Alaofe 2019.

Study characteristics
Methods	Study design: PCS How were missing data handled? The analytic sample was restricted to HHs or mothers with complete data at baseline and endline for a given indicator. Pregnant women at baseline (n = 3) or endline (n = 8) excluded from analyses. ITT protocol used for analysis as 19 SMG WGs at baseline became SMG NWGs at follow‐up and 2 SMG NWGs became SMG WGs. Randomisation ratio: N/A Recruitment method: villages: before the baseline evaluation, villages in the district of Kalalé were identified for possible inclusion in the SMG. WG HHs: only 1 mother or carer of childbearing age (15–49 years) who had a child aged 6–59 months at time of baseline survey (January–March 2014) was invited to participate. NWG HHs: delegates/leaders of the selected villages were contacted to obtain a complete listing of all NWG HHs with a target mother–child. From that list, a single HH was selected as a starting point, using a random number between 1 and the required number of HHs in village. Sample size justification and outcome used: sample size based on available funds with expectation that it would be able to show differences in agricultural production and changes in food security when scaled up from the original pilot study. Sample size was not based on changes in nutritional status. Sampling method: purposive (villages) and random (control HHs): participating villages needed to have potential water sources (as determined by geophysical survey to map groundwater) to support production during dry season. 16 eligible villages identified. Delegates/leaders of selected villages were contacted to obtain a complete listing of all NWG HHs with a target mother–child. From that list, a single HH was selected as a starting point, using a random number between 1 and the required number of HHs in village. Study aim or objective: to examine the impact of a 1‐year solar‐powered drip irrigation SMG programme in Kalalé district of northern Benin on mothers' nutritional status and micronutrient levels. Study period: January–March 2014 to February–March 2015 Unit of allocation or exposure: cluster: villages (16 eligible villages identified, matched and assigned to 1 of 2 groups)
Participants	Baseline characteristics Intervention: WG Age: mother/carer, years, mean: 31.92 (SD 7.73) Place of residence: village in Kalale district, northern Benin Sex, %: female: 100 Ethnicity and language, %: Gondo 30.22; Boo 39.56; Peulh 17.58; Bariba 11.54; other: 1.1 Occupation: mother, %: agricultural/other labour: 80.33; service/business: 18.03; other: 1.64 Education: mother, %: no formal education 90.5; primary or less 4.47; secondary 5.03; university or more 0 SES, %: low 17.79, middle 49.08, high 33.13. Electricity connection, %: 13.21. HH size, mean: 7.21 (SD 3.05) Social capital: NR Nutritional status: food insecurity, %: 17.32. HDDS, mean: 6.07 (SD 1.26). WDDS‐10, mean: 4.06 (SD 1.06). BMI, mean: 21.89 (SD 2.93). Prevalence of underweight, %: 9.16. Iron deficiency, %: 15.32. Iron‐deficiency anaemia, %: 6.56. Vitamin A deficiency, %: 14.29. Morbidities: NR Concomitant or previous care: NR Control: WG Age: mother/carer, years (mean): 29.69 (SD 6.49) Place of residence: village in Kalale district, northern Benin Sex, %: female: 100 Ethnicity and language, %: Gondo 34.4; Boo 35.2; Peulh 17.6; Bariba 9.6; other 3.2 Occupation: mother, %: agricultural/other labour: 83.18; service/business: 13.64; other: 3.18 Education: mother, %: no formal education 89.43, primary or less 7.32, secondary 3.25, university or more 0 SES, %: low 24.11, middle 53.57, high 22.32. Electricity connection, %: 2.48; HH size, mean: 8.59 (SD 4.32) Social capital: NR Nutritional status: food insecurity, %: 16.00. HDDS, mean: 6.05 (SD 1.26). WDDS‐10, mean: 4.87 (SD 0.98). BMI, mean: 21.72 (SD 2.94). Prevalence of underweight, %: 12.79. Anaemia, %: 49.0. Iron deficiency, %: 17.98. Iron‐deficiency anaemia, %: 13.79. Vitamin A deficiency, %: 20.22. Morbidities: anaemia, %: 49.0 Concomitant or previous care: NR Intervention: NWG Age: mother/carer, years (mean): 29.41 (SD 6.25) Place of residence: village in Kalale district, northern Benin Sex, %: female: 100 Ethnicity and language, %: Gondo 32.39; Boo 35.68; Peulh 18.31; Bariba 8.92; other 4.692 Occupation: mother, %: agricultural/other labour: 75.2; service/business 20.8; other: 4.0 Education: mother's, %: no formal education 89.29, primary or less 5.36, secondary 5.36, university or more 0 SES, %: low 21.57, middle 48.04, high 30.39. Electricity connection, %: 10.61 Social capital: NR Nutritional status: food insecurity, %: 12.02. HDDS, mean: 6.62 (SD 1.17). WDDS‐10, mean: 4.58 (SD 1.04). BMI, mean: 23.01 (SD 3.97). Prevalence of underweight, %: 4.88. Iron deficiency, %: 21.83. iron‐deficiency anaemia, %: 12.23. Vitamin A deficiency, %: 16.67. Morbidities: anaemia, %: 44.23 Concomitant or previous care: NR Control: NWG Age: mother/carer, years (mean): 28.74 (SD 6.03) Place of residence: village in Kalale district, northern Benin Sex, %: female: 100 