Kroeger 2002 VEN.

Methods	Trial design: Cluster‐RCT. Unit of randomization: City sectors. Number of clusters: 14. Entomological data collection: Cross sectional estimates of the density of the vectors using light traps in the main room of 565 houses for 150 nights at baseline (pre‐intervention) and during the three months after the intervention (post‐intervention). Clinical data collection: Cross sectional questionnaire survey of 569 houses with 2913 inhabitants plus examination for past or current CL (pre‐intervention) at baseline and repeated at 12 months post‐intervention. Length of follow‐up: 12 months. Analysis: Analysed at individual (population) and cluster level (sector/houses).
Participants	Baseline data on 2913 people living in 569 houses, follow‐up data on similar number. (The original sample size was 578 but 1.6% did not respond). The population was described as having moderate levels of poverty, 31% < 15 years old, 9% > 60 years old, average of 5 people per household, 21% were engaged in domestic activities, 21% were students, 13% were manual workers, self employed artisans, or secretaries, 7% were unemployed, 7% had an academic profession, and only 2% were farmers. Estimated annual incidence of leishmaniasis: 0.5% or above. Endemic disease: CL caused by unknown Leishmania spp (main vector: Lu. youngi and Lu. ovallesi).
Interventions	ITCs. The windows of all houses were covered with polyester curtains (mesh size: 0.05 mm), impregnated with lambdacyhalothrin (12.5 mg/m²; ICON 2.5CS, Syngenta, Basle) at baseline and at 6 months. Control. 7 sectors had unimpregnated curtains and 1 randomly selected sector had no curtains.
Outcomes	Number of new cases of CL assessed at 12 months. Mean number of houseflies per traps per night pre and post intervention.
Notes	Country: Venezuela (Trujillo). Trial dates: January 2000 to August 2001. Trial sponsor: Funded by the European Commission (contract Alfa Programme 6­0011­9 and INCO­DEV IC18CT 980339). The insecticide was donated by Syngenta. Sample size: Not calculated. Compliance assessment: Not reported.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"For each of the seven matched pairs we randomly allocated one sector (using computer created random numbers) to the intervention group and the other to the control group".
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) participants	High risk	"the population being “blind” towards the group allocation". One sector did not receive unimpregnated curtains so would be aware of their allocation.
Blinding (performance bias and detection bias) investigators	Unclear risk	Not reported.
Blinding (performance bias and detection bias) Assessors	Unclear risk	This was not stated.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No clusters were lost. The final number of participants were increased in 25 persons (see below). ITT analyses and dropouts per group and reasons described.
Selective reporting (reporting bias)	Unclear risk	Not reported.
Baseline measurements	Unclear risk	Only number cases (%) of CL and mean number of phlebotomine sandflies per traps captured, at baseline (other info not provided by groups).
Statistical adjustment for clustering	High risk	No statistical adjustment for clustering was performed in the primary trial. However, sensitivity analysis was done at a range of ICCs in this review and it was concluded that if the ICC had been as high as 0.05, the CIs would have crossed over 1. Statistical analysis: The trial authors used EpiInfo, SPSS, and Stata v6 for analysis. Before the main analysis: Fisher's exact tests to compare cumulative incidence between intervention and control sectors for each pair. They used cumulative incidence rates of CL and the average number of flies per trap (house) for each sector as the units of analysis. They compared data at baseline and then at follow‐up between the intervention and control groups using a paired t test, weighting the data according to the sector size. They also used Wilcoxon's matched pairs test because the small number of pairs made it difficult to assess whether the underlying distribution of the differences was normal (necessary for the validity of the t test), and the Wilcoxon test does not require this assumption. Differences rather than ratios are presented as the estimates of effect because zeroes for the main outcome, CL, precluded the use of ratios.
Other bias	Unclear risk	Trial authors declared no competing interests. The founders had no role in the trial.