length of stay).

• Adherence related factors (RUTF, routine antibiotics, deworming, follow up visits).

Data analysis

Data were checked for completeness before analysis. Records with missing information on outcomes or predictor variables were excluded from the final analysis. Data were analysed using SPSS. Descriptive statistics, such as frequency and percentage, were conducted to summarize and present various characteristics of study participants. Median and Interquartile range were used for data not normally distributed.

Predictors of recovery from uncomplicated SAM were analysed using bivariable and multivariable logistic regression analyses. The outcome or dependent variable was coded as 1 = recovered from uncomplicated SAM and 0 = non-recovered from uncomplicated SAM, therefore, odds ratios greater than one indicate increased odds of recovery. After an extensive literature review [8–12] and understanding of the management of uncomplicated SAM in Afghanistan, we included relevant independent variables to see their potential association with recovery from uncomplicated SAM using bivariable analysis.

This was followed by multivariable logistic regression analysis, but before that potential confounders were identified based on prior literature, and causal reasoning. These variables were then evaluated using the change-in-estimate method, in which the crude and adjusted odds ratios for main independent variables were compared. Each potential confounder was added to crude model one at a time, if the adjusted odds ratio changed by 10% or more, the variable was considered a confounder. In addition to confounders, variables with a p-value of < 0.25 in bivariable analysis, along with variables supported by prior literature [5, 6, 9, 12, 18, 19, 21, 22], were included in the final multivariable logistic regression model.

The final model included the following independent variables; walking distance to health centre, father’s employment, father’s literacy, maternal literacy, household income, family size, family’s access to improved sanitation facility, family’s access to improved water source, admission criteria, presence of comorbidity, breastfeeding, adherence to antibiotics, administration of measles vaccination, caregiver nutrition couselling, and missed follow-ups.

Multicollinearity was checked using the variance inflation factor (< 5) and tolerance test (> 0.1) in a linear regression model. None of the variables exceeded the threshold and were retained in the final logistic regression model. Calibration of the model was assessed using the Hosmer–Lemeshow test. The test showed a p-value of 0.505, which indicated that the model adequately fits the data. The discriminatory power of the model was assessed using the receiver operating characteristic (ROC) curve which resulted in the area under curve (AUC) to be 0.837, demonstrating excellent ability of the model to discriminate between children who recovered from uncomplicated SAM and those who did not.

The odds ratio and 95% confidence interval (OR, 95% CI) were used to measure the degree of association, which were taken from the bivariable and multivariable analyses. Statistical significance was considered at a p-value of < 0.05.

Use of artificial intelligence (AI) tools

ChatGPT was used for brainstorming and quickly scanning literature concerning predictors of recovery from uncomplicated SAM, with a particular focus on research conducted in Afghanistan. This information was then cross-checked by conducting an in-depth search of scholarly databases such as PubMed and Google Scholar.

During the writing of introduction and materials and methods sections of this paper, DeepL was used to improve the language. To ensure the content’s original meaning was not altered and that it was accurate, everything was proofread carefully and corrections were made where necessary.

We confirm AI was used with the necessary care and caution. We have fully disclosed its use and take full responsibility for the content of this paper.

Ethical issues

Ethical consideration

The study was conducted in accordance with local ethical guidelines and in compliance with the Declaration of Helsinki.

Privacy / confidentiality of study participants

Data were collected in a protected space (room), respecting privacy. The dataset (in SPSS) and the hard copies are without participants’ names, or health centre names. However, it includes participants’ gender, age and some medical data so it is a pseudonymized data and cannot be linked to specific participants without having additional information.

Hard copies of the data are stored and locked in a file cabinet, located in the archives of Kandahar University, accessible only by the head of IRB. The stored data will be kept for five years.

Results

Descriptive

In this prospective cohort study, a total of 800 children with uncomplicated SAM were recruited across the four study sites. Of these, (713, 89.1%) had complete data and documented discharge outcomes within three months of treatment initiation and were therefore included in the final study analysis.

