Maluccio 2005.

Study characteristics
Methods	Study design: cRCT Study grouping: parallel group How were missing data handled? 90% (1581) of the stratified random sample was interviewed in the first round. In 2002, just over 90% of these were re‐interviewed. The sample for which there is a complete set of observations (1 in each of the 3 survey rounds) was 1396, smaller than the 1434 shown in the first row of the third column of Table 3. The HHs were about evenly divided between intervention and control groups, indicating that at least the level of attrition was not significantly different between them. Similarly, when the sample was limited to those interviewed in all 3 rounds as a partial control for attrition bias, estimated effects changed only slightly, with no systematic bias. Another partial remedy to control for attrition bias is to estimate a HH fixed‐effects model, particularly if one suspects that unobserved persistent heterogeneity is leading to attrition. However, as with the other robustness checks, when the models were estimated with these controls, the results differed little. The number of HHs was about evenly divided between intervention and control groups, suggesting that attrition was not significantly different between groups. Combining this with the evidence from the robustness checks just described, it was concluded that attrition bias was not driving the results presented. Randomisation ratio: 1:1 Recruitment method: NR Sample size justification and outcome used: 42 HHs were randomly selected in each comarca using a census carried out by RPS 3 months prior to the survey as the sample frame, yielding an initial target sample of 1764 HHs. The sample size calculation was based on assessing the necessary sample sizes for the indicators listed in Appendix B, Table 26. Assuming a random sample, the indicator that required the largest sample size, using a significance level of 5% and a power of 80%, was enrolment for Grades 1–4 (indicator 5 in Appendix B, Table 26). To detect a minimum, statistically significant difference of 8 pps between intervention and control groups, a sample size of 549 students for each group was required. Not all HHs had children in this age group. According to the 2000 RPS population census, 63% of HHs had ≥ 1 child aged 6–12 years. Therefore, to obtain a sample of 549 children (in different HHs), it was necessary to interview 871 HHs in each group (549 divided by 0.63) or 1742 in total. Thus, the study authors arrived at a target sample of 1764 HHs. Sampling method: in the design phase of RPS, rural areas in all 17 departments of Nicaragua were eligible for the programme. In addition, these departments had easy physical access and communication (including being < 1‐day drive from the capital, Managua, where RPS is headquartered), relatively strong institutional capacity and local co‐ordination, and reasonably good coverage of health posts and schools. By purposively targeting, RPS avoided devoting a disproportionate share of its resources during the pilot to increasing the supply of educational and health services. In the next stage of geographic targeting, all 6 (out of 20) municipalities that had the participatory development programme Microplanificación Participativa (Participatory Micro‐planning), run by the national Fondo de Inversión Social de Emergencia (FISE), were chosen. In the last stage of geographic targeting, a marginality index based on information from the 1995 National Population and Housing Census was constructed, and an index score was calculated for all 59 rural census comarcas2 in the selected municipalities. The index was a weighted mean of the following set of poverty indicators (with respective weights in parentheses) known to be highly associated with poverty (Arcia 1999): 1. family size (10%), 2. access to potable water (50%), 3. access to latrines (30%), and 4. illiteracy rates (10%). Higher index scores were associated with more impoverished areas. Recognising that the index could not reliably distinguish between 2 comarcas with similar scores, rather than use the scores directly, the 59 rural comarcas were grouped into 4 priority levels after renormalising the highest index score to 100: a score > 85 was given highest priority (priority 1); 70–85, priority 2; 60–70, priority 3 and below 60, lowest priority. The 42 comarcas with the priority scores 1 and 2 were eligible for the pilot phase's first stage. Comarcas are administrative areas within municipalities that include 1–5 small communities averaging 100 HHs each. Study aim or objective: to determine the impact evaluation of a randomised community‐based intervention, RPS, against a broad range of outcomes related to the programme's primary objectives, including 1. HH (food) expenditures, 2. child schooling and child labour, 3. preventive health care of children aged < 5 years, and 4. nutritional status of children. Study period: baseline: 2000; follow‐up: 2001–2002 Unit of allocation or exposure: randomisation was at the comarca level. Intervention at HH level
