The study was conducted in the rural community of Matlab sub-district, Chandpur, Bangladesh, located 56 kilometers southeast of Dhaka, the capital city of Bangladesh. The International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b) has been conducting a Health and Demographic Surveillance System (HDSS) in a population of about 220,000 since 1966. At Matlab, a comprehensive maternal, neonatal & child health (MNCH) programme was initiated in March 2007. Under this programme, all pregnancies in the area are being identified and followed-up bi-monthly at the household level. In addition, two additional visits are also being done (at 12-14 weeks and at 33-34 weeks of pregnancy) and counselling are also being provided to the mothers about facility-based delivery, antenatal care, and birth preparedness (21). For the present study, infants who completed six months of age were selected along with their mothers from the MNCH database. This database was upgraded bi-weekly basis after receiving information from Community Health Research Worker (CHRWs) and cross-checked by the field supervisors.
Implementation
Mothers and infants of completed six months were identified from the MNCH database for both areas between March 2011 and June 2011 randomly. At the beginning, baseline information was collected for each participant. If an infant was not present during initial household visit a schedule reattempted was carried on. Infant with severe illnesses or handicaps affecting development, feeding, or activity and absent during enrolment were excluded. The details of the enrolment have been described in Figure-1 (Trial profile).
All enrolled mothers along with other family members received two training (at enrollment and at nine months) on standard CF practices at designated sub-center health facilities for the intervention area. A training manual was developed following WHO and Bangladesh Breast Feeding Foundation guidelines.
Ethical Approval
An institutional review board (IRB) of icddr,b reviewed the protocol and provided approval. The protocol was registered (NCT03024710) at the clinical trial registration website.
Data collection and quality control
Trained community health research workers (CHRW), Field Research Assistants (FRAs) and Field Research Supervisors (FRSs) were assigned to monitor all the study activities. Children’s weight and length (both intervention and comparison groups) were measured during enrolment at 6th, 8th, 9th, 10th and 12th months of age following standardized methods and their weight-for-age (WAZ; underweight <-2SD), height-for-age (HAZ; stunting <-2SD) and weight-for-height (WAZ; wasting <-2SD) were estimated following standard method and by WHO-Anthro 2005 software. The Field Research Assistants (FRAs) and Field Research Supervisors (FRSs) received training on study questionnaires and anthropometric measurements. The FRAs were responsible for all supervision and monitoring for data collection. Day to day supervision and monitoring was carried by the FRSs using the SOP checklist. Any inconsistency or queries were re-assessed upon discussing with the respective CHRW and FRA. For quality control, 5% participants were randomly chosen and were re-interviewed by the trained FRAs within two weeks after routine interview.
Table:1 Variables consider under Dependent and Independent variables
Dependent
|
Independent
|
HAZ (Height-for-age Z score)
|
Mother Age
|
WAZ (Weight-for-age Z score)
|
Education
|
WHZ (Weight-for-height Z score)
|
|
Stunting (HAZ <-2SD)
|
Birth order
|
Underweight (WAZ <-2SD)
Wasting (WHZ <-2SD)
|
Child sex
Socio-economic status
|
CF Index (variables)
|
Religion
|
• Breastfeeding
|
|
• Bottle feeding
|
|
• Initiation of CF
|
|
• Dietary diversity (last 24 hours)
|
|
• Food group frequency (Past 7 days)
|
|
• Meal frequency (last 24 hours)
|
|
Complementary feeding index score
Table-2 describes the scoring pattern of six variables for CF (CF) practices of infants aged 6-12 months, such as continued breastfeeding, avoiding bottle feeding, timely initiation of CF, dietary diversity (past 24-hours), food group frequency, meal frequency (past 24-hours). The following seven food groups were consumed: starch staple (rice, kichuri, potato, roti, suzi etc.), pulses, milk (other than breast milk), meat/eggs, vit-A rich fruit, vegetables, other fruits and vegetables and others.
The dietary diversity (last 24 hours), food group frequency (Past 7 days), meal frequency (last 24 hours) were categorized and scored following ICFI guideline(23). The detail scorning for each ICFI items was given under Table 4.
Theoretically, the CFI score ranges from 3 to 11 (Table-2). For analysis we combined the two age-groups and grouped into terciles to form 3 categories of CF practices- low (7 or below), medium (8-9), and high (10 or above).
Table-2: Variables and scoring system used to construct the ICFI
|
Scores
|
Variables
|
6-8 months
|
9-12 months
|
Breastfeeding
|
No=0, Yes=2
|
No=0, Yes=2
|
Bottle feeding
|
No=0, Yes=1
|
No=0, Yes=1
|
Initiation of CF
|
No=0, Yes=2
|
No=0, Yes=2
|
Dietary diversity (last 24 hours)
|
Low (no diversity) =0, Medium (1-2 diversity) =1, High (≥3 diversity) =2
|
Low (no diversity) =0, Medium (1-3 diversity) =1, High (≥4 diversity) =2
|
Food group frequency (past 7 days)
|
Nil=0, 1-2 food group =1,
≥3 food group =2
|
Nil=0, 1-3 food group =1,
≥4 food group =2
|
Meal frequency (last 24 hours)
|
No meal was given = 0, Only single meal given=1, ≥2 meals given =2
|
No meal was given =0, 1-2 meals given =1, ≥3 meals given =2
|
Total maximum CFI score
|
11
|
11
|
Co-variates
Several maternal socio-demographic data were collected. Such as asset quintile (poor, middle, rich); education of the mother (collected from HDSS database), age of the mother (calculated from the date of birth to the date of entry in study and grouped a <20 years, 20-24 years, 25-29 years, ≥30 years); religion (Muslim and others); sex of the infant; education of the mother [no education, up to primary (1-5 class), up to secondary (6-10 class) and above secondary (≥11 class)]; birth order (1, 2-3 and ≥4). To estimate asset quintile assets included durable goods (e.g., table, chair, watch, television, or bicycle), housing facilities (e.g., type of toilet, or source of drinking water), housing materials (e.g., type of wall or roof), and possession of farming land.
Analysis plan and outcome measures
Baseline characteristics of the two groups were examined at enrollment. A Chi-square and t-test were performed to observe any association between the groups for categorical and continuous data, respectively. A p-value of <0.05 was considered significant. The primary outcome measure was to assess the change in all CFI indices in three time points at baseline (6 months), mid-time point (9 months) and at endline (12 months). The secondary outcomes were to assess the impact of CFI score (exposure) in changes in three nutritional outcomes (HAZ, WAZ and WHZ) over time. A Generalized Estimating Equation (GEE) was used separately for their nutritional indices for intervention and control group l (24). We considered three separate models to understand the potential confounding and modifying effect. Model 1 only CFI; model 2=Model-1 with child age and sex; and Model-3=Model-2 with maternal education, and asset score and religion.
Data cleaning was carried out by Statistical package for Social Sciences (SPSS, version 19) and per-protocol analyses were performed using stata (version 13).