Improved methods of community/home management and follow-up of severely undernourished children need to be developed in low-income countries like Bangladesh. The proposed study will be conducted at ICDDR,B Hospital and in different urban primary health care centres within Dhaka city. The efficacy of four different interventions will be evaluated/compared with control with respect to the rate of completion of clinical follow-up and growth, morbidity, and changes in psychomotor development in 500 children aged 6-24 months presenting to ICDDR,B with initial weight-for-age (WA) <-3 Z score. The interventions are:

1. Usual follow-up (fortnightly for 1st 3 months) service at the hospital nutrition follow-up unit (HNFU) including growth monitoring and promotion + health education + micronutrient supplementation and treatment of intercurrent illness (current standard practice; control group).
2. All management like control group, but follow up at community-based nutrition follow-up unit (CNFU) (fortnightly for 1st 3 months) (intervention A).
3. All management like intervention A at CNFU + supplementary food (SF) for the first 3 months (intervention B).
4. All management like intervention A at CNFU + psychosocial stimulation (PS) for the first 3 months (intervention C).
5. All management like intervention A at CNFU + SF + PS for the first 3 months (intervention D).

After the initial three months intervention period, children in all groups will be followed on fortnightly basis for next three months in their respective follow up unit.

Follow-up rate, nutritional status, mental and psychomotor development, behavior, morbidities (including intestinal permeability in a sub-sample) of the children, and child rearing practices, depression and self esteem of mothers will be assessed in different groups.

By ascertaining the most cost-effective/best management package, and by extending the package to other systems in the country, it will be possible to rehabilitate a greater number of severely malnourished children in their communities.

Study Design. A prospective, randomized, intervention trial.

Study site. The study will be conducted in Dhaka, Bangladesh. The children will be recruited on discharge from the Dhaka Hospital of ICDDR,B after treatment of their acute illnesses, and the follow up visits will be conducted at HNFU or CMFU (will be established) at four PHC or community clinics of Demra, Gulshan, Sabujbagh and Mirpur thanas of Dhaka city. The rationale for choosing these four thanas is that higher proportions of children with severe malnutrition come from these four thanas or nearby areas (unpublished data; Ahmed et al., ICDDR,B, 2001).

Initial hospital-based treatment protocol. At ICDDRB, severely malnourished children are treated in either the general ward or the special care unit of the Dhaka Hospital, according to a standardized protocol [50], which is similar to that recommended by WHO [5]. Implementation of this protocol has resulted in a 47% reduction in the case-fatality rate (CFR) among severely malnourished children [51] and according to recent observation there was further reduction of the case-fatality rate to 68 % (the current CFR is 5.4%).

Criteria for enrollment. Once the children complete the initial phase of management of any acute problems associated with severe malnutrition, they will be considered for recruitment, according to the following inclusion criteria:

1. Age 6 - 24 months
2. Either sex
3. Severely malnourished [weight for age (WA) < -3 Z score]
4. Resolution of acute illnesses
5. Not planning to leave the current residence within next six months (for the purpose of follow-up)
6. Informed consent granted from the guardian.

Exclusion Criteria. Children with any of the following clinical problems or other limitations will be excluded from study:

1. Persistent anorexia, fever, severe wasting (WH < -3 Z score) or edema
2. Clinically apparent congenital/acquired disorders that may affect growth
3. Other acute or chronic diseases requiring hospitalization and/or affecting growth
4. Lack of fixed address (to avoid difficulties in tracing for follow up examinations)
5. Caregiver's condition: children accompanied by caregivers who are not able to give any time for child care or unable to give stimulation due to any condition e.g. old age, disease, physical or mental handicap.

Randomization. Once the children fulfill all the criteria for entry into the study and written consent to participate is granted by their parents/guardian, the children will be randomly assigned to one of the five treatment packages (control or interventions A, B, C or D). Four sets of randomization will be done separately (one set each for the children coming from four thanas- Demra, Gulshan, Sabujbagh and Mirpur or respective nearby thanas) according to computer-generated random numbers using permuted blocks of block lengths of five and ten. The assigned treatment packages will be kept in closed opaque envelopes until the same serial number is assigned to an enrolled child.

