Cohen 1995

Methods	Study design: Randomised controlled trial Funding: “Supported jointly by the Thrasher Research Fund; the World Health Organization; UNICEF/Honduras and the Institute for Reproductive Health (formerly the Institute for International Studies in Natural Family Planning), Georgetown University, under a Co‐operative Agreement with the U.S. Agency for International Development (A.I.D.) (DPE‐3040‐A‐00‐5064‐00 and DPE‐3061‐A‐00‐1029‐00).”s
Participants	Description: Low income, first time mothers and their infants N (Randomised): 152 children Age: Infants: Infants were randomised at 16 weeks of age Mother (mean): 20.2 years % Female: 55% SES and ethnicity: “Subjects came from low income neighborhoods in which environmental sanitation was poor (only 60% of the households had indoor piped water). Mean household income was $120/mo.” Inclusion/exclusion criteria: Inclusion criteria: “Selection criteria were that mothers be primiparous, willing to exclusively breast‐feed for 26 wk, not employed outside the home prior to 6 mo postpartum, low income (less than $150/mo), at least 16 years old and healthy (not taking medication on a regular basis), and that infants be healthy, term, and weigh at least 2000g at birth.” Exclusion criteria: “Multiparous and working mothers were excluded because the intervention required a 3‐d stay in the La Leche League unit on three occasions to measure breast milk intake.” Recruitment: “Subjects were recruited from two public hospitals in San Pedro Sula, Honduras” Recruitment rate: 86% (152/176) Region: San Pedro Sula, Honduras
Interventions	Number of experimental conditions: 3 Number of participants (analysed): Solid foods: 42 Solid foods + maintenance: 39 Exclusive breast‐feeding: 42 Description of intervention: Solid foods: introduction of solid foods at 4 months, with breast‐feeding as required 4‐6 months. Solids foods + maintenance: introduction of solid foods at 4 months, with mothers told to continue breast‐feeding as often as they had prior to the intervention. Exclusive breast‐feeding: exclusive breast‐feeding to 6 months; no other liquids (water, milk, formula) or solids In addition all mothers: 1. Stayed at the la Leche League unit at 16 weeks for 3‐days and returned to the unit at weeks 21 and 26 weeks for repeated measurements. 2. Received weekly home visits during the intervention period to collect data on breast‐feeding and infant morbidity. In the solid food groups these weekly visits also were used to monitor use of the foods provided and encourage mothers in the maintenance group to maintain their re‐intervention breast‐feeding frequency. Duration: 2 months Number of contacts: 13 (10 weekly home visits + 5 hospital visits) Setting: Home + hospital Modality: Face‐to‐face Interventionist: Mothers Integrity: “To encourage compliance with study procedures, mothers recorded the number of breastfeeds each day from 16 to 26 weeks on a simple form provided weekly. This was especially important for the SF‐M mothers, who were asked to maintain breastfeeding frequency. At 19 and 24 weeks, 12‐hour in‐home observations were conducted to record breastfeeding frequency and duration, and adherence to the feeding instructions.” Date of study: Recruited from October 1991‐January 1993 Description of control: N/A
Outcomes	Outcome relating to children's fruit and vegetable consumption: Consumption of fruit (grams): “the amount of food offered and consumed at the midday meal was measured (using an electronic scale, to the nearest gram)” Outcome relating to absolute costs/cost‐effectiveness of interventions: Not reported Outcome relating to reported adverse events: Not reported Length of follow‐up from baseline: 9 and 12 months Length of follow‐up post‐intervention: 2.5 and 5.5 months Subgroup analyses: None Loss to follow‐up (at 9 and 12 months): Unclear ‐ states “for a subsample of infants, n=60 at 9 months and n= 123 at 12 months” Analysis: Unknown if sample size calculation was performed
Notes	First reported outcome (frequency of consuming fruit) at 9 months for the < 12 months was extracted for inclusion in meta‐analysis. Sensitivity analysis ‐ primary outcome: primary outcome not stated, fruit or vegetable consumption was not first reported outcome (first reported outcome was dairy).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	"At 16 wk, subjects were randomly assigned, by week of infant's birth, to one of three groups: 1) Control: Exclusive breast‐feeding to 26 wk (EBF); 2) Solid Foods: Introduction of solid foods at 16 wk (SF), with ad libitum breast‐feeding; or 3) Solid Foods‐M: Introduction of solid foods at 16 wk (SF‐M), with mothers told to continue breast‐feeding as often as they had prior to the intervention.” Allocated to group by week of birth.
Allocation concealment (selection bias)	Unclear risk	There is no information provided about allocation concealment and therefore it is unclear if allocation was concealed from those conducting the research.
Blinding of participants and personnel (performance bias) All outcomes	High risk	”Subjects were not informed of their assignment until they had completed the first 16 wk of the study.” “All women were visited weekly during the first 4 mo postpartum to assist them in maintaining exclusive breast‐feeding.” Due to the nature of the intervention, both participants and personnel were aware of group allocation after 16 weeks.
Blinding of outcome assessment (detection bias) All outcomes	High risk	“During the 9‐ and 12‐mo visits, the amount of food offered and consumed at the midday meal was measured (using an electronic scale, to the nearest gram) for a subsample of infants (n = 60 at 9 mo, n = 123 at 12 mo), and their mothers were interviewed regarding the infants' usual daily food intake and acceptance and frequency of consumption of a variety of common foods.” It is unclear whether outcome assessors visiting the home were aware of group allocation. Mothers self‐reported food intake and acceptance.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	“Home visits were conducted for a subsample only (total n=141). 9mth n=60; 12 mth n=123.” Unclear if this is actual subsample or if this reflects attrition/non‐response It is unclear whether the n value for the subsample represents everyone who was eligible (i.e. had infants younger than 12 months prior to May 1993) with 100% consent rate, or if there were refusals.
Selective reporting (reporting bias)	Unclear risk	There is no trial registration or protocol.
Other bias	Unclear risk	It is unclear how the women were recruited, what the consent rate was, or how representative the sample was of the target population.