In the present study, the proportion of children who were female was found to be greater than half of the sample (402, 56.4%). The median age of our study participants was 9 (IQR 7–11) months. Most of the study participants (483, 67.7%) resided within 30 min of walking distance from the health centre. The caregivers of the children were mostly mothers (597, 83.7%) followed by grandmothers (47, 6.6%), fathers (41, 5.8%) and sisters (28, 3.9%). Approximately two third (474, 66.5%) of the participants’ fathers were employed, yet the majority of them (566, 79.4%) were illiterate. A significant proportion of the mothers (698, 97.9%) were unemployed, with only (123, 17.3%) of them being literate. Furthermore, (550, 77.1%) children were from households with lower monthly incomes (< 150 USD). More than half (399, 56%) of the children belonged to smaller sized families (≤ 7 members). The number of children belonging to families without access to an improved sanitation facility was (429, 60.2%) and those without access to an improved water source was (274, 38.4%) (Table 3).

Table 3.

Socio-demographic and household related characteristics of the study participants

Variables	N	(%)
Gender
Male	311	(43.6)
Female	402	(56.4)
Age at admission (months)
6–23	618	(86.7)
24–59	95	(13.3)
Residence
Urban	623	(87.4)
Rural	90	(12.6)
Walking distance (minutes) to health centre
≤ 30	483	(67.7)
> 30	230	(32.3)
Primary caregiver accompanying the child
Mother	597	(83.7)
Grandmother	47	(6.6)
Father	41	(5.8)
Sister	28	(3.9)
Father’s employment
Unemployed	239	(33.5)
Employed	474	(66.5)
Father’s education
Illiterate	566	(79.4)
Read and write	119	(16.7)
School graduate or higher	28	(3.9)
Maternal employment
Unemployed	698	(97.9)
Employed	15	(2.1)
Maternal education
Illiterate	590	(82.7)
Read and write	111	(15.6)
School graduate or higher	12	(1.7)
Birth interval (years)
≤ 2	477	(66.9)
> 2	236	(33.1)
Monthly household income (USD*)
≥ 150	163	(22.9)
< 150	550	(77.1)
Family size
≤ 7	399	(56.0)
> 7	314	(44.0)
Family’s access to improved sanitation facility
Yes	284	(39.8)
No	429	(60.2)
Family’s access to improved water source
Yes	439	(61.6)
No	274	(38.4)
Handwashing practice in the family
Yes	298	(41.8)
No	415	(58.2)

*1 USD = 65 Afghanis – February 2026 [23]

Majority of the children (690, 96.8%) admitted to the OTFP sites were Marasmic (MUAC and WHZ/WLZ) and only (23, 3.2%) children were suffering from Kwashiorkor (bilateral pitting oedema), with the most frequent admission type being newly admitted children (671, 94.1%). The numbers of children presenting with fever, cough, diarrhoea, and vomiting were (166, 23.3%), (143, 20.1%), (110, 15.4%), and (107, 15%), respectively (Table 4).

Table 4.

Anthropometric and clinical status at admission of the study participants

Variables	N	(%)
Weight (Kg) at admission
3–4.9	239	(33.5)
5–6.9	448	(62.8)
7–10	26	(3.7)
Bilateral pitting oedema at admission
None	690	(96.8)
+	16	(2.2)
++	7	(1.0)
Admission criteria (Malnutrition type)
MUAC (Marasmus)	568	(79.7)
WHZ/WLZ (Marasmus)	122	(17.1)
Oedema (Kwashiorkor)	23	(3.2)
Admission type
New	671	(94.1)
Returned loss to follow up	9	(1.3)
Referred in	33	(4.6)
Presence of fever
Yes	166	(23.3)
No	547	(76.7)
Presence of cough
Yes	143	(20.1)
No	570	(79.9)
Presence of diarrhoea
Yes	110	(15.4)
No	603	(84.6)
Presence of vomiting
Yes	107	(15.0)
No	606	(85.0)
Presence of at least one comorbidity
Yes	331	(46.4)
No	382	(53.6)

