Implementation fidelity of heat-mitigation infrastructure interventions in primary schools in selected climate-vulnerable districts of Bangladesh: A mixed-method study

Abstract

Background: Climate change is driving more frequent and intense heatwaves across South Asia, with children often experiencing the worst impacts. In low-income community settings of climate-vulnerable districts in Bangladesh, classrooms are often equipped with single-layer iron-tin roofs that frequently overheat, harming children's health, concentration, and learning. Passive cooling interventions such as solar-reflective paint (SRP) and bamboo ceilings with polyethylene foam (BCPF) are promising low-cost, electricity-free solutions. However, there is limited evidence regarding their effectiveness in reducing classroom temperatures and improving the health and well-being of children.

Objectives: The objective of this study was to assess the implementation fidelity of SRP and BCPF interventions in 69 BRAC primary schools in the Kurigram and Gaibandha districts of Bangladesh. Specific objectives were to (i) measure fidelity levels for each intervention and (ii) identify school, implementer, and context-level factors influencing fidelity.

Methods: A convergent parallel mixed-methods study was conducted nested within a pilot cluster randomised controlled trial. The pilot trial (parent study) assesses the impact of implementing the infrastructure interventions (SRP only, BCPF only, and combined) in 69 BRAC primary schools across eight sub-districts in Kurigram and Gaibandha districts of Bangladesh. The quantitative component of the study included 92 intervention-event observations (46 SRP + 46 BCPF) through structured observation checklists and a school-level questionnaire administered to monitoring officers (one per school) in 69 schools. The questionnaire and checklists were administered electronically via KoBoToolbox. Fidelity was operationalised across three dimensions, content adherence, coverage, and quality of delivery, using an adapted Carroll–Hasson fidelity framework. The qualitative component included conducting 15 key informant interviews with painters, carpenters, monitoring officers, and the project coordinator, who were purposively sampled across districts and intervention types. Quantitative analysis included descriptive statistics and hierarchical ordinary least squares (OLS) linear regression with Huber-White heteroscedasticity robust standard errors in which school readiness, implementer-level, and contextual determinants of composite fidelity were entered sequentially as nested blocks. Qualitative data were analysed thematically using a hybrid deductive–inductive approach guided by the Carroll–Hasson framework and the Theoretical Domains Framework. Integration of quantitative (fidelity dimensions and determinants) and qualitative (themes) findings followed a side-by-side joint display approach for data triangulation.

Results: Mean composite fidelity was high for both interventions (SRP: 86.2%, SD = 6.4; BCPF: 88.8%, SD = 9.9). In both interventions, the fidelity dimensions ranked in the same order. Content adherence was the highest scoring dimension (SRP: 96.7%; BCPF: 92.2%); coverage was the second highest (SRP: 85.5%; BCPF: 91.3%); and quality of delivery was lowest in both interventions (SRP: 76.6%; BCPF: 82.9%). Qualitative findings showed that intensive supervision, structured training, community engagement, and adaptive responses to weather disruption helped maintain implementation fidelity for both interventions. However, inconsistent use of personal protective equipment (PPE) is identified as a major challenge affecting the quality-of-delivery dimension. This inconsistency accounted for a lower fidelity score on this dimension.

Conclusions: SRP and BCPF were implemented with high implementation fidelity in tin-shed primary schools in climate-vulnerable districts of Bangladesh, showing potential as low-cost, electricity-free heat-mitigation strategies. Enhancing the use of PPE among intervention implementation workers remains the main area for strengthening fidelity. Key elements that supported successful implementation in this pilot, such as strong supervision, flexibