Darrouzet Nardi 2016.

Study characteristics
Methods	Study design: cRCT How were missing data handled? NR Randomisation ratio: 1:1 Recruitment method: HHs were identified via 'Village Development Committees,' and all were invited to participate. A parent from each HH provided informed consent. Sample size justification and outcome used: original trial sample size was calculated to detect a difference of > 0.25 in mean WAZ with a power of 87% and a 2‐sided significance level of 0.05. Sampling method: 3 pairs of comparable communities in each district were identified, based on geographic location (including altitude), size, local natural resources, employment opportunities, availability of health care, type of agriculture practiced and other demographic features. Paired communities were randomly assigned to receive Heifer development activities either first (intervention communities) or second (control communities). Study aim or objective: to evaluate the effect of a holistic community‐level nutrition‐sensitive intervention on child dietary diversity and animal source food consumption in rural Nepal. Study period: NR Unit of allocation or exposure: communities
Participants	Baseline characteristics Intervention or exposure group: Age: NR Place of residence: HHs, n: Chitwan 59; Nuwakot 59; Nawalparasi 62 Sex: female‐headed HH, %: 13 Ethnicity and language: NR Occupation: NR Education: NR SES: HH members, n, mean: 6.44 (SD 0.21); SES score, mean: 1.63 (SD 0.08); animal‐ownership score, mean: 1.71 (SD 0.12); water source in HH, %: 23; treating drinking water, %: 4; annual income, mean: NPR 69,386 (SD 3365) Social capital: NR Nutritional status: had a kitchen garden, %: 42; HAZ, mean: –1.47 (SD 0.07); prevalence of stunting, %: 31; WAZ, mean: –2.04 (SD 0.07); prevalence of underweight, %: 49; WHZ, mean: –1.44 (SD 0.07); prevalence of wasting, %: 25; MUACZ, mean: –1.47 (SD 0.05) Morbidities: diarrhoea in previous 2 weeks, %: 8.24; fever in previous 2 weeks, %: 27.6; respiratory illness in previous 2 weeks, %: 28.67; ill days in previous 2 weeks, n, mean: 0.36 (SD 0.21); 'Health score: (maximum 6), mean: 5.35 (SD 0.05) Concomitant or previous care: NR Control Age: NR Place of residence: HHs, n: Chitwan 44; Nuwakot 72; Nawalparasi 68 Sex: NR Ethnicity and language: NR Occupation: NR Education: NR SES: HH members, n, mean: 6.46 (SD NR); SES score, mean: 1.69 (SD 0.08); animal‐ownership score, mean: 2.42 (SD 0.13); water source in HH, %: 15; treating drinking water, %: 2; annual income, mean: NPR 65,273 (SD 3310) Social capital: NR Nutritional status: had a kitchen garden, %: 48; HAZ, mean: –1.48 (SD 0.06); prevalence of stunting, %: 33; WAZ, mean: –1.94 (SD 0.06); prevalence of underweight, %: 45; WHZ, mean: –1.26 (SD 0.06); prevalence of wasting, %: 23; MUACZ: –1.51 (SD 0.05) Morbidities: diarrhoea in previous 2 weeks, %: 10.38; fever in previous 2 weeks, %: 29.87; respiratory illness in previous 2 weeks, %: 29.56; ill days in previous 2 weeks, n, mean: 2.42 (SD 0.20); 'Health score' (maximum 6), mean: 5.30 (SD 0.04) Concomitant or previous care: NR Overall Age: newborn to 92 years Place of residence: NA Sex: 1469 males; 1508 females (NR for 17 participants) Ethnicity and language: NR Occupation: both areas largely populated by low‐income subsistence farmers. Education: NR SES: NR Social capital: NR Nutritional status: NR Morbidities: NR Concomitant or previous care: NR Inclusion criteria: all members of each participating HH. Exclusion criteria: children with physical or neurological disabilities that prevented ingestion of a normal diet for age or children with severe intercurrent illnesses at time of survey Pretreatment: animal ownership was greater in control HHs (2.42 (SD 0.13) than in intervention HHs (1.71 (SD 0.12) at baseline (P = 0.0001). Attrition per relevant group: unclear, as the participatory HHs increased over duration of study. Description of subgroups measured and reported: gender; age (6–12 months; > 12 months); season (hungry vs harvest); region for dietary quality outcomes. Total number completed and analysed per relevant group: for growth outcomes at 24 months: intervention group: 305 children aged 6–60 months; control group: 306 children aged 6–60 months. Number of HHs unclear. For dietary outcomes at 24 months: 533 children in total (unclear how many per group). Number of HHs unclear. Total number enrolled per relevant group: intervention group: 201 HHs (including 283 children aged 6–60 months); control group: 214 HHs (including 324 children aged 6–60 months) Total number randomised per relevant group: intervention group: 3 study sites; control group: 3 study sites
