Pellerano 2014.

Study characteristics
Methods	Study design: cRCT Study grouping: parallel How were missing data handled? HHs that were living elsewhere or unavailable for interview for other reasons were dropped from the study. Analysis was limited to panelled HHs that were observed both at baseline and follow‐up (25% of the original sample not analysed). Randomisation ratio: 1:1 (clusters = EDs) Recruitment method: field workers visited each randomly selected HH, where the head of each HH was interviewed. In case the head of the HH/carer was N/A, any knowledgeable member of the HH aged > 18 years qualified for the interview. Sample size justification and outcome used: NR Sampling method: multistage stratified random cluster sampling. Within 10 community councils, 96 EDs were randomly allocated to the CGP programme or not (primary sampling units) in public lottery events that took place in each ED. EDs that were selected for the programme were paired with EDs that were not (48 pairs). Of these, 40 pairs were randomly selected. Within each selected ED, 2 villages (or clusters) were randomly selected (secondary sampling units). In every cluster, a random sample of 20 HHs (10 eligible and 10 non‐eligible at baseline) were randomly selected and interviewed. Study aim or objective: to assess the direct impact of the CGP on the well‐being of eligible HHs. To assess the indirect impact of the CGP on non‐eligible HHs (data not extracted). Study period: 2 years; September 2011–2013 Unit of allocation or exposure: EDs (HHs within these were then selected based on eligibility criteria).
Participants	Baseline characteristics Intervention or exposure Age: mean, years: 23.6 (table 9). Mean number of children (aged 0–17): 2.9. Proportion of HHs with elderly people (aged > 59 years): 39 Place of residence: rural areas Sex: female, %: 51.5 (table 9) Ethnicity and language: Sesotho Occupation: NR Education: NR SES: mean HH size: 5.8; proportion of HHs with: single orphans: 17.3; double orphans: 35.5; no able bodied adult: 23; proportion of HH heads that are: children (aged < 18 years): 0.1; elderly (aged > 59 years): 38.6; able‐bodied adult: 44.6; female: 45.6 Social capital: proportion of HHs that borrowed or received support from other family members, friends or neighbours: cash: 72.4; in‐kind: 71.2; labour: 11.4; agricultural: 47.5 Nutritional status: DDI: 4; HHs with poor food consumption: 20.9; HH food expenditure per month: LSL 553.5 Morbidities: proportion of HHs with chronically ill members: 39.9; disabled members: 21.1; proportion of HH heads that are chronically ill or disabled: 14.9; proportion of children aged 0–5 years with any illness in the last month: 38.9; mean number of days ill in the last month (children aged 0–5 years that were ill): 6.7; mean number of days ill in the last month (all children aged 0–5 years): 2.4 Concomitant or previous care: NR Control Age: mean: 24.2. Mean number of children (aged 0–17): 2.7. Proportion of HHs with elderly people (> 59 years) Place of residence: rural areas Sex: female, %: 51.9 Ethnicity and language: Sesotho Occupation: NR Education: NR SES: mean HH size: 5.5; proportion of HHs with: single orphans: 17.9; double orphans: 37.3; no able‐bodied adult: 22.1; proportion of HH heads that are: children (aged < 18 years): 0; elderly people (aged > 59 years): 39; able‐bodied adult: 45; female: 49.8 Social capital: proportion of HHs that borrowed or received support from other family members, friends or neighbours: cash: 76; in‐kind: 80.1; labour: 11; agricultural: 49.7 Nutritional status: DDI: 4; HHs with poor food consumption: 20.3; HH food expenditure per month: LSL 570.1. Morbidities: proportion of HHs with chronically ill members: 38.2; disabled members: 18.1; proportion of HH heads that were chronically ill or disabled: 14.4; proportion of children aged 0–5 years with any illness in the last month: 36.7; mean number of days ill in the last month (children aged 0–5 years that were ill): 7.6; mean number of days ill in the last month (all children aged 0–5 years): 2.7 Concomitant or previous care: NR Overall: NR Inclusion criteria: ultra‐poor HHs with any child aged 0–17 years in 10 selected community councils spread across 5 districts. Ultra‐poor HHs were identified by members of their community and by collection of proxy indicators of HH wealth, captured in the National Information System for Social Assistance (NISSA – HHs had to be categorised as NISSA 1 or NISSA 2) Exclusion criteria: NR Pretreatment: baseline differences between HHs in the intervention and control group included number of children aged 0–5 years (P < 0.01), females aged 18–59 years (P < 0.05), and proportion of HHs that borrowed or received support from other family members, friends or neighbours (P < 0.05) (data included eligible and non‐eligible HHs in treatment and control groups – not disaggregated for only eligible HHs). Attrition per relevant group: intervention group: 5% (41/747); control group: (12%; 92/739); main reason for LTFU: moved outside the cluster. Description of subgroups measured and reported: none reported Total number completed and analysed per relevant group: total: 1353 HHs (91%); intervention group: 706 HHs; control group: 647 HHs. Only 75% of children were in both baseline and follow‐up surveys (attrition: 25%) Total number enrolled per relevant group: intervention group: 747 HHs; control group: 739 HHs Total number randomised per relevant group: total: 48 EDs in 5 districts. Intervention: 24 EDs; control: 24 EDs.
