Kamala 2018a.

Methods	Study design: cluster‐randomised controlled trial Country: Tanzania Setting: smallholder farming households producing maize, primarily for home consumption, from villages in the Northern highland, Eastern lowland, and South‐Western highland Study aim: to evaluate the effectiveness of a post‐harvest intervention package as a mitigation strategy for reducing aflatoxin and fumonisin contamination in maize, and subsequently, dietary exposure of infants Study dates: May to November 2013 (Hanang and Kilosa districts); February to August 2014 (Rungwe district) Recruitment: eligible infants aged 0 to 6 months were identified from their registration number and date of birth, and their households were recruited one month before the harvesting period Sampling: purposive sampling of three maize‐producing districts; followed by random sampling of 30 villages, and stratified sampling of 10 households from each village Sample size justification: yes; using a difference of 50% in the proportion of children exposed to fumonisin B1 above the provisional maximum tolerable daily intake (PMTDI) of 2 ug/kg bw/day; alpha of 0.05 (two‐tailed test), variation in cluster (km) at 0.25, and power at 80%. 30 clusters with 10 households per cluster required Unit of randomisation: villages Total number randomised: 30 villages (15 intervention; 15 control); 300 households (150 intervention households; 10 households per village and 150 control households; 10 households per village) Unit of analysis: households; study authors adjusted for the effect of clustering using multilevel mixed‐effect models Intention‐to‐treat analysis: quotation: "Analysis of the primary outcomes was by intention‐to‐treat, and all missing households were analysed as part of the community in which they were enrolled." Attrition: no loss of clusters. WAZ outcome: 51 children (22 intervention, 29 control); total aflatoxin and fumonisin content of maize outcome: 39 households (14 intervention, 25 control) Relevant study limitations as reported by study authors: linear growth of children not being measured due to fieldwork limitations; lack of baseline data on various outcomes; differential dropout; lack of morbidity information for the children using maize as the only dietary source of aflatoxins and fumonisins, and the use of insecticides and consequent exposure
Participants	Inclusion criteria: households with a breastfed infant aged 0 to 6 months, born to parents who were local residents at the time of recruitment; maize producing households with capacity of storing maize for at least six months after harvest Exclusion criteria: households with an infant having a known congenital malformation or chronic condition that could affect weight Group differences: NR Subgroups: NR Baseline characteristics of participants Intervention group (N = 150) Mother Mother, age: in years, mean (SD): 26.9 (6.5) Marital status: single, living with parents, %: 23; married, %: 74; widowed, %: 1; divorced, %: 2 Mother, educational level: never attended school, %: 10; primary education, %: 75; secondary education, %: 15 Occupation: NR Mother, nutritional status: NR Mother, dietary diversity or food security measure: NR Mother, biomarker of aflatoxin exposure: NR Child Child, age: in months, mean (SD), 3.4 (2.1) Child, gender: Female, %: 44 Child, nutritional status: NR Child, dietary diversity or food security measure: dietary diversity score, mean (SD): 2.3 (0.9); no. of meals per day, mean (SD): 2.4 (1) Child, morbidity: NR Child, biomarker of aflatoxin exposure: NR Household Place of residence: NR Race, ethnicity, and religion: NR Socio‐economic status: number of children in household, mean (SD): 4 (2) Social capital: NR Household head, age: NR Household head, gender: NR Household head, educational level: NR Household, food security or dietary diversity measure: NR Control group (N = 150) Mother Mother, age: in years, mean (SD): 26.0 (5.9) Marital status: single, living with parents, %: 24; married, %: 75; widowed, %: 1; divorced, %: 0 Mother, educational level: never attended school, %: 15; primary education, %: 71; secondary education, %: 14 Occupation: NR Mother, nutritional status: NR Mother, dietary diversity or food security measure: NR Mother, biomarker of aflatoxin exposure: NR Child Child, age: in months, mean (SD), 3.9 (2.1) Child, gender: female, %: 46 Child, nutritional status: NR Child, dietary diversity or food security measure: dietary diversity score, mean (SD): 2.2 (0.9); no. of meals per day, mean (SD): 2.6 (0.9) Child, morbidity: NR Child, biomarker of aflatoxin exposure: NR Household Place of residence: NR Race, ethnicity, and religion: NR Socio‐economic status: number of children in household, mean (SD): 3 (2) Social capital: NR Household head, age: NR Household head, gender: NR Household head, educational level: NR Household, food security or dietary diversity measure: NR