Abbreviations: MUAC Mid-upper arm circumference, WHZ Weight-for-height Z-score, WLZ Weight-for-length Z-score

As shown in Table 5, (607, 85.1%) children had good adherence to routine antibiotics. The majority (564, 79.1%) were administered the measles vaccine, however, less than one-third (219, 30.7%) of the children were breastfed during treatment. A large proportion (631, 88.5%) had good adherence to the consumption of RUTF. Most of the caregivers (548, 76.9%) received nutrition counselling. Overall (567, 79.5%) of the study participants recovered, followed by (111, 15.6%) with no response to treatment, (23, 3.2%) loss to follow up and (12, 1.7%) with death outcome. The median length of stay in the program was 49 (IQR 42–54) days (Table 5).

Table 5.

Adherence to treatment, follow up and outcomes of the study participants

Variables	N	(%)
Adherence to routine antibiotics
Good	607	(85.1)
Poor	106	(14.9)
Adherence to deworming
Good	146	(20.5)
Poor	68	(9.5)
Not applicable	499	(70.0)
Measles vaccination
Yes	564	(79.1)
No	149	(20.9)
Child currently breastfeeding
Yes	219	(30.7)
No	494	(69.3)
Adherence to RUTF
Good	631	(88.5)
Poor	82	(11.5)
Caregiver receiving nutrition counselling
Yes	548	(76.9)
No	165	(23.1)
Waiting time (minutes) at health centre
< 60	638	(89.5)
≥ 60	75	(10.5)
Treatment outcome
Recovered	567	(79.5)
No response to treatment	111	(15.6)
Died	12	(1.7)
Loss to follow up	23	(3.2)
Length of stay (days) in the OTFP
< 42	190	(26.6)
≥ 42	523	(73.4)
Any missed follow-ups during the stay in OTFP
Yes	132	(18.5)
No	581	(81.5)

Abbreviations: RUTF Ready-to-use therapeutic food, OTFP Outpatient therapeutic feeding program

Predictors of recovery from uncomplicated SAM

The results of multivariable logistic regression analysis showed that having a literate mother (AOR 6.03, 95% CI 1.76–20.57, p = 0.004), belonging to a smaller sized family with ≤ 7 members (AOR 4.05, 95% CI 2.56–6.41, p < 0.001), family having access to an improved water source (AOR 4.24, 95% CI 2.63–6.82, p < 0.001), being adherent to routine antibiotics (AOR 4.28, 95% CI 2.49–7.34, p < 0.001), and caregivers receiving nutrition counselling (AOR 2.80, 95% CI 1.72–4.58, p < 0.001) were independently associated with recovery from uncomplicated SAM (Table 6). A sensitivity analysis which excluded children who died or were lost to follow-up produced findings consistent with the primary analysis. This supported the robustness of the results.

Table 6.

Results of the bivariable and multivariable logistic regression analyses (n = 713)