Interventions	Intervention characteristics Intervention or exposure Food access intervention category: increase buying power Intervention type: UCT Description: UCT targeted to poor and vulnerable HHs. Provided regular transfer of LSL 360–7506 every quarter. Transfer value for CGP was originally set at a flat rate of LSL 120 (USD 12) per month per HH and was disbursed every quarter. Effective from April 2013, the cash transfer was indexed to number of children: 1. HHs with 1 or 2 children LSL 360 quarterly; 2. HHs with 3 or 4 children LSL 600 quarterly; and 3. HHs with ≥ 5 children LSL 750 quarterly Duration of intervention period: 24 months. HHs enrolled in July and August 2011 and the first payments started in September 2011. Quantitative panel HH survey with a baseline in 2011 and a follow‐up in 2013, in control and intervention locations and covering CGP eligible and non‐eligible HHs. Frequency: quarterly payments. However, payments were not made as scheduled; payment schedule was unpredictable and the transfers were made in more irregular disbursements than expected. Based on the administrative records, the mean number of payments received per CGP beneficiary HH in the sample was 6 to 7, while based on the operational design, the intended number of quarterly payments should have been 8. Number of study contacts: 2. Baseline survey in June–August 2011 and follow‐up survey in June–August 2013 Providers: programme run by the Ministry of Social Development of the government of Lesotho, with financial support from the European Commission and technical support from UNICEF‐Lesotho. In the pilot stage, technical assistance to the implementation was provided by Ayala Co. and World Vision. Delivery: a cash‐in‐transit firm provided payments at pay points in each community. Majority of HHs received the total intended amount of funds, despite irregular payments (mean of 6.6 instead of 10 payments made per HH). Co‐interventions: a Food Emergency Grant was also disbursed to CGP beneficiaries in 2012 and 2013. A bi‐monthly top‐up of LSL 400 (LSL 200/month) that was disbursed together with the CGP, but in a separate envelope. Resource requirements: respondents reported spending on average around 3 hours travelling to and from the pay point on pay days (return journey on foot). Almost all the respondents walked to the pay point where they on average spent 2.3 hours waiting. On average, respondents spent LSL 9 to collect the payment. Economic indicators: costs of overall programme reported elsewhere but for a different time period. Control: no intervention
Outcomes	Real monthly total consumption expenditure; monthly amount spent on food Food security: proportion of children aged 0–5 and 6–17 years with severe food deprivation (FSI); number of months in which HHs had sufficient/some shortage/extreme shortage food to meet their needs Dietary diversity: DDI; FCS; proportion of HHs with poor/borderline/acceptable food consumption Anthropometry: weight; underweight Morbidity: proportion of children ill in previous month; mean number of days children ill in previous month. Adverse event: overweight
Identification	Sponsorship source: Oxford Policy Management (OPM) was contracted by UNICEF to design and undertake an independent evaluation of Round 2 Phase 1 of the CGP pilot. Country: Lesotho Setting: rural HHs Author's name: Luca Pellerano Email: luca.pellerano@opml.co.uk Declarations of interest: NR Study or programme name and acronym: Lesotho Child Grants Programme (CGP) Type of record: report
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	Low risk	Randomisation took place through public lottery events in each community council.
Allocation concealment (Selection bias)	Low risk	Allocation was at the ED level, in public lottery events.
Baseline characteristics similar (Selection bias)	Low risk	A number of baseline differences reported between HHs in the intervention and control group despite the matching of electoral districts (primary sampling unit), followed by random sampling of clusters (secondary sampling units). These include number of children aged 0–5 years (P < 0.01), women aged 18–59 years (P < 0.05), price of rubber boots in the community (P < 0.05), and proportion of HHs that borrowed or received support from other family members, friends or neighbours (P < 0.05). The study authors used the DID method and adjusted for baseline imbalances in their analyses.
Baseline outcome measurements similar (Selection bias)	Low risk	HHs in both groups were similar at baseline in terms of outcome measurements such as HH food consumption, HH food security and expenditure on food.
Blinding of participants and personnel (Performance bias)	Low risk	No blinding possible. This was unlikely to introduce performance bias.
Blinding of outcome assessment (Detection bias)	High risk	Outcomes were measured by self‐report in questionnaire. Self‐reported outcomes could have been influenced by knowledge of treatment allocation.
Protection against contamination (Performance bias)	Low risk	CGP administrative records indicated that no eligible HHs in control areas received the intervention.
Incomplete outcome data (Attrition bias)	Low risk	Differential attrition (12% in control group; 8% in intervention group), mainly due to more HHs in the control group that moved outside their clusters. The study authors adjusted sampling weights for selective non‐response in their analysis by calculating the probability of HHs being retained in the sample on the basis of key HH characteristics at baseline. Loss of clusters (cRCT): low risk. No loss of complete clusters (villages) reported.
Selective outcome reporting (Reporting bias)	Unclear risk	No study protocol available.
Other bias	Unclear risk	Misclassification bias: low risk. Measurement bias: low risk. Trained fieldworkers. Used a 7‐day dietary recall 8 food groups at baseline and follow‐up. Incorrect analysis: low risk. Estimates adjusted for clustering. Recruitment bias: low risk. Villages (clusters) were sampled before randomisation of electoral districts to the intervention or control group. Seasonality bias: low risk.