Azeredo 2003.

Methods	CBA
Participants	School children in grades kindergarten to grade 5 and their families in Muskogee, Oklahoma from private and state schools located in both rural and urban settings. Number of participants: not reported. 4750 observations of children's behaviour were made and 6300 pre‐ and postintervention questionnaires were distributed.
Interventions	Intervention: 18 or 27 lesson curricula stratified by grade (30‐ to 45‐minute sessions). Smoke alarm giveaway. School cycle fairs with helmet giveaways. Safety pen pal letters. Letters to parents. Injury prevention talks at parent‐teacher meetings. Control: no injury prevention curriculum. Opportunity for smoke alarm to be installed.
Outcomes	Observed seat‐belt use of occupants in the front seat of a vehicle and cycle helmet use during and 2 weeks after the intervention. Self‐reported behaviour, including driver and passenger seat‐belt use and cycle helmet use. Safety knowledge measured using written questions for children in kindergarten to grade 1, and true or false and multiple choice questionnaire for children in grades 2 and 3 and grades 4 and 5.
Injury mechanisms	Vehicle safety; smoke alarms and fire; cycle safety helmet use; brain and spinal cord injuries; home safety; pedestrian safety; first aid; traffic signs and signals; intersections and railroad crossings; water safety.
Notes	Did not present the characteristics of the control and intervention groups. Measurement of observed seat‐belt use 3 months' post‐intervention did not occur as many of the students had emigrated or immigrated, leading to a change in the population.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation to intervention/control (selection bias) (for non‐RCT and CBA studies)	High risk	Schools chose to be in the intervention group.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Self‐reported outcomes.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Number allocated at baseline not reported.
Selective reporting (reporting bias)	High risk	Seat‐belt use outcome reported incompletely.
Other bias	Unclear risk	No baseline data presented.
Risk of bias due to confounding (for non‐RCTs and CBA studies)	High risk	Did not adjust for confounding. Pre‐programme, more control school pupils used seat belts.