Bartholomew 2006.

Methods	Included as outcome evaluation Study design: clustered parallel‐group design with schools selected as the unit of randomisation Setting: study was conducted across inner city elementary schools in Texas, USA. All schools from a single district were invited to participate, with 84 agreeing to participate. 60 schools were selected because two‐thirds of students were in receipt of free school meals Period: study conducted and data collected from 1997 to 2000
Participants	Eligible sample frame: 982 pupils eligible and consented Randomised: 946 pupils randomised at school level: 515 into treatment group and 431 into control group Completed (intervention): 503 pupils were available at follow‐up; 16 had actively withdrawn, 325 were lost to follow‐up, and 102 had graduated Inclusion criteria: not clearly stated ‐ case detection procedure implemented to discover asthmatic students Exclusion criteria: not clearly stated at child level (school‐level criteria included that schools would have two‐thirds of pupils in receipt of free school meals) Baseline characteristics Age of children: based on reports from 88.3% of pupils: mean age 7.7 years Ethnicity: based on reports from 88.3% of pupils: 45% of children were African American, 51% Hispanic, 3% Caucasian, and 1% from other ethnic groups Socio‐economic status: most students were from households with incomes < $20,000 per year; 28% reported incomes < $9999 Gender: based on reports from 88.3% of pupils: 400 girls (47.9%) and 435 boys (52.1%) Asthma status: indicators of asthma severity not provided
Interventions	Intervention: multi‐component intervention involving direct delivery to children, care providers, and parents/guardians. Children received self‐management education through the "Watch, Discover, Think, and Act" interactive computer programme, which was based on National Institutes of Health guidelines and pedagogical and self‐management theories. Children participated in the intervention for approximately 1 year, during which time they played the computer game and their parents received training materials on managing asthma. School nurses received training to improve communication with community primary care providers and to encourage greater monitoring of children's asthma status. Children with persistent asthma in 15 schools received an enhanced intervention, which involved meetings with a project physician to develop an Asthma Action Plan and receipt of a month's supply of medication, with the plan sent to the child's community primary care provider. Researchers also assessed the quality of the school environment with regards to asthma triggers in intervention schools and communicated these findings with recommendations to schools Control: not clearly stated (usual care) Intensity: not clearly described: children were involved in the intervention over the course of a year, and all children completed all levels of the intervention computer programme, although patterns and intensity of usage were not described Instructor: main standardised instruction provided through a computer programme; other components involved school nurses Theoretical framework: a logic model was provided and the computer programme was reported as being based on social cognitive theory Parental engagement: not reported directly. Problems reported in gaining consent: "only about 64% of parents returned case detection surveys, and about half of the families of children with probable asthma agreed to participate in the study" Child satisfaction: not reported Timing of intervention in school day: not reported
Outcomes	Withdrawal from study
Notes	Several primary outcomes of interest were collected but were not disaggregated by treatment status (exacerbations leading to admission to hospital and absence from school) or were not presented in the results (asthma symptoms leading to emergency department visits) Also considered for possible inclusion as a process evaluation ‐ while the study purports to include a process evaluation section, this was not deemed to include the core components of a process evaluation (i.e. at least a partial focus on implementation outcomes and the relationship with context) Funding source: National Heart, Lung, and Blood Institute, National Institutes of Health
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information was provided on how randomisation occurred, except schools were randomly assigned to treatment and control groups
Allocation concealment (selection bias)	Unclear risk	No information was provided on how allocation of schools occurred, except that schools were randomly assigned to treatment and control groups Also little detail on how schools were additionally allocated to enhanced intervention: in 15 schools, an enhanced intervention allowed children and their parents to meet with a project physician, develop an asthma action plan, and receive a 1‐month supply of medication; the project physician then followed up with the child's community physician
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No measures to limit performance and detection biases were described (not considered by trial authors)
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No measures to blind outcome assessment were described (not considered by trial authors)
Incomplete outcome data (attrition bias) All outcomes	High risk	Very high levels of attrition were noted in both outcome and control groups: 43% in control and 49.5% in treatment at post test. Some evidence provided by study authors indicated that "attrition did not create any significant group differences in the variables measured in the study sample"
Selective reporting (reporting bias)	High risk	Some outcomes were not reported or results were not disaggregated by intervention and control (e.g. levels of hospitalisation)
Other bias	High risk	Missingness ‐ children were not tracked from school to school because of problems with treatment and control group migration. Around 10% of children had missing data at baseline or were not considered to have "usable data", and no imputation strategies were described Baseline imbalance ‐ no data on baseline demographic characteristics were given to illuminate the split between control and intervention groups. No differences between groups were found in health status variables, school performance, attendance, or levels of environmental allergens in schools Risk of contamination ‐ low ‐ randomisation occurred at the school level
Transparent and clearly stated aims	Unclear risk	N/A
Explicit theories underpinning and/or literature review	Unclear risk	N/A
Transparent and clearly stated methods and tools	Unclear risk	N/A
Selective reporting	Unclear risk	N/A
Harmful effects	Unclear risk	N/A
Population and sample described well	Unclear risk	N/A
Continuous evaluation	Unclear risk	N/A
Evaluation participation equity and sampling	Unclear risk	N/A
Design and methods overall approach	Unclear risk	N/A
Tools and methods of data collection reliable/credible	Unclear risk	N/A
Tools and methods of data analysis reliable/credible	Unclear risk	N/A
Performance bias/neutrality/credibility/conformability	Unclear risk	N/A
Reliability of findings and recommendations	Unclear risk	N/A
Transferability of findings	Unclear risk	N/A
Overall risk of bias of process evaluation	Unclear risk	N/A