Murshed E Jahan 2011.

Study characteristics
Methods	Study design: PCS How were missing data handled? No missing data reported by study authors. Randomisation ratio: N/A Recruitment method: NR Sample size justification and outcome used: NR Sampling method: stratified random sampling. Respondents were selected in such a way that farmers of different wealth ranks were included by means of a participatory wealth ranking exercise undertaken in the study sites. The participants identified land holding as the proxy measure of wealth for ranking the HHs. All HHs were ranked based on their access to different natural resources. Study aim or objective: to assess the farm productivity, profitability, efficiency and HH nutrition of participants in integrated aquaculture‐agriculture training (with and without a small financial grant) and of a group of control farmers. Study period: 3 years; 2002/2003 to 2005/2006 Unit of allocation or exposure: individuals (farmers)
Participants	Baseline characteristics Intervention or exposure group: Age: NR Place of residence: districts in Bangladesh Sex: NR Ethnicity and language: NR Occupation, %: farmers: 100 Education: NR SES: n (%) of farmers, according to farm area: poor (0.20 ha): 73 (28.1); marginal (0.20–0.60 ha): 81 (31.2); medium (0.61–1.21 ha): 64 (24.6); rich (> 1.21 ha): 42 (16.2) Social capital: access to local government institutions such as the Department of Fisheries (DOF), the Department of Agriculture Extension (DAE), or the Department of Livestock (DOL), which enables farmers to seek technical support: 10% of project farmers. Enabled the project farmers to take leading roles in community organisations such as fish farmer groups, mosque committees and school committees: 7% Nutritional status: NR Morbidities: NR Concomitant or previous care: NR Control group: Age: NR Place of residence: districts in Bangladesh Sex: NR Ethnicity and language: NR Occupation, %: farmers: 100 Education: NR SES: n (%) of farmers according to farm area: poor (0.20 ha): 31 (24.6); marginal (0.20–0.60 ha): 41 (32.5); medium (0.61–1.21 ha): 32 (25.4); rich (> 1.21 ha): 22 (17.5) Social capital: access to local government institutions such as the Department of Fisheries (DOF), the Department of Agriculture Extension (DAE), or the Department of Livestock (DOL), which enables farmers to seek technical support: 10% farmers. Enabled the project farmers to take leading roles in community organisations such as fish farmer groups, mosque committees and school committees: NR (4% in 2005/2006) Nutritional status: NR Morbidities: NR Concomitant or previous care: NR Overall: NR Inclusion criteria: farmers from 4 districts (Mymensingh, Comilla, Magura and Bogra). No other criteria reported. Exclusion criteria: farmers who had previously received training in aquaculture production or integrated aquaculture‐agriculture and those who did not agree to participate were excluded from the pool. Pretreatment: baseline characteristics were not assessed for the control group; table 2 presented characteristics of both groups assessed in 2006, which showed no differences between them. At baseline the table showed that there was no significant difference between different groups of farmers included in the study at the time of selection in terms of farm area, i.e. number of hectares owned (land holding was proxy measure for wealth). Attrition per relevant group: attrition NR. Description of subgroups measured and reported: none Total number completed and analysed per relevant group: intervention (project farmers) 260 (grant farmer: 127; non‐grant farmers: 133); control farmers 126 Total number enrolled per relevant group: intervention (project farmers) 260 (grant farmer: 127; non‐grant farmers: 133); control farmers = 126 Total number randomised per relevant group: N/A
Interventions	Intervention characteristics Intervention or exposure group: integrated agriculture‐aquaculture intervention Food access intervention category: increase buying power Intervention type: income generation Description: integrated agriculture‐aquaculture intervention with and without a small financial grant. Farmers were provided with 3 years of continuous training and extension support to enable them to become efficient in utilising inputs such as labour, organic fertiliser and capital. The project tested the adoption of integrated aquaculture‐agriculture under 2 different models: 1 with the provision of a small financial grant (which was provided for purchasing inputs especially fish seed but also other inputs), and 1 without. The project aimed at improving resource‐use efficiency to increase farm‐level productivity in a sustainable manner through diffusion of low‐cost integrated aquaculture‐agriculture approaches suitable for poor farmers with limited resources. Long‐term training and close extension support were provided to the project farmers. These farmers received 3 training sessions during the first year, 2 during the second year, and 1 follow‐up training in the final year. Formal training was complemented by regular informal training sessions, such as group meetings at the pond/plot site using the Participatory Adaptive Learning approach (PAL), and annual participatory evaluation sessions. Under the PAL approach community members and field staff learn together while going through the process. Duration of intervention period: 3 years; 2002/2003 to 2005/2006 Frequency: 3 training sessions during the first year, 2 during the second year, and 1 follow‐up training in the final year. The grant seemed to have been provided one‐off but this was unclear. Number of study contacts: "Research assistants visited each family on a bi‐monthly basis to collect the information, help complete the form where necessary and answer technical questions. Providers: NGO Project staff; 48 partner NGOs participated in the programme and disseminated low‐cost aquaculture technologies. Delivery: Farmers received technical and extension support from 2003/2004 to 2005/2006. They were trained in filling in the record book provided to them to monitor all on‐farm production activities over the duration of study. Research assistants visited each family on a bi‐monthly basis to collect the information, help complete the form where necessary and answer technical question. Respondents kept daily records of consumption in a consumption diary from 2003/2004 onwards. Co‐interventions: NR Resource requirements: NR Economic indicators: NR Control group: no intervention
Outcomes	Total and per capita HH food consumption – total/for individual foods
Identification	Sponsorship source: USAID; World Fish Center Country: Bangladesh Setting: small‐scale farms in 4 Bangladesh districts Author's name: Khondker Murshed‐E‐Jahan Email: k.jahan@cgiar.org; d.pemsl@cgiar.org Declarations of interest: NR Study or programme name and acronym: Development of Sustainable Aquaculture Project (DSAP) Type of record: journal article
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	High risk	CBA; no randomisation done.
Allocation concealment (Selection bias)	High risk	CBA; no allocation concealment done.
Baseline characteristics similar (Selection bias)	Unclear risk	Study authors reported similar farm areas and access to social capital for both groups at baseline; however, other baseline characteristics were NR.
Baseline outcome measurements similar (Selection bias)	Unclear risk	No baseline measurements reported.
Blinding of participants and personnel (Performance bias)	Low risk	No blinding but this was unlikely to have influenced participant or personnel behaviour.
Blinding of outcome assessment (Detection bias)	High risk	There was no blinding. Measures were based on self‐report, and lack of blinding could have influenced reporting.
Protection against contamination (Performance bias)	Unclear risk	Control farmers were selected from the same districts as project farmers but it was not clear what their distance was from project farmers and whether any of these started any integrated aquaculture‐agriculture projects by themselves.
Incomplete outcome data (Attrition bias)	Low risk	All farmers completed the study; there were no missing data.
Selective outcome reporting (Reporting bias)	Unclear risk	No protocol available.
Other bias	Unclear risk	Misclassification bias: unlikely. NGO assigned farmers to intervention. Measurement bias: low risk. HHs completed daily food consumption diary. Seasonality bias: low risk.