Ethnicity and language, %: Gando 32.46, Boo 29.82, Peulh 16.23, Bariba 14.9 other 6.58 Occupation: mother, %: agricultural/other labour: 80.52, service/business: 16.88, other: 2.60 Education: mother, %: no formal education 89.61, primary or less 4.33, secondary 5.63, university or more 0.43 SES, %: low 29.15, middle 54.27, high 16.58. Electricity connection, %: 2.23 Social capital: NR Nutritional status: food insecurity, %: 20.09. HDDS, mean: 6.51 (SD 1.12). WDDS‐10, mean: 4.83 (SD 0.97). BMI, mean: 22.03 (SD 3.14). Prevalence of underweight, %: 6.57. Iron deficiency, %: 16.56. Iron‐deficiency anaemia, %: 7.91. Vitamin A deficiency, %: 25.17. Morbidities: anaemia, %: 45.73 Concomitant or previous care: NR Overall Age: NR Place of residence: NR Sex, %: female: 100 Ethnicity and language: NR Occupation: mother, %: 80.3 agricultural/other labour (all 4 groups) Education: mother, %: no formal education 90.3 (all 4 groups) SES: NR Social capital: NR Nutritional status: NR Morbidities: NR Concomitant or previous care: NR Inclusion criteria: villages: participating villages needed to have potential water sources (as determined by geophysical survey to map groundwater) to support production during dry season. Control group: similarity along several variables, including pre‐existing local WG, location along the same roads, administrative status and size. HH: women in an agricultural group, each of whom farmed her own 120 m2 plot (SMG WG and control WG); women NOT in an agricultural group (SMG NWG and control NWG). Women: in each investigated HH, only 1 mother or carer of childbearing age (15–49 years) who had a child aged 6–59 months at time of baseline survey (January–March 2014) was invited to participate in the impact evaluation. Exclusion criteria: NR Pretreatment: at baseline, there was no significant difference in HH religion, ethnicity, access to an improved source of water, self‐reported food insecurity, mothers' education level and occupation between SMG and control groups. However, a greater proportion of SMG WG HHs had older mothers, access to latrines, healthcare insecurity and high SES compared with the other 3 groups (P < 0.05). In addition, HH size in SMG NWG was lowest compared with the other 3 groups while the prevalence of access to electricity was greatest (Table 1). Attrition per relevant group: outcome: BMI: intervention group: total 161/415 (38.8%) (WG women 56/187 (30.0%); NWG women 105/228 (46.1%)); control group: total 136/359 (37.9%) (WG women 40/126/(31.7%); NWG women 96/233 (41.2%)). Outcome: HDDS: intervention group: total 111/415 (26.7%) (WG women 39/187 (20.9%); NWG women 72/228 (31.6%)); control group: total 58/359 (16.1%) (WG women 14/126 (11.1%); NWG women 44/233 (18.9%). Outcome: Hb: intervention group: total 111/415 (26.7%) (WG women 39/187 (20.9%); NWG women 72/228 (31.6%)); control group: total 95/359 (26.5%) (WG women 26/126 (20.6%); NWG women 69/233 (29.6%)). Outcome: iron: intervention group: total 148/415 (35.7%) (WG women 62/187 (33.2%); NWG women 86/228 (37.7%)). Control group: total 119/359 (33.1%) (WG women 37/126 (29.4%); NWG women 82/233 (35.1%)). Outcome: Vitamin A: intervention group: total 145/415 (34.9%) (WG women 61/187 (32.6%); NWG women 84/228 (36.8%)); control group: total 119/359 (33.1%) (WG women 37/126 (29.4%); NWG women 82/233 (35.1%)). There was some attrition from baseline to follow‐up (4.3%) that was spread across villages, with no structural differences in terms of who was most likely to dropout. Most common reason was that mothers/carers were working on their land or moved/travelled out of village on day of data collection. In addition, some blood samples were unsuitable for further processing: 5.74% due to haemolysis, 0.47% were specimens without proper requisition slips and 3.23% had insufficient sample quantity. Description of subgroups measured and reported: intervention villages: WG: HHs with women who participated in a local women's agricultural group and NWG: HHs where none of the women belonged to a women's agricultural group. Control villages: WG: HHs with women who participated in a local women's agricultural group and NWG: HHs where none of the women belonged to a women's agricultural group. Total number completed and analysed per relevant group: depended on outcome. SMG WG: BMI, n = 131 (83.44%); HDDS, n = 148 (81.32%); Hb, n = 148 (86.05%); iron, n = 125 (79.625%); vitamin A, n = 126 (79.25%). SMG NWG: BMI, n = 123 (72.35%); HDDS, n = 156 (71.23%); Hb, n = 156 (77.61%); iron, n = 142 (78.45%); vitamin A, n = 144 (81.82%). Control WG: BMI, n = 86 (90.53%); HDDS, n = 112 (90.32%); Hb, n =100 (90.91%); iron, n = 89 (89.90%), vitamin A, n = 89 (78.76%). Control NWG women: BMI, n = 137 (84.05%); HDDS, n = 189 (81.82%); Hb, n = 164 (83.67%); iron, n = 151 (83.43%), vitamin A, n = 151 (75.50%) Total number enrolled per relevant group: total: 771 women (intervention villages: 415 women (184 WG women; 228 NWG women); control villages: 359 women (126 WG women; 233 NWG women) Total number randomised per relevant group: N/A
