There were 10 studies of the effects of school environment interventions that did not include health education or services as components: six randomised controlled trials and four non-randomised studies.
Most of the studies had methodological limitations; the quality of the studies was considered to be generally poor, particularly in terms of attrition and adjustment for clustering in the analysis and small sample sizes. Compared to standard practice, interventions that encouraged staff and students to develop school climates characterised by a stronger sense of community and better relationships seemed to show benefits for some but not all health outcomes.
The most methodologically sound evaluation reported the fewest significant benefits. Overall, the authors stated that the studies suggested potential for school environment interventions to bring about benefits, particularly for measures concerned with violence and aggression. There was no evidence of any interventions causing harms.
Two randomised controlled trials of interventions that enabled students to advocate for changes in school catering and physical activity reported benefits for physical activity but not diet. Three evaluations of improvements to school playgrounds offered weak evidence of effects on physical activity.
There were four studies of feasibility and acceptability of school environment interventions. All four studies described their sampling methods, two defined a clear question and all four reported quantitative data; one also reported qualitative findings. Results were mainly positive for process evaluations in terms of intervention feasibility, fidelity, reach and acceptability; differences in methods precluded a comparison of the delivery and uptake of interventions.
There 19 studies of processes through which school-level influences could occur. Fourteen studies were considered high quality. Studies focused on aspects of the school environment that most concerned students, including the physical environment, lack of teacher involvement in pastoral care and outsourcing of discipline to security guards, and new technologies; there was little about how schools were organised and managed and how teaching was delivered.
Schools in which students felt educationally marginalised or unsafe seemed to have increased school disengagement and overall a feeling of lack of safety. A strong positive teacher-student relationship seemed to benefit student well-being and avoidance of risk behaviour. It seemed that organisational features of secondary schools in USA and UK undermined teacher-student relationships so that students felt a loss of control within their school community and this increased the likelihood that they will look for a sense of identity and social support through some of the risk behaviours. The extent to which schools protected and engaged students and ensured that teachers fulfil a broader role in young people’s development seemed to be important; there were higher rates of violence, drug use and other risk behaviours where there were socially and economically disadvantaged families and communities, inequities in school funding and high staff turnover.