Splett 2006.

Methods	Included as outcome evaluation and process evaluation Study design: clustered parallel‐group design, randomised at the school level Setting: K‐8 schools in Minneapolis, Minnesota, USA Period: in 2000 and 2001, the HLAI was implemented in schools and was tested for effectiveness; in 2001 and 2002, the HLAI was expanded to all K‐8 schools
Participants	Eligible sample frame: not reported; however 700 students with asthma were required in each group to detect a positive change in attendance Randomised: 916 in intervention schools and 645 in control schools Completed (intervention): not reported Inclusion criteria: not reported Exclusion criteria: not reported Baseline characteristics Age of children: not reported Ethnicity: 66% were African American, 6% were Hispanic, 5% were American‐Indian, and 20% were white Socio‐economic status: 73% of participants were eligible for free or reduced price lunches Gender: males represented 58% of participants Asthma status: not reported
Interventions	Intervention: in participating schools, licensed school nurses, licensed practical nurses, and health service assistants received coaching and reinforcement by asthma resource nurses. Clinics also received training on NIH guidelines and guidance on implementing standards of care for asthma. Study authors reported: "staff followed 'Core Components of Asthma Management in the School Health Office' (Core Components), including case identification, nursing care procedures, care coordination, emergency care, and student education, to provide more systematic and consistent care to students with asthma and improve communication with school staff, parents, and health care providers", although further details of the student education component were not provided Control: usual care Intensity: not reported Instructor: school nurse Theoretical framework: not reported Parental engagement: not reported Child satisfaction: not reported Timing of intervention in school day: not reported
Outcomes	Extractable outcomes were collected for: Absence from school Unplanned visit to hospital or GP due to asthma symptoms Core processes evaluated (child level): no information (other outcomes considered around sustainability)
Notes	This study conducted an ecological analysis Process evaluation category: stand‐alone, named section (2 papers) Breadth and depth: neither broad nor deep Voice of children given prominence: not featured Funding source: Member Organisations of the Healthy Learners Board, Controlling Asthma in American Cities Grant
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Study authors reported: "A random sequence of treatment codes, stratified by school system, was generated using the SAS System (Version 9.1, Cary, North Carolina) by the statistician"
Allocation concealment (selection bias)	Low risk	Centrally allocated: study authors reported: "A random sequence of treatment codes, stratified by school system, was generated using the SAS System (Version 9.1, Cary, North Carolina) by the statistician"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No measures were described as implemented around blinding
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No measures were described as implemented around blinding
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No direct reports describe attrition
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting was found
Other bias	Unclear risk	Missingness ‐ low risk ‐ all those who were followed up submitted information Baseline imbalance ‐ low risk ‐ no evidence of baseline imbalances between intervention and control groups Risk of contamination ‐ high ‐ children were the unit of randomisation; potential was present for children with different treatment allocations to share materials/information, etc
Transparent and clearly stated aims	Low risk	Study aims were clearly stated
Explicit theories underpinning and/or literature review	High risk	No theoretical framework or supporting literature was provided
Transparent and clearly stated methods and tools	Low risk	Methods and tools were clearly stated
Selective reporting	High risk	A focus group is mentioned in the DuPlessis paper, but whether this occurred before or after or during the intervention is not clear. In addition, relevant results were not presented
Harmful effects	High risk	Very broad study; harmful effects were not directly considered
Population and sample described well	High risk	Some information is missing, including age of participants
Continuous evaluation	Low risk	Process evaluation data were apparently collected throughout
Evaluation participation equity and sampling	High risk	No information was collected from nurses
Design and methods overall approach	Unclear risk	Some core elements are missing
Tools and methods of data collection reliable/credible	Low risk	Based on administrative records ‐ straightforward constructs
Tools and methods of data analysis reliable/credible	Unclear risk	Lack of age data makes it difficult to interpret some outcomes, although the models include controls for age
Performance bias/neutrality/credibility/conformability	Low risk	As based on administrative records, little reason was provided to assign anything but low risk of bias
Reliability of findings and recommendations	Unclear risk	Reliability of findings was compromised by the K‐12 age group
Transferability of findings	High risk	Details are lacking, and standardisation introduced difficulties related to transferability ‐ particularly the wide age range
Overall risk of bias of process evaluation	High risk	Details around the ages of children and other key factors that could influence outcomes are lacking