Criteria for considering studies for this review
Types of studies
Randomized and quasi-randomized controlled trials. A quasi-randomized controlled trial is where the intervention is allocated in a way that is not truly random; for example, allocation by date of birth, day of the week, medical record number, month of the year, or the order in which participants are recruited into the study (e.g. alternation).
The unit of randomization may include individuals, families, households, communities, or other clusters.
Types of participants
Children aged 0-5, born at term (> or = 37 completed weeks of gestation). Studies involving children with undernutrition due to secondary causes, e.g. cystic fibrosis, metabolic and endocrine disorders, etc., will be excluded.
Parents, mothers, or primary caregivers of preschool children receiving supplemental food randomised to receive a nutritional education intervention.
We will exclude studies that provide food supplements during pregnancy as the objective of this review is to assess the impact of two nutritional interventions targeted to young children after birth.
Types of interventions
Supplementary feeding to children combined with nutritional education for their caregivers.
Supplementary feeding will be defined as the provision of extra food to children or families beyond the normal rations of their home diets. The intervention has to be "community-based", i.e. young children could consume the supplementary food at home, at a supervised feeding centre, or other places adapted for this purpose.
Trials in hospital and emergency food aid and refugee settings will be excluded in view of the different family dynamics that may not be comparable to those of the majority of families/children receiving supplementary feeding at the community level.
Supplementary feeding could comprise calorie/protein with or without micronutrient supplementation and could consist of:
-meals (local or imported foods)
-drinks (juices or milk)
-snacks (including both food and milk snacks)
Nutritional education for caregivers will be defined as any activity to ensure that carers of young children received nutrition advice about child food and feeding practices for their pre-school children. Personalised (individual) and family-focused educational strategies will be included. Different approaches include: sessions on nutrition and health education (e.g. weekly mandatory sessions), structured long-term educational programs with motivational enhancement interventions, participative methods or techniques (e.g. problem solving, case discussion, role-playing), etc. The intervention should be delivered by community health or nutrition workers.
Excluded: Mass-marketing campaigns (television, radio, and newspaper), video-based training or prompts, and mailed messages (e.g. newsletter, tip cards) and educational interventions aimed exclusively at promoting breastfeeding. Nutritional education in rehabilitation health centres for severe undernutrition will be also excluded.
Controls will include only children receiving no treatment (i.e. home diet, no extra feeding or nutritional education).
The following comparisons groups will be included:
Supplementary feeding and nutritional education versus no treatment
Supplementary feeding and nutritional education versus supplementary feeding
Supplementary feeding and nutritional education versus nutritional education
Types of outcome measures
Primary outcome measures (at the end of the intervention):
1) Weight expressed in kg or weight-for-age (W-F-A) z-score, a measure of underweight status (WFA tends to fluctuate seasonally in many countries);
2) Length /height expressed in cm or length/height-for-age (L-F-A) z-score, or level of stunting (LFA is a seasonally-stable measure of a child's growth, but does not generally reflect short-term effects);
3) Weight-for-height (W-F-H) z-score, or level of wasting (WFH tends to fluctuate seasonally in many countries);
4) Child development measured using standardised testing in any of the following linked domains: sensori-motor, cognitive-language, and social-emotional function.
Secondary outcomes measures will include:
5) Morbidity: number of episodes of cough, diarrhoea or fever within the study groups provided that a comprehensible definition is reported in the original study.
6) Caregivers' satisfaction could include, for example, views on the frequency and duration of nutritional advice, opinion on the information provided, applicability of learned recommendations, etc.
Search methods for identification of studies
In addition to searching the Specialised Register of clinical trials maintained by the Cochrane Developmental, Psychosocial and Learning Problems Review Group, the following search terms will be used to search the Cochrane Controlled Trials Register (CENTRAL), MEDLINE, EMBASE, CINAHL, LILACS, Social Science Citation Index, Dissertation Abstracts International, National Research Register and ClinicalTrials.gov. Where necessary, the search terms will be modified to suit the requirements of particular databases. No language limitations will be applied. Appropriate trials filters will be added to the strategy where necessary.