Independent Variable	Category	Recovery from uncomplicated SAM				Unadjusted COR (95% CI)	P-value	Adjusted AOR (95% CI)	P-value
		Recovered		Non-recovered
		N	%	N	%
Gender	Male	250	80.4	61	19.6	1.09 (0.76–1.58)	0.616	-	-
	Female	317	78.9	85	21.1
Age (months)	6–23	490	79.3	128	20.7	0.89 (0.51–1.54)	0.692	-	-
	24–59	77	81.1	18	18.9
Residence	Urban	497	79.8	126	20.2	1.12 (0.66–1.92)	0.661	-	-
	Rural	70	77.8	20	22.2
Walking distance (minutes) to the health centre	≤ 30	390	80.7	93	19.3	1.25 (0.85–1.83)	0.241	1.54 (0.97–2.43)	0.064
	> 30	177	77.0	53	23.0
Father’s employment	Employed	385	81.2	89	18.8	1.35 (0.93–1.97)	0.113	0.95 (0.57–1.59)	0.858
	Unemployed	182	76.2	57	23.8
Father’s literacy	Literate	126	85.7	21	14.3	1.70 (1.02–2.81)	0.037	0.99 (0.51–1.93)	0.985
	Illiterate	441	77.9	125	22.1
Maternal employment	Employed	12	80.0	3	20.0	1.03 (0.28–3.70)	0.963	-	-
	Unemployed	555	79.5	143	20.5
Maternal literacy	Literate	120	97.6	3	2.4	12.79 (4.00-40.86)	< 0.001	6.03 (1.76–20.57)	0.004
	Illiterate	447	75.8	143	24.2
Birth interval (years)	≥ 2	191	80.9	45	19.1	1.14 (0.77–1.68)	0.512	-	-
	< 2	376	78.8	101	21.2
Monthly household income (USD)	≥ 150	139	85.3	24	14.7	1.65 (1.02–2.66)	0.038	1.45 (0.76–2.79)	0.257
	< 150	428	77.8	122	22.2
Family size	≤ 7 members	355	89.0	44	11.0	3.88 (2.62–5.74)	< 0.001	4.05 (2.56–6.41)	< 0.001
	> 7 members	212	67.5	102	32.5
Family’s access to improved sanitation facility	Yes	236	83.1	48	16.9	1.45 (0.99–2.13)	0.054	1.44 (0.89–2.34)	0.133
	No	331	77.2	98	22.8
Family’s access to improved water source	Yes	388	88.4	51	11.6	4.03 (2.75–5.92)	< 0.001	4.24 (2.63–6.82)	< 0.001
	No	179	65.3	95	34.7
Handwashing practice in the family	Yes	238	79.9	60	20.1	1.03 (0.71–1.50)	0.848	-	-
	No	329	79.3	86	20.7
Admission criteria	MUAC	461	81.2	107	18.8	1.58 (1.03–2.42)	0.032	1.37 (0.80–2.33)	0.243
	WHZ/Oedema	106	73.1	39	26.9
Presence of a comorbidity	No	320	83.8	62	16.2	1.75 (1.21–2.53)	0.003	0.86 (0.54–1.37)	0.541
	Yes	247	74.6	84	25.4
Child currently breastfeeding	Yes	190	86.8	29	13.2	2.03 (1.30–3.16)	0.001	1.53 (0.91–2.57)	0.105
	No	377	76.3	117	23.7
Adherence to routine antibiotics	Good	509	83.9	98	16.1	4.29 (2.77–6.67)	< 0.001	4.28 (2.49–7.34)	< 0.001
	Poor	58	54.7	48	45.3
Adherence to deworming	Good	127	86.9	19	13.1	1.15 (0.50–2.63)	0.831	-	-
	Poor	58	85.3	10	14.7
Administration of measles vaccination	Yes	455	80.7	109	19.3	1.37 (0.90–2.11)	0.139	0.69 (0.40–1.18)	0.184
	No	112	75.2	37	24.8
Adherence to RUTF	Good	505	80.0	126	20.0	1.29 (0.75–2.22)	0.352	-	-
	Poor	62	75.6	20	24.4
Caregiver receiving nutrition counselling	Yes	459	83.8	89	16.2	2.72 (1.83–4.03)	< 0.001	2.80 (1.72–4.58)	< 0.001
	No	108	65.5	57	34.5
Waiting time (minutes) at health centre	< 60	509	79.8	129	20.2	1.15 (0.65–2.05)	0.619	-	-
	≥ 60	58	77.3	17	22.7
Missed follow-up visits during the stay in OTFP	No	469	80.7	112	19.3	1.45 (0.93–2.25)	0.096	1.23 (0.73–2.06)	0.431
	Yes	98	74.2	34	25.8

Abbreviations: SAM Severe acute malnutrition, COR Crude Odds Ratio, AOR Adjusted Odds Ratio, CI Confidence Interval, WHZ Weight-for-height Z-score, MUAC Mid-upper arm circumference, RUTF Ready-to-use therapeutic food, OTFP Outpatient therapeutic feeding program

Discussion

This is the first prospective cohort study conducted on recovery from uncomplicated SAM in Afghanistan. In this study, 713 children with uncomplicated SAM were enrolled in OTFP, and the overall recovery was 79.5%. Multivariable logistic regression analyses showed that various factors (i.e., maternal literacy, routine antibiotics adherence, nutrition counselling for caregiver, family size, and family’s access to improved water source) were significantly associated with recovery from uncomplicated SAM.