Interventions	Intervention characteristics Intervention or exposure group: multicomponent agriculture training Food access intervention category: increase buying power Intervention type: income generation Description: agricultural‐related interventions (livestock training); sharing resources; community development. The Heifer training curriculum focused on poverty alleviation, citizen empowerment and community development, with a strong emphasis on optimisation of livestock management as a means to income generation. Training was based on the '12 Cornerstones' for holistic community development. No emphasis on child health and nutrition. In each community, local leaders were invited to serve on an advisory panel and as liaisons to the population about the project activities. Duration of intervention period: 12 months Frequency: bi‐weekly Number of study contacts: 5 (baseline and follow‐up after 6, 12, 18 and 24 months) Providers: Heifer International trained staff provided the intervention. Data collection by a local field research NGO (the Nepal Technical Assistance Group), that was not connected to Heifer. Delivery: women's self‐help groups that met with a trained facilitator, supplemented by specific interactive instruction, workshops, guidance and training. Co‐interventions: NR Resource requirements: trained staff, staff time, training material for participants, livestock Economic indicators: NR Control group: no intervention
Outcomes	Dietary diversity: HH DDI, child MDD Dietary intake: % children consuming specific foods – oil, dal, milk, meat, eggs, vitamin A rich foods, other Anthropometry: HAZ, WAZ
Identification	Sponsorship source: Heifer International Country: Nepal Setting: HHs of communities mainly populated by low‐income subsistence farmers. 2 regions involved: Hills region is conducive to livestock production; Terai region is conducive to crop production. Authors' names: Amelia F Darrouzet‐Nardi; Laurie C Miller Email: adarrouzetnardi@allegheny.edu; lmiller1@tuftsmedicalcenter.org Declarations of interest: yes; no conflicts of interest. Study or programme name and acronym: Heifer International Intervention Type of record: journal articles
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	Unclear risk	Method of randomisation NR.
Allocation concealment (Selection bias)	Unclear risk	NR.
Baseline characteristics similar (Selection bias)	Low risk	The mean HH size was larger in 1 control community compared to the intervention community. Animal ownership was greater in 1 control community. The study authors controlled for baseline HH characteristics in their analysis.
Baseline outcome measurements similar (Selection bias)	Unclear risk	Study authors reported no differences in growth parameters or morbidity of children aged 6–60 months between the intervention and control groups at baseline (Miller 2014). DDSs, e.g. DDS, MDD at baseline between the intervention and control groups were NR (Darrouzet‐Nardi 2016).
Blinding of participants and personnel (Performance bias)	Low risk	Although blinding was not possible, it was likely that a lack of blinding influenced the children's nutritional outcomes.
Blinding of outcome assessment (Detection bias)	High risk	The enumerators who collected baseline data were blinded to the allocation of the intervention but not those collecting data during the subsequent 12 months. Outcomes were based on self‐report, which could have been influenced by participant's knowledge of treatment allocation.
Protection against contamination (Performance bias)	Low risk	The intervention and control communities were not adjacent to each other to minimise spillover effects.
Incomplete outcome data (Attrition bias)	Unclear risk	NR how many HHs and participants dropped out per group; and it seems that HHs were added after randomisation and initial enrolment. It is also unclear how missing data were handled. No clusters were lost.
Selective outcome reporting (Reporting bias)	Unclear risk	Study protocol N/A.
Other bias	Unclear risk	Misclassification bias: unlikely. Measurement bias: low risk. Single 24‐hour dietary recall at baseline and repeated during follow‐up. Bias between intervention clusters due to level of participation in intervention. Quote: "Villages were randomly assigned to either Group 1 or Group 2, but within villages, HHs could choose their level of participation in the intervention, creating the potential for selection bias." Incorrect analysis: high risk. Study authors did not adjust for clustering in their analysis. Recruitment bias: unclear. Unclear whether HHs were recruited before or after randomisation.