The following electronic search strategy will be used to search MEDLINE:
1 exp CHILD/
2 exp Infant/
3 (child$ or infan$ or baby or babies or pre-school$ or preschool$).tw.
5 ((supplement$ or extra) adj5 (food or feed$ or diet$ or nutrition$)).tw.
6 exp Education/
15 advice$ or advise$
References of retrieved articles and relevant reviews will be scanned for potentially eligible studies. Letters will be sent to the authors of included trials asking for help in clarifying relevant and missing data and to identify reports of unpublished or ongoing trials. Up to three different contact attempts will be made per identified author (e-mail, fax, telephone).
We anticipate that a great deal of relevant literature may be unpublished, or published only as in-house reports, and will attempt to make contact by email or letter with relevant NGOs and experts in the field to locate reports of unpublished or ongoing studies. In addition, we will strive to consult websites and databases of the following organisations: WHO, WFP, International Food Policy Research Institute (IFPRI), Famine Early Warning System (FEWS), The Centre for the Study of African Economies, the Institute for Agriculture and Trade Policy, InterAction.org, and USAID.
Data collection and analysis
Selection of trials
Titles and abstracts of articles retrieved will be assessed independently by two authors (YS and JD) to determine whether they might meet the inclusion criteria. A specially developed data extraction form will be developed for this purpose. Reviewers will not be blinded to the names of the authors, institutions or journal of publication. Differences of opinion about suitability for inclusion will be resolved by discussion. Any doubt during this process will be resolved by consultation with a third reviewer (EA) or through discussion with the editorial base. If necessary, further information will be sought from trialists.
Data will be extracted independently by two authors (YS and JD). The following data will be recorded on data extraction forms: method of random allocation to treatment/control groups, details about participants, description and length of the intervention, description of co-interventions, data on outcomes related to child growth, development and morbidity, and rates of withdrawals. Any disagreement between the reviewers will be discussed and resolved by consultation with a third reviewer (EA).
Two reviewers (YS and JD) will independently assess each included study on a number of criteria, in particular:
Allocation bias (Was group assignment determined randomly and so concealed from that clinicians and participants remained unaware of upcoming assignments?)
Allocation concealment will be defined as below, as described Cochrane Reviewer's Handbook (Higgins 2005):
(A) Indicates adequate concealment of allocation (e.g. for example, by telephone randomisation, or use of consecutively numbered, sealed, opaque envelopes).
(B) Indicates uncertainty about whether the allocation was adequately concealed (e.g. where the method of allocation concealment is unknown).
(C) Indicates that the allocation was definitely not adequately concealed (e.g. open random number lists or quasi-randomisation methods such as alternate days, odd/even date of birth, or hospital number).
(D) Indicates random allocation was not used.
Detection bias (Were outcomes influenced by anything other than the constructs of interest, including biased assessment or the influence of exposure on detection?)
Blinding in outcome assessment will be rated as below:
-MET: assessor unaware of the assigned treatment when collecting outcome measures.
-UNCLEAR: blinding of assessor not reported and cannot be verified by contacting investigators.
-NOT MET: assessor aware of the assigned treatment when collecting outcome measures.
Attrition bias (Could deviations from protocol, including missing data and dropout, have influenced the results?) (Delgado-Rodriguez 04, Juni 2001).
In addition, losses to follow-up will be classified as follows:
ADEQUATE: loss to follow-up less than 20% in each of the comparison groups.
UNCLEAR: losses to follow-up not reported.
INADEQUATE: losses to follow-up equal or greater than 20% in any of the comparison groups.
Outcome validity (Were the outcome measures on a standardised scale?)
Two authors (YS and JD) will independently code all studies. Citations and data will be entered and organized in RevMan 4.2 (RevMan 2003).