The present study showed that the proportion of children who recovered from uncomplicated SAM met the minimum national (IMAM) standard which recommends an overall recovery of more than 75% from uncomplicated SAM managed in OTFP [6]. This proportion of recovery is in line with findings from various other studies conducted on uncomplicated SAM [8, 9, 24]. However, the recovery proportion in this study is higher compared to the 36.3% in Uganda and the 65% in Ethiopia [12, 18]. The difference from those studies can be explained by the percentage of loss to follow up (56% in the Uganda study and 17% in the Ethiopia study), which are very high compared to the loss to follow up (3.2%) in our study. The high percentage of loss to follow up in those two studies may reflect the long walking distances to reach the health centres and the absence of nutrition counselling [12, 18]. Conversely, in our study more than two-thirds (67.7%) of the participants resided within 30 min of walking distance from the health centres. Research indicates that 30 min walking time to the nearest hospital is generally accepted [25]. Besides this our study participants may have benefited from the provision of nutrition counselling and regular follow up through community health workers. Evidence shows that counselling enhances children’s recovery from acute malnutrition [26].

In this study, children with literate mothers were more likely to recover from uncomplicated SAM compared to those with illiterate mothers. This supports previously published findings from studies, which have shown that literate mothers are more likely to adhere to feeding instructions, recognize danger signs and seek timely care [19, 24]. However, a study conducted in India on predictors of recovery from uncomplicated SAM reported no significant association between maternal education and recovery from uncomplicated SAM [11]. The difference between the two studies may be attributed to contextual variations in literacy levels, healthcare accessibility and socioeconomic conditions. In Afghanistan, female literacy remains substantially lower and access to health education is limited [27], therefore, maternal literacy may play a comparatively greater role in influencing recovery from uncomplicated SAM. These findings are particularly important in the context of Afghanistan, where female literacy levels remain low and access to educational opportunities varies across settings [27]. As caregiving responsibilities for children are predominantly undertaken by women (mothers, sisters, or grandmothers), maternal literacy may substantially influence child health outcomes [28, 29]. The observed association between maternal literacy and recovery suggests that female literacy and health education could be an important consideration for future programs and interventions aimed at improving child health outcomes.

We observed that our study participants with smaller sized family (≤ 7 members) had higher chance of recovery from uncomplicated SAM than their counterparts. This result is consistent with other relevant studies [33, 34]. However, a multicentre study from India did not find family size to be significantly associated with recovery from uncomplicated SAM [11]. This discrepancy may reflect differences in household structure, economic conditions, and cultural caregiving practices between the two settings. In Afghanistan, families usually live in extended and multigenerational houses. Additionally, the country has one of the highest fertility rates in Asia [35]. Moreover, due to the ongoing economic stagnation and limited housing availability, many families are not able to inhabit in separate living spaces which has resulted in large size families. This may contribute to increased risk of infections, lower per-child resource availability, and reduced caregiver attention, all of which may delay a child’s recovery and affect treatment adherence [36–38]. Furthermore, research shows that families sometimes see therapeutic food as something to be shared among family members and when needed, can be sold for collective benefits of the household [39]. These findings highlight the importance of targeted support in nutrition programs for children from larger size families, which could include community-based outreach services such as home visits, additional counselling, or supplementary feeding packages to families with multiple dependents [40, 41].

