Study area and period
Institution based descriptive, retrospective chart review was conducted from January 2017 to June 2018. The study was conducted at The Indus Hospital, which is a tertiary care facility in the suburbs of Karachi, Pakistan. Around 1200–1300 malnourished children are seen annually in this hospital, which include a mix of severe acute malnutrition (SAM) and moderate acute malnutrition (MAM). Community based rehabilitation of uncomplicated malnourished children is done in outpatient department of Indus Hospital by counseling and recommendation of fortified home-based staple diet plans with regular follow up.
Sample size and sampling procedure
Sample size determination
The sample size was computed using OpenEpi version 3.01 Statistical software with the following assumptions: Proportion of children attained desired weight using home based fortified diet as 69% , confidence level of 95% with 5% margin of error. The minimum required sample size required for the study was 329 children.
All children with ages ranging from 6 to 59 months with malnutrition, who were treated at Indus nutrition rehabilitation clinic (NRC), from January 2017 to June 2018 were included in the study.
Children who did not have proper records were excluded from the study. Children with secondary malnutrition due to other medical conditions or children who had edema due to other causes were also excluded from the study.
Nutrition rehabilitation clinic (NRC) was conducted twice weekly in the outpatient department of Indus Hospital. The clinic catered to malnourished children with ages ranging from 6 month- 5 years. Malnutrition was diagnosed on basis of weight, height and mid upper arm circumference cut-off values prescribed by WHO. At enrollment nutritional details of each child were recorded on pilot tested, predesigned questionnaire which included nutritional history and details of physical examination. Examination was done for anthropometry and clinical features like edema, dermatosis, anemia, rickets and eye changes. If there were clinical signs of micronutrient deficiencies, then relevant labs were sent. Children with rickets and anemia were treated with oral iron and vitamin D3 supplements. The dosages of iron and vitamin D were prescribed according to WHO protocols. Blood transfusion was done in cases of severe anemia. Vitamin B12 deficiency was treated with oral Cobalamin according to institutional protocol. All the malnourished children were given multiple micronutrient powder (MNP) and zinc supplements. Antibiotics were given when needed.
Mothers were counseled on age appropriate feeding practices and hygiene strategies through Infant young child feeding practices (IYCF) counseling cards.. The diet plans were made by the nutritionist at Indus hospital using ingredients which were indigenous and available in normal households. Cooking of fortified recipes was demonstrated to the mothers in the cooking area. Quantities of the ingredients were shown by using spoon and measuring cups. After demonstration of recipes, brochures containing pictorial and written instructions in Urdu were f given to the mothers for reminder at home. Meal frequencies ranging from 2 to 6 times per day were advised based on age of child. For non-breast-fed children milk and milk products were added. About 150–220 Kcal/kg/day of calories and 3–5 g/kg/day of proteins were advised. Calories and proteins were gradually escalated in the diet.
Children were regularly followed according to severity of malnutrition. Moderately malnourished children were followed 3 weekly whereas severely malnourished children were called fortnightly. At every follow up visit mothers were asked about cooking techniques, re-counseling by the nutritionists was done if there was any variation in recipe or cooking method. Height, weight and MUAC were recorded at every follow up visit. Weight gain was calculated on basis of weight for height ratio and MUAC cutoffs. Adequate weight gain was considered when weight for height/length Z-score was equal or more than 1 standard deviation or MUAC was more than 12.5 cms. Absenteeism from follow up for more than r 6 consecutive weeks was considered as default. The children after recovery were followed for 2 months to ensure continuous weight gain.
Severe acute malnutrition (SAM)
SAM was labeled if any of the three criteria was present (i) weight for height/length Z- score < − 3.0, or (ii) mid upper arm circumference < 11.5 cm, or (iii) pitting pedal edema .
Moderate acute malnutrition (MAM)
MAM was labeled when weight for height/length Z- score was <− 2.0, or (ii) mid upper arm circumference was between 11.5–12.5 cm .
Severe when hemoglobin is less than 6 g/dl and moderate when hemoglobin level is between 6.1–11 g/dl .
Vitamin B12 deficiency
Plasma vitamin B 12 level < 203 pg/mL .
Serum 25(OH) D levels at < 30 nmol/L with or without clinical signs of Rickets .
Weight gain was calculated in g/Kg/day. 5 g/kg/day is considered adequate weight gain .
Presence of pitting edema on dorsum of feet or shin of legs or peri-orbital edema.
Child was considered cured when weight for height/length SD score was > − 1.0 SD, or mid upper arm circumference was > 12.5 cm, whichever came first.
Data collection procedure
A structured data extraction form was used for data collection. Data were gathered for demographic characteristics, feeding, and micronutrient, anthropometric and nutritional details at enrollment and on follow-up. Data were collected by nutritionists and doctor. The data extraction form was adopted from WHO guidelines  and Sphere standard for management of severe acute malnutrition .
Data management and analysis
The statistical analysis was performed using Stata 16.0 software. Normality assessment of continuous variables was done on the basis of skewness and kurtosis. Normally distributed variables were reported as mean [SD] whereas median (IQR) was calculated for skewed variables. Paired T test was applied to compare the entry and exit variables for normally distributed variables, while Wilcoxon sign rank test was used for non-normal data. The categorical data was presented as frequencies and percentages. McNemars test was applied to measure the difference between (2 × 2) variables, whereas others with more than two categories were assed via McNemars Bowker test. P value < 0.05 was considered to be significant.