Participants	Baseline characteristics Intervention or exposure: NR Control: NR Overall Age: children 0–5 years and 7–13 years Place of residence: rural Central Region of Nicaragua Sex: NR Ethnicity and language: NR Occupation: NR Education: NR SES: within the 42 comarcas selected for the programme evaluation, 42% of the population was extremely poor before the programme, i.e. their total expenditures were less than the amount necessary to purchase a food basket providing minimum caloric requirements (World Bank 2003) and 80% extremely poor or poor. Social capital: NR Nutritional status: NR Morbidities: NR Concomitant or previous care: participatory development programme Microplanificación Participativa (Participatory Micro‐planning), run by the national Fondo de Inversión Social de Emergencia (FISE), were chosen. The goal of that programme was to develop the capacity of municipal governments to select, implement and monitor social infrastructure projects such as school and health post construction, with an emphasis on local participation. Inclusion criteria: in the design phase of RPS, rural areas in all 17 departments of Nicaragua were eligible for the programme. The focus on rural areas reflects the distribution of poverty in Nicaragua of the 48% of Nicaraguans designated as poor in 1998, 75% resided in rural areas. For the pilot, the Government of Nicaragua selected the departments of Madriz and Matagalpa from the northern part of the Central Region, on the basis of poverty as well as on their capacity to implement the programme. Table 1 of the publication summarises the eligibility requirements and demand and supply‐side benefits of RPS. Exclusion criteria: HH not extremely poor satisfying 1 or both of the following: 1. own a vehicle, truck, pickup truck or jeep; 2. own > 20 manzanas (14.1 hectares) of land. Based on these criteria, 169 HHs (2.9% of HHs living in the intervention areas as reported in the May 2000 RPS census population) were excluded from the programme. In addition to these HHs, 219 (3.8%) HHs were excluded after the orientation assemblies and programme registration for ≥ 1 of the following reasons: 1. HH comprising a single man or woman who was not disabled, 2. HH with significant economic resources or a business, 3. HH that omitted or falsified information during the RPS population census. Finally, 240 (4.2%) HHs did not attend the orientation assembly or chose not to participate. Pretreatment: there were few significant differences between HHs (or individuals) in intervention and control groups at baseline. Differences in baseline 2000 study: RPS mean effect on annual total HH: intervention: NIO 20,903 control; NIO 20,695 in control Attrition per relevant group: 90% (1581) of the stratified random sample was interviewed in the first round (see Table 3 of publication). In a few comarcas, the coverage was 100%, but in 6, it was < 80%. For the follow‐up surveys in October 2001 and October 2002, the target sample was limited to these 1581 first‐round interviews. In 2002, just over 90% of these were re‐interviewed, on a par with surveys of similar magnitude in other developing countries. Again, however, coverage in 6 of the comarcas was substantially worse, with < 80% successfully re‐interviewed (and 1 was 1 of the 6 from above with high first‐round non‐response rate). This attrition was unlikely to have been random. Description of subgroups measured and reported: Grades 1, 2, 3 and 4; extreme poor, poor and non‐poor; children aged 0–3 years, 12–23 months; HHs with children aged 0–5 years; HHs with children aged 7–13 years; Total number completed and analysed per relevant group: total completed interview: 1581 in 2000, 1490 in 2001 and 1434 in 2002. Total completed interview in all 3 rounds: 1396. Total number enrolled per relevant group: total enrolled: 1764 in 2000, 1581 in 2001 and 1581 in 2002. Total number randomised per relevant group: evaluation design was based on a randomised, community‐based intervention with measurements before and after the intervention in both treatment and control communities. One‐half of the 42 comarcas (targeted in the first stage) were randomly selected into the programme. Thus, there are 21 comarcas in the intervention group and 21 in the control group.
Interventions	Intervention characteristics Intervention or exposure Food access intervention category: increase buying power Intervention type: CCT Description: modelled after PROGRESA, RPS is designed to address both current and future poverty via cash transfers targeted to HHs living in poverty in rural Nicaragua. The transfers were conditional, and HHs were monitored to ensure that children were, among other things, attending school and making visits to preventive healthcare providers. RPSs specific objectives included: supplementing HH income for up to 3 years to increase expenditures on food, reducing school desertion during the first 4 years of primary school, and increasing the health care and nutritional status of children aged < 5 years. Duration of intervention period: 2 years Frequency: every other month Number of study contacts: 3 (baseline, follow‐up 2001 and follow‐up 2002) Providers: IADB loan financing the project and the Government of Nicaragua and RPS provided the service. IFPRI conducted the quantitative impact evaluation Delivery: to ensure adequate supply, RPS trained and paid private providers to deliver the specific healthcare services required by the programme. Cash transfer delivery method NR. Co‐interventions: none reported Resource requirements: to ensure adequate supply, RPS trained and paid private providers to deliver the specific healthcare services required by the programme. These services provided free of charge to beneficiary HHs, included growth and development monitoring; vaccination; and provision of antiparasites, vitamins, and iron supplements. The monitoring was done using the MIS designed specifically for and by RPS. It comprises a continuously updated, relational database of beneficiaries, healthcare providers and schools. The MIS is also used to 1. select beneficiaries and prepare invitations to programme incorporation assemblies, 2. calculate transfer payments, 3. compile requests to the Ministry of Health for vaccines and other materials, and 4. monitor whether service providers were meeting their responsibilities. Decision rules capturing the requirements were programmed directly into the MIS. Substantial time was dedicated to designing data forms for the various programme participants that fed into this system (including