Intervention: On the day of recruitment (i.e., on the day of discharge from the ICDDR,B hospital after completion of acute phase management) a health worker (HW) will accompany the child and mother to their home for identification of the home for any future home visits if required, and next day the HW and a research assistant will go to the home to collect the data about the quality of stimulation at home. At the beginning of the study, the importance of bringing the child for clinical/nutritional follow up at the follow up unit (HNFU or CNFU as assigned) at set intervals (according to assignment) after enrollment will be explained to the mothers/caregivers. Among all scheduled follow-up visits, two visits i.e. at the end of third and sixth month will be at HNFU and mothers/caregivers of all groups will be advised to bring their children to HNFU during these two visits. The specific components of treatment packages are described in the following sections.

Standard therapy: Standard therapy includes growth monitoring, health education, multiple micronutrient supplementation, and clinical check up. In addition, self-referral will be encouraged and entertained at any hour and day for intercurrent illnesses; children with suspected serious illness, (e.g. dehydrating diarrhea, pneumonia, ear or throat infection, fever, loss of appetite) will be admitted to the hospital, where standard treatment will be provided.

Growth monitoring and promotion: During each follow up visit, the children's nutritional status and general condition will be assessed. Anthropometrics will include weight, length/height, mid-upper arm and head circumference, and the measures will be recorded on a standardized form. The health workers will discuss child's overall condition, change in weight and height and diet with respective mothers/caretakers.

Health education of mothers/caregivers: The major components of the process of education are providing information, showing examples, practicing and checking understanding. During each follow-up, mothers or caregivers will be given structured lessons (standardized previously) by health worker on the causes and prevention of malnutrition, the importance of breast feeding, preparation of nutritious low-cost diets with locally available food items, home management of diarrhea, recognition of referral sign/symptoms of common illnesses, such as pneumonia, and when to seek help, use of safe water and hygienic practices, and counseling on birth spacing.

Micronutrient Supplementation. On recruitment and at each follow up visit during the first three moths of the study, multivitamin drops, folic acid (in iron tablet) and zinc (standard practice of ICDDR,B) will be provided for oral administration. Iron will be provided during weeks 2-12 at a dose of 3 mg of elemental iron/kg/d. Caregivers will be asked to bring back any leftover micronutrient/medicine at the time of each follow up visit to check the compliance.

Other services: Other components of the treatment package that will be offered to children in all study groups include: immunization according to the Expanded Programme on Immunization guidelines, deworming of children >1 year of age if such treatment has not been provided in the last six months, and counseling of parents on birth-spacing and contraceptives, which will be provided free of cost to the mothers. At the end of six months, if any child still has a WH < 90% (<-1 Z score), he/she will be invited to continue attending the nutrition follow up clinic (regular service component) of ICDDR,B.

Follow up at the community-based clinics: Available data of 2001 showed that a good proportion (~ 70%) of under-5 severely malnourished children were coming from four thanas (Demra, Gulshan, Sabujbagh and Mirpur) or nearby areas of Dhaka city. One CNFU will be established at each of these thanas within an already functioning PHC or NGO clinic. One health worker contracted for the study will conduct the CNFUs on specified times and days of the week.

Provision of supplementary food (SF): Children assigned to treatment groups B and D will additionally receive the SF packets for first three months of the study as take-home supplementary food in addition to their usual meals. The SF packets will be distributed on recruitment i.e. at the time of discharge from the hospital and at each follow up visit from the CNFU to the respective groups. A sufficient supply of packets will be provided to ensure 1 and 2 packets/day for 6 -11 and 12-24 months old children respectively until the next planned follow-up, which we consider will be effective based on our previous experience. The composition of each packet of the SF is (based on supplements commonly available through a national supplementary feeding program in rural areas): 1) toasted rice powder 20g; 2) toasted lentil powder 10g; 3) molasses 5g; and 4) soy bean oil 3g, and total energy per packet is ~ 150 kcal (~ 630 k joules) with 11% energy from protein.

The ingredients will be bought from the local market and prepared in the HNFU-kitchen. The health workers will measure the appropriate amounts into polythene packets under the guidance of a research assistant (rice powder and lentil powder in one compartment, and molasses and soybean oil in two separate compartments of each packet). Previous testing revealed that the locally packaged supplements are stable for at least two months. Caregivers will be asked to offer each packet mixed with (not more than) 30 ml (~ 6 teaspoonfuls) of water, and bring back any leftover food and/or empty packets at the time of each follow up visit to check the compliance. Under-five siblings of the index children will also receive the SF (according to their age, as stated above) for the 1st three months (to minimize sharing of SF assigned to the study subject). Mothers/caregivers of other treatment groups will receive education about how to prepare the SF at home and be encouraged to give it to their children.

