Leroy 2008 (PROGRESA).

Study characteristics
Methods	Study design: PCS How were missing data handled? children LTFU or with missing data were excluded from the analysis. Randomisation ratio: N/A Recruitment method: mass media advertised the programme and invited families to solicit a socioeconomic screening questionnaire at the enrolment centres. Assessment utilised HH assets, housing quality, years of education and HH composition. Programme staff visited all HHs that appeared eligible to validate the results of the screening questionnaire. Eligible applicants subsequently had to return to the programme office to register. HH enrolment in urban areas started in 2002 and was staged over 2 years. Sample size justification and outcome used: NR Sampling method: For the evaluation sample, 149 manzanas (the smallest administrative unit within an urban area) in 17 of Mexico's 31 states were selected through probabilistic stratified sampling from the pool of localities where Oportunidades would be implemented in 2002. The localities were selected based on density of low‐income HHs. The survey included eligible HH that enrolled and eligible HHs that did not enrol in the programme. Study aim or objective: to evaluate the impact of Mexico's CCT programme, Oportunidades, on the growth of children aged 24 months living in urban areas. Study period: 2 years (baseline: 2002; follow‐up: 2004) Unit of allocation or exposure: HHs
Participants	Baseline characteristics Intervention or exposure Age: child, mean, months: 12.6 (SD 6.7) Place of residence: poor urban areas in Mexico Sex: female, %: 50 Ethnicity and language: NR Occupation: NR Education: head of HH completed primary school, %: 60 SES: rooms in house, mean, n: 1.0 (SD 0.7); HH income, mean: MXN 1540 (SD 2011.41) Social capital: NR Nutritional status: HAZ, mean: –1.29 (SD 1.36); length, mean, cm: 70.9 (SD 8.5); WHZ, mean: 0.30 (SD 1.07); weight, mean, kg: 8.62 (SD 2.05); maternal height, mean, cm: 149.1 (SD 5.6) Morbidities: NR Concomitant or previous care: NR Control Age: child, mean, months: 12.4 (SD 6.9) Place of residence: poor urban areas in Mexico Sex: female, %: 52 Ethnicity and language: NR Occupation: NR Education: head of HH completed primary school, %: 64 SES: rooms in house, mean, n: 1.0 (SD 0.7); HH income, mean: MXN 1708.33 (SD 2934.36) Social capital: NR Nutritional status: HAZ, mean: –1.40 (SD 1.16); length, mean, cm: 70.2 (SD 8.7); WHZ, mean: 0.33 (SD 1.0); weight, mean. kg: 8.46 (SD 2.08); maternal height, mean, cm: 149.8 (SD 5.5) Morbidities: NR Concomitant or previous care: NR Overall Age: NR Place of residence: poor urban areas in Mexico Sex: NR Ethnicity and language: NR Occupation: NR Education: NR SES: all HHs were within the poorest 20th percentile of the Mexican population. Social capital: NR Nutritional status: NR Morbidities: NR Concomitant or previous care: NR Inclusion criteria: poorest HHs in an urban block in urban centres (based on a cut‐off of the national HH Income and Expenditure Survey). None other reported. Exclusion criteria: NR Pretreatment: baseline characteristics did not differ between treatment and comparison HHs. However, the propensity score was substantially higher in intervention HHs (P < 0.05). The score was based on SE variables and likelihood of enrolment. Attrition per relevant group: total: 301/733 (41.1%) children (263 LTFU and 38 with missing data); intervention: 230/574 (40.1%) children (202 LTFU and 28 with missing data); control: 71/159 (44.7%) children (61 LTFU and 10 with missing data). Description of subgroups measured and reported: age of child at baseline: 0–6 months; 6–12 months and 12–24 months. Socioeconomic tertile at baseline: tertile 1, 2 and 3. Total number completed and analysed per relevant group: intervention group (HHs who enrolled in the programme): 344 children; control group (HHs who did not enrol in the programme): 88 children Total number enrolled per relevant group: total: 733 HHs; intervention: 574 children; control: 159 children Total number randomised per relevant group: N/A
Interventions	Intervention characteristics Intervention or exposure Food access intervention category: increase buying power Intervention type: CCT Description: cash transfers linked to children's school enrolment and regular school and clinic attendance. 3 types of cash transfers: scholarships linked to school attendance; money for school supplies and a cash transfer for food (the alimento). The programme also provided in‐kind health benefits (nutritional supplements for: children aged 6–23 months; low‐weight children (WAZ –1SD) aged 2–4 years, and pregnant or lactating women); and instructional meetings on health and nutrition issues. Typically, HHs received the equivalent of USD 32.5–41.3, constituting 19–24% of mean HH consumption. Duration of intervention period: 2 years (2002–2004) Frequency: monthly Number of study contacts: 2 study contacts: September–December 2002 (baseline); July–November 2004 (follow‐up). Providers: federal government of Mexico Delivery: growth monitoring and health and nutrition education components of the programme were obligatory and hence compliance was > 90%. Cash transfer to female head of HHs. Co‐interventions: none reported Resource requirements: NR Economic indicators: NR Control: no intervention
Outcomes	Anthropometry: height; weight; HAZ, WHZ
Identification	Sponsorship source: National Coordination of the Oportunidades programme of the Mexican secretary of Social Development Country: Mexico Setting: urban poor HHs participating in a national anti‐poverty programme Authors' names: Lynnette M Neufeld; first author: Jef L Leroy Email: neufeld@insp.mx Declarations of interest: no conflicts of interest Study or programme name and acronym: PROGRESA/Oportunidades 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 randomisation done. By excluding 26% of all HHs surveyed, this could have introduced other forms of bias in their sample. Study authors did not compare these HHs to included HHs.
Baseline characteristics similar (Selection bias)	Low risk	Although intervention HHs had a higher baseline mean propensity score (probability of enrolment in the Oportunidades programme) than those in the control group, the study authors used PSM to compare changes in HHs in both groups during their analysis.
Baseline outcome measurements similar (Selection bias)	Low risk	The anthropometric parameters of the children at baseline did not differ. The study authors also adjusted for baseline child anthropometric measurements and maternal height in their analyses.
Blinding of participants and personnel (Performance bias)	Low risk	Blinding not possible but lack of blinding unlikely to influence participant or personnel behaviour.
Blinding of outcome assessment (Detection bias)	Low risk	Blinding of outcome assessors NR. However, outcomes of interest (anthropometric measures) were objective and unlikely to have been influenced by lack of blinding.
Protection against contamination (Performance bias)	Low risk	Eligible HHs were enrolled in the programme and control HHs were eligible HHs that did not enrol in the programme. Therefore, it was not possible for control HHs to receive any programme benefits.
Incomplete outcome data (Attrition bias)	High risk	High levels of missing data overall (41%) and in intervention (40.1%) and control (44.7%) groups. Although for most HHs reasons for LTFU were because of moving to another area, this information was only available for some of these HHs (91/263). Furthermore, by excluding 26% of all HHs surveyed this could have introduced other forms of bias in their sample. The study authors did not compare these HHs to included HHs.
Selective outcome reporting (Reporting bias)	Unclear risk	No protocol available.
Other bias	Low risk	Misclassification bias: low risk. The study authors only used data from included HHs that were consistently classified as either intervention or control HHs. Measurement bias: low risk. Incorrect analysis: N/A. Seasonality bias: unknown risk (time of year not stated).