When necessary, the corresponding author will be contacted to supply any unreported data (e.g., group means and standard deviations (SDs), details of dropouts, and details of interventions received by the control group). If a study reports outcomes only for participants completing the trial or only for participants who followed the protocol, authors will be contacted and asked to provide additional information to facilitate an intention-to-treat analyses.
Data synthesis and analysis
Data synthesis will be conducted with RevMan 4.2, or the latest version of the Cochrane Collaboration's meta-analysis software (RevMan 2003). Meta-analyses may be conducted to combine comparable outcome measures across studies, if applicable (significant heterogeneity resulting from variations in participant characteristics, study settings and types of educational programs may preclude this).
a. Continuous data
Continuous data will be analyzed if means and standard deviations were available and there is no clear evidence of skew in the distribution. If continuous outcomes are measured identically across studies, an overall weighted mean difference (WMD) and 95% confidence interval (CI) will be calculated. If the same continuous outcome is measured differently across studies, an overall standardised mean difference (SMD) and 95% CI will be calculated (Higgins 2005). SMDs will be calculated using Hedges g.
b. Binary data
Binary outcomes will be analyzed by calculating odds ratios with 95% confidence intervals. RevMan 4.2 uses Mantel-Haenszel methods for combining binary outcome data across studies.
If some primary studies report an outcome as a dichotomous measure and others use a continuous measure of the same construct, two separate meta-analyses will be used (one for odds ratios and another for SMDs). When a primary outcome study reports multiple measures of the same construct at different points in time, we will use a single measure that is closest to a one-year follow-up.
Statistical methods for cluster-randomised trials to be used in the review are described in section 8.11.2 of the Cochrane Reviewer's Handbook (Higgins 2005).
For binary outcomes an intention-to-treat analysis will be performed with two assumptions (best-case scenario: none of the dropped-out grew better; worst-case scenario: all of the dropped-out did not grow better) and sensitivity analyses will be performed to test these assumptions.
The consistency of results will be assessed using the I2 statistic (Higgins 2002, Higgins 2003). If there is evidence of heterogeneity (p value from test of heterogeneity < 0.1 coupled with an I2 value of 25% or greater), we will consider sources according to pre-specified subgroup and sensitivity analyses (see below), but will not calculate an overall effect size estimate. If the primary studies are judged to be substantially heterogeneous even within these sub-groupings, only a descriptive analysis will be conducted.
Depending on the data reported in the included studies, the following sub-group analyses will be undertaken:
(i) the differential impact of studies including children younger than 24 months
(ii) the differential impact of studies conducted in children formally assessed as malnourished at baseline, defined as underweight (weight-for-age below -2 standard deviation from the reference median value of the NCHS/WHO), stunting (height-for-age below -2 standard deviation from the reference median value of NCHS/WHO), and wasting (weight-for-height below -2 standard deviation from the reference median value of the NCHS/WHO)
(iii) the differential impact of interventions lasting more than 12 months.
If appropriate, sensitivity analyses will be performed to explore the influence of the following factors on effect size:
1. Study quality, as specified above.
2. Studies considering cluster and non cluster RCTs.
The robustness of the results will also be tested by repeating the analysis using different statistic models (fixed effect and random effect models).
Assessment of bias
If sufficient studies are found, funnel plots will be drawn to assess the presence of possible publication bias. Whilst funnel plot asymmetry may indicate publication bias, this is not inevitably the case (Egger 1997), and possible explanations for any asymmetry found will be considered and discussed in the text of the review. As the best weapon against publication bias is a robust search strategy, every effort will be made to contact relevant organisations and experts in the field to identify unpublished or ongoing studies.
Qualitative data from included studies may be considered to better understand the influence of contextual factors such as: type of supplementation (calorie/protein with or without micronutrient supplementation), delivery of interventions, age of participants, study setting.