This study demonstrated that children from families with access to an improved water source are more likely to recover from uncomplicated SAM. This finding is backed by extensive evidence showing that access to an improved water source reduces the burden of diarrhoea, which contributes to malnutrition and poor treatment outcomes in children [34, 42, 43]. This is particularly critical for Afghanistan, where estimations show that one-third of the population don’t have access to safe water [44]. Most of the rural areas and places affected by decades of conflicts depend on water sources that are not protected, such as rivers or open wells, which are often contaminated [45]. Based on the observed link between access to improved water sources and recovery from uncomplicated SAM, integration of basic water, sanitation and hygiene (WASH) components within nutrition programs may warrant further consideration. Potential approaches such as providing water treatment options (e.g., chlorine tablets), hygiene promotion, and safe water storage containers can be implemented alongside therapeutic feeding services [46]. However, intervention studies are needed to determine whether such approaches improve recovery outcomes among children receiving therapeutic feeding services.

Our findings demonstrated that good adherence to routine antibiotics is associated with recovery from uncomplicated SAM. Consistently, extensive evidence has shown that not being adherent to routine antibiotics decreases the likelihood of recovery and that children taking routine antibiotics during treatment have higher chances of recovery from uncomplicated SAM [9–12, 19]. Furthermore, we observed that nutrition counselling for caregivers also played a considerable role in recovery from uncomplicated SAM. These findings highlight that both adherence to routine antibiotics and nutrition counselling for caregivers are linked with recovery from uncomplicated SAM. They are closely related, as both immensely depend on the caregiver’s understanding of treatment procedure and their active participation in the child’s care [26]. Caregivers who receive counselling are more likely to understand the importance of adhering to treatment recommendations [30, 31]. In Afghanistan, where literacy levels are low [32], routine counselling may therefore serve as an essential intervention. These findings represent important considerations such as improving adherence to routine antibiotics and strengthening nutrition counselling in OTFP which may be helpful in improving recovery for children with uncomplicated SAM. However, further research is required to determine whether interventions aimed at strengthening treatment adherence and counselling lead to improved recovery outcomes.

Some of the variables that are commonly regarded as clinically or programmatically important, such as comorbidity status, household income, access to an improved sanitation facility and paternal employment status were not independently associated with recovery after adjusting for potential confounders. In contrast, studies conducted in Ethiopia, Pakistan and India reported significant associations between these factors and recovery from uncomplicated SAM [9–12]. The absence of significant associations in the present study may be attributable to relatively uniform access to OTFP services among participants, limited variation in certain exposures or the influence of unmeasured contextual factors. These findings do not necessarily imply that such factors are unimportant for child nutrition and health, but rather that they did not independently predict recovery from uncomplicated SAM within our study population.

Overall, this study adds important evidence to existing literature by studying the association of several key factors (i.e. child related demographic and clinical, parental, behavioural, environmental and household) with recovery from uncomplicated SAM in a high-burden, conflict-affected, low-resource and under-researched setting. From a clinical and programmatic standpoint, these findings may help inform future research and the development of more targeted and patient-centred interventions that address not only the child’s clinical condition, but also the caregiver’s education, behaviours and household factors. However, further intervention based and longitudinal studies are needed before making broader policy recommendations or establishing causal relationships. In addition, qualitative research is required to explore the underlying barriers and facilitators influencing service uptake, treatment adherence, and caregiver engagement, which are critical for improving the effectiveness of therapeutic feeding programs. Furthermore, interaction effects between certain socioeconomic and program related variables were not explored and may warrant further investigation in future studies.

The study strengths can be summarized as follows; A prospective design was used, with data collected at three main points during the treatment period. This helped in reducing the recall bias, facilitated precise tracking of treatment progress and ensured that potential predictors were measured before recovery outcomes occurred, thereby establishing the temporal sequence between exposures and outcomes. However, causal inference remains limited because of the possibility of residual confounding, selection bias and other unmeasured factors. Furthermore, the anthropometric measurements of children were performed by experienced nutrition nurses using standard procedures in line with national guidelines. This minimized measurement errors, and improved data consistency and reliability. Additionally, the context-specific focus enhances the reliability and applicability of the findings to programmatic decision making in Afghanistan and similar settings.