Interventions	Intervention characteristics Intervention group: WG Food access intervention category: increase buying power Intervention type: income generation Description: SMG system: installation of low‐pressure drip irrigation system, combined with a solar‐powered water pump in each village. Each SMG was used jointly by 30–35 women belonging to the local women's agriculture group (each woman farmed her own land of 120 m2). Duration of intervention period: 1 year Frequency: continuous Number of study contacts: 2 (January–March 2014; February–March 2015) Providers: Solar Electric Light Fund Delivery: study expanded the installation of SMG systems, from previous pilot study (Alaofe 2016) Co‐interventions: women's agriculture group activities Resource requirements: see Alaofe 2016 Economic indicators: see Alaofe 2016 Intervention group: NWG Food access intervention category: increase buying power Intervention type: income generation Description: SMG system: installation of low‐pressure drip irrigation system, combined with solar‐powered water pump in each intervention village. Women who were not part of local women's agriculture groups did not have direct access to use of SMG. Duration of intervention period: 1 year Frequency: continuous Number of study contacts: 2: baseline (January–March 20140) and endline (February–March 2015) Providers: Solar Electric Light Fund Delivery: study expanded installation of SMG systems, from previous pilot study (Alaofe 2016) Co‐interventions: no WG Resource requirements: see Alaofe 2016 Economic indicators: see Alaofe 2016 Control group: no intervention
Outcomes	Dietary diversity: HDDS (0–12); Women's DDS (0–10) Anthropometry: BMI (mothers); prevalence of underweight (mothers) (BMI < 18.5 kg/m2) Biochemical: iron deficiency; vitamin A deficiency Morbidity: anaemia; iron‐deficiency anaemia
Identification	Sponsorship source: University of Stanford, the Hellman Fellows Programme at the University of California, San Diego and the University of Arizona. Country: Benin Setting: rural villages in Kalalé district, northern Benin with sufficient groundwater sources to sustain agricultural production during the dry season. Comments: none Author's name: Halimatou Alaofè Institution: NR Email: halaofe@email.arizona.edu Address: NR Declarations of interest: none Study or programme name and acronym: Solar Market Garden (SMG) Type of record: journal article
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	High risk	PCS; no randomisation done
Allocation concealment (Selection bias)	High risk	Quasi‐experimental design; allocation not concealed.
Baseline characteristics similar (Selection bias)	Low risk	Groups were different in many characteristics, but these were adjusted for in the analyses.
Baseline outcome measurements similar (Selection bias)	Low risk	Groups were different in outcomes at baseline, but these were adjusted for in the analyses.
Blinding of participants and personnel (Performance bias)	Low risk	No blinding, but no major impact expected on outcome measurement as only some outcomes were subjective.
Blinding of outcome assessment (Detection bias)	High risk	Unclear whether the field workers who collected the dietary intake data were blinded; however, dietary recall data were self‐reported and thus at high risk for reporting bias. Local health workers conducted the anthropometric measurements; and, therefore, there was a risk of detection bias.
Protection against contamination (Performance bias)	Low risk	SMG systems were only installed in intervention villages, in conjunction with local women's agriculture groups. In their analysis, the study authors reported that there was little evidence of a 'spillover' effect of these systems to NWG HHs in intervention villages, as well as to WG and NWG HHs in control villages. Note: 19 SMG WG at baseline became SMG NWG at follow‐up and 2 SMG NWG became SMG WG.
Incomplete outcome data (Attrition bias)	Unclear risk	High percentages of incompleteness of outcome data (e.g. BMI (37.9–38.8%); HDDS (16.1–26.7%); Hb (26.5–26.7%); iron intake (33.1–35.7%) and Vitamin A intake (33.1–34.9%). However, the study authors reported no significant differences in BMI, anaemia, iron deficiency or vitamin A insufficiency between women who dropped out compared to those who were not (data not shown). They do not report whether there were any differences in HDDS between those who dropped out and those who did not. There was some attrition from baseline to follow‐up (4.3%) that was spread across villages, with no structural differences in terms of who was most likely to dropout.
Selective outcome reporting (Reporting bias)	Unclear risk	No protocol available. All the important outcomes in the methods section were reported in the results section.
Other bias	High risk	Measurement bias: high risk. Although standardised scores were calculated, dietary data consisted of 2 recalls of 1 day each (at baseline and endline). Seasonality bias: low risk. Baseline and endline surveys conducted during the same season.