the HH registry or census forms, school forms, and healthcare provider forms that were all sent to the main office where they were entered into the computer). Economic indicators: RPS comprised 2 phases over 5 years, starting in 2000. The pilot phase (also known as Phase I) lasted 3 years and had a budget of USD 11 million, representing approximately 0.2% of GDP or 2% of annual recurring government spending on health and education. The value of the supply‐side services, as measured by how much RPS paid to the providers, was also substantial. On an annual basis, the education workshops cost approximately USD 50 per beneficiary and the health services for children aged < 5 years, approximately USD 110, including the value of the vaccines, antiparasites, vitamins and iron supplements, all of which were provided by the Ministry of Health. To enforce compliance with programme requirements, beneficiaries did not receive the food or education component of the transfer if they failed to carry out any of the conditions. Control: no intervention
Outcomes	Proportion of HH expenditure on food: per capita food expenditure (annual); percentage of HH food expenditure; food expenditure for different food groups/items Anthropometry: underweight (WAZ < –2SD); stunted (HAZ < 2SD); wasted (WHZ < –2SD); HAZ
Identification	Sponsorship source: IFPRI Country: Nicaragua Setting: 42 rural comarcas areas in rural departments of Madriz and Matagalpa in the northern part of the Central Region in Nicaragua Authors' names: John A Maluccio and Rafael Flores Email: NR Declarations of interest: NR Study or programme name and acronym: Red de Protección Social (RPS) or 'Social Safety Net' Type of record: report
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	Low risk	Randomisation done through lottery process. Given the geography of the programme area, control and intervention comarcas were at times adjacent to one another. The selection was done at a public event with representatives from the comarcas, the Government of Nicaragua, IADB, IFPRI and the media present. The 42 comarcas were ordered by their marginality index scores and stratified into 7 groups of 6 each. Within each stratum, randomisation was achieved by blindly drawing 1 of 6 coloured balls (3 blue for intervention, 3 white for control) from a box after the name of each comarca was called out. Thus, 3 comarcas from each group were randomly selected for inclusion in the programme, while the other 3 were selected as controls. The survey sample is a stratified random‐sample at the comarca level from all 42 comarcas. The areas represented comprise a relatively poor part of the rural Central Region in Nicaragua, but the sample is not statistically representative of the 6 municipalities (or other areas of Nicaragua, for that matter). 42 HHs were randomly selected in each comarca using a census carried out by RPS 3 months prior to the survey as the sample frame, yielding an initial target sample of 1764 HHs.
Allocation concealment (Selection bias)	Low risk	Allocation was at the comarca level at the beginning of the study. Randomisation process was done at a public event with representatives from the comarcas, the Government of Nicaragua, IADB, IFPRI, and the media present. No recruitment was done after randomisation.
Baseline characteristics similar (Selection bias)	Low risk	Baseline outcome data are presented. On the whole, there were few significant differences between HHs (or individuals) in intervention and control groups at baseline.
Baseline outcome measurements similar (Selection bias)	Low risk	Quote: "… double‐difference estimates of the effects of the program presented later in this report all show differences at baseline for the entire range of outcomes analyzed. In no instance were any of those measures significantly different at baseline."
Blinding of participants and personnel (Performance bias)	Low risk	Although authors stated that, "Within each stratum, randomisation was achieved by blindly drawing one of six colored balls (three blue for intervention, three white for control) from a box after the name of each comarca was called out." Unclear who was blinded.
Blinding of outcome assessment (Detection bias)	High risk	No information on whether or not the assessments were done blindly but some outcomes were self‐reported which could have been influenced by lack of blinding.
Protection against contamination (Performance bias)	Unclear risk	Although this study had a community cluster randomised design, the clusters in the intervention and control comarcas were at times adjacent to one another. A HH may be a beneficiary while its neighbour is a non‐beneficiary, particularly in a few cases where boundaries such as roads divide 2 comarcas. Seeing the activity and the emphasis placed on the RPS objectives may lead non‐beneficiaries to undertake behaviour they would not have otherwise.
Incomplete outcome data (Attrition bias)	Low risk	Quote: "We now document non‐response in the 2000 baseline survey and attrition in the follow‐up surveys. Overall, 90 percent (1,581) of the stratified random sample was interviewed in the first round (see Table 3)." The authors also state that, "Recall that the number of households is about evenly divided between intervention and control groups, suggesting that attrition was not significantly different between intervention and control groups. Combining this with the evidence from the robustness checks just described, we conclude that attrition bias is not driving the results presented here."
Selective outcome reporting (Reporting bias)	Unclear risk	No protocol is available
Other bias	Unclear risk	Misclassification bias: unlikely. Measurement bias: unlikely. Incorrect analysis: unlikely. Seasonality bias: low. Authors statistically controlled for seasonality.