This cohort study, however, also has some limitations. First, causal relationships between the predictors and recovery from uncomplicated SAM cannot be established due to the observational design of the study. Second, the study was conducted in four OTFP sites within Kandahar City and therefore primarily reflects children with uncomplicated SAM who accessed treatment services in these facilities. Therefore, the findings may not be generalizable to children living in remote, insecure or underserved areas of other regions of Afghanistan with different demographic characteristics or cultural norms. Third, consecutive sampling within selected health facilities may have introduced selection bias and may limit the generalizability of the findings beyond similar outpatient therapeutic feeding settings. Fourth, exclusion of children with incomplete data may have introduced bias and potentially affected the representativeness of the study findings. Fifth, the number of variables included in the multivariable model compared to the number of outcome events may have introduced some risk of overfitting, so cautious interpretation of the model estimates is recommended. Sixth, some estimates, particularly maternal literacy, showed relatively wide confidence intervals, which suggests limited precision and the need for cautious interpretation. Seventh, logistic regression analyses were conducted based on recovery status rather than time to event methods, despite participants being prospectively followed until program exit and therefore the variations in recovery timing were not fully incorporated into the analysis. Furthermore, outcome dates were based on scheduled follow-up visits and could not always be precisely established, particularly for children who died outside health facilities. Future studies with more precise event-time recording may benefit from survival analysis approaches to better examine factors influencing the timing of recovery. Finally, certain potential confounding factors such as maternal nutritional status, household food security, dietary diversity and tuberculosis screening were not assessed and therefore could not be controlled for in the analysis. These factors may have influenced recovery outcomes and should be considered in future studies.

Conclusions

The overall recovery from uncomplicated SAM was 79.5%, which demonstrated the good performance of the OTFP in Afghanistan. Maternal literacy, routine antibiotics adherence, nutrition counselling for caregivers, smaller sized family and access to improved water sources were associated with better recovery from uncomplicated SAM. These findings can be important considerations for program planning aimed at improving recovery from uncomplicated SAM. However, there is need for intervention studies before making specific policy recommendations.

Abbreviations

AOR

Adjusted Odds Ration

AI

Artificial Intelligence

CHC

Comprehensive Health Centres

CI

Confidence Interval

COR

Crude Odds Ratio

IMAM

Integrated Management of Acute Malnutrition

LTFU

Loss To Follow Up

MUAC

Mid Upper Arm Circumference

OTFP

Outpatient Therapeutic Feeding Program

RUTF

Ready-to-Use Therapeutic Food

SAM

Severe Acute Malnutrition

WASH

Water, Sanitation, and Hygiene

WHZ

Weight-for-Height Z-score

WLZ

Weight-for-Length Z-score

WHO

World Health Organization

Authors’ contributions

Conceptualization and design: AHB, GS, AM. Analysis: AHB, GS, AM. Writing- original draft: AHB. Writing- review & editing: GS, AM and NM. All authors have read and approved the final manuscript.

Funding

There was no funding, sponsoring or financial support for this research.

Data availability

The datasets used and/or analysed during the current study are not publicly available due to ethical and confidentiality considerations involving the health related data from children. However, an anonymized dataset may be made available from the corresponding author upon reasonable request and with permission from the relevant institutional authorities.

Declarations

Ethics approval and consent to participate

Ethical approval certificate was given by Research and Ethics Committee of Faculty of Medicine, Kandahar University, Kandahar, Afghanistan (Letter No: 165, Date: 27/September/2023). Approval from the Institute of Tropical Medicine (ITM) Institute Review Board (IRB) was also obtained (IRB code: 1824/24, Date: 11/November/2024).

An informed written consent was taken, prior to data collection, from the caregivers of children included in study. In case the caregiver was illiterate or minor, then a literate and non-minor witness was also added to the consent form process.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.