Psychosocial stimulation (PS): The intervention program is comprised of child stimulation and parental counseling, which will be conducted by female health workers trained on how to give psychosocial stimulation to the children and parental counseling, how to attract the groups of attending mothers/caregivers, how to interact with the children and attract them to the play materials, etc.

Data collection: At the beginning of the study information will be sought on the families' wealth, standard of housing, family structure and parental characteristics. At the time of each follow-up visit the children will be evaluated with regard to physical growth (anthropometry), morbidity from infections and cognitive development and behavior. The data collection methods and timing of each assessment are described below.

Anthropometry: At baseline and each subsequent scheduled follow-up visit the health worker will record the children's nude weight, using a digital scale (Seca, model-345) with 10g precision, length using a calibrated length board, and head and mid upper-arm circumference to the nearest mm (non-stretch insertion tape), and triceps skinfold thickness by Harpenden calipers to nearest 0.2mm. All anthropometrics will be taken twice; if they vary, a third measurement will be done, and the average of the nearest two measures will be recorded. The visits at the end of three and six month will be held at HNFU and the anthropometric measurement taken there.

Morbidity: At each follow-up, the health worker will record child's morbidity since the previous visit on a pre-coded form. Diarrhea will be defined as the passage of three or more watery or liquid stools per day. An episode will be considered new if at least two diarrhea-free days intervene between the current and the previous episodes. Acute lower respiratory infection/pneumonia will be defined if there is cough + rapid respiratory rate (> 50/min for 6-12 months group and > 40/min for older than 12 months group). Fever will be noted if the mother thinks that her child's temperature is elevated, and cough will be registered if reported to result in at least some discomfort to the child. Although the definitions are somewhat subjective, we are planning to accept these definitions considering the constraints of the field situation. Children will be admitted to the hospital if they have any serious illness requiring hospital-based treatment.

Lactulose:Mannitol test: The test will be done in a randomly selected sub-sample of 15 children from each of the four intervention groups (A, B, C and D). Four sets of randomization will be done separately (one set each for the children receiving interventions A, B, C and D). There will be 100 opaque envelopes with serial number for each intervention group and indication of Lactulose:Mannitol test be mentioned inside the envelope randomly in 15 out of 100. Incase any mother refuses the immediate next mother in the same group will be asked for the test. Three ml sugar solutions per kg body weight (each 3 ml containing 400 mg Lactulose + 100 mg Mannitol + up to 3 ml water) will be ingested in the morning of the day of enrolment. After 30 minutes, they will be allowed to consume their usual meal and breast milk. A urine collection bag will be placed just before the children consume the test dose of sugars, and all urine will be collected for the next 5 hours. Every time the child voides, the urine volume will be measured and recorded. One drop of 20% (wt/vol) chlorhexidine gluconate will be added to each sample to prevent bacterial growth, and approximately 3 ml of the pooled urine sample will be stored at -20OC until analysis. The same procedure will be done at the end of the 3 months when the children will come to HNFU for their follow-up visit. Urinary Lactulose and Mannitol (L:M) concentrations will be measured in the Clinical Biochemistry Lab of ICDDR,B using an automated enzyme assay [55, 56]. For the sake of comparison we will also collect L:M data from 20 non-malnourished children (one time) in the same community.

Outcome Measures:

Main outcomes:

• Rate of follow up
• Rate of weight gain
• Urinary Lactulose: Mannitol (ratio)

Other outcomes:

• Psychomotor development and behavior
• Maternal depression and self esteem
• Child rearing practices of mothers/caregivers
• Morbidities: episodes and percent of time reported with diarrhea, pneumonia, fever, cough, and other illnesses
• Changes in length or height, head circumference and mid upper arm circumference
• Mortality rate, % of children remaining <-3Z score for WA

Inclusion Criteria:

1. Age 6 - 24 months
2. Either sex
3. Severely malnourished [weight for age (WA) < -3 Z score]
4. Resolution of acute illnesses
5. Not planning to leave the current residence within next six months (for the purpose of follow-up)
6. Informed consent granted from the guardian.

Exclusion Criteria:

1. Persistent anorexia, fever, severe wasting (WH < -3 Z score) or edema
2. Clinically apparent congenital/acquired disorders that may affect growth
3. Other acute or chronic diseases requiring hospitalization and/or affecting growth
4. Lack of fixed address (to avoid difficulties in tracing for follow up examinations)
5. Caregiver's condition: children accompanied by caregivers who are not able to give any time for child care or unable to give stimulation due to any condition e.g. old age, disease, physical or mental handicap.