Interventions	Intervention Type: post‐harvest intervention package and agricultural extension services Description: routine agricultural extension on good practice for handling crops plus training for identifying and sorting infected or damaged maize kernels, demonstrations on drying maize on raised platforms or sheets, provision of sheets for drying, education on adequate sun drying and how to test for dryness, advice to de‐hull maize before milling, training on the use of insecticides with maize kernels and the provision of insecticides Delivery: farmers gathered at nearby health facility or school in the particular village for informal meetings; intervention and agricultural extension provided from harvest until six months post‐harvest, but frequency and duration not specified Providers: Duration of intervention: six months Duration of follow‐up: six months Co‐intervention: NR Cost: NR Resource requirements: sheets for drying provided to participants, bottles and salt for testing maize dryness, and insecticides Control Type: routine agricultural extension services Description: routine agricultural extension on good practice around the handling of farmers' crop Delivery: agricultural extension was provided from harvest until six months post‐harvest. Frequency and duration of intervention was not specified, but the study authors referred to a regular service for farmers at the village level. Providers: agricultural extension (AE) workers Duration of intervention: six months Duration of follow‐up: six months Co‐intervention: NR Cost: NR Resource requirements: NR
Outcomes	Primary outcomes: Total aflatoxin content of maize (µg/kg) at 6 months Total fumonisin content of maize (µg/kg) at 6 months Secondary outcomes: Estimated aflatoxin intake of children (ng/kg/day) at 6 months Estimated fumonisin intake of children ((µg/kg/day) at 6 months Underweight‐for‐age z‐score (WAZ) at 6 months Underweight (WAZ ≥ 2 SD below the reference median value) at 6 months
Notes	Sponsorship source: Flemish Interuniversity Council‐Institutional University cooperation (VLIR‐UOS); grant number ZEIN2011PR388 Study title or name (study acronym): Intervention in minimizing aflatoxins and fumonisins exposure to children through food and breastfeeding in Tanzania Author’s name: Analice Kamala; Martin E. Kimanya Email: analice.kamala@tfda.go.tz; martin.kimanya@nm‐aist.ac.tz; Mekimanya@yahoo.co.uk Declarations of interest: no Trial registration: NCT02438774
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbers
Allocation concealment (selection bias)	Low risk	All clusters assigned simultaneously by one investigator (not involved in field work) according to randomisation sequence
Baseline outcome measurements similar	Unclear risk	We do not know if there were important group differences in the review outcome, WAZ, at baseline, as these measurements were not done.
Baseline characteristics similar	Low risk	Baseline characteristics were reported as comparable for all assessed variables.
Incomplete outcome data (attrition bias) All outcomes	High risk	No loss of clusters. Overall attrition for the WAZ outcome was 17% (intervention 14.6% (22/150); control 19.3% (29/150)). Among the reasons cited was the category 'unwilling to participate' (intervention: 9/150; control: 13/150), which was not specific enough to assess whether this attrition impacted the reported effect. Furthermore, reasons were not provided for 12 trial participants with missing WAZ outcome data (6 participants per group). Missing data for the secondary outcome, WAZ, were not handled appropriately.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not clear whether personnel or participants were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The data collection team was separate from the intervention support team, and they were blinded to the village allocation.
Protection against contamination	Unclear risk	It is possible that agricultural extension officers, based within a specific geographical area, communicated with each other.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes indicated in NCT02438774 were reported, with the exception of aflatoxin and fumonisin levels in breast milk in the main citation. However, the study authors reported these outcomes in earlier conference proceedings.
Other bias	High risk	Recruitment bias (cluster‐RCT): high risk; recruitment followed randomisation. Incorrect analysis (cluster‐RCT): low risk; the study authors used multilevel mixed‐effects models, with fixed‐effect for study group, and random‐effects for the village.

NR = not reported; n/a = not applicable; RCT = randomised controlled trial; SE = standard error; SD = standard deviation; KSh = Kenyan Shillings; WAZ = weight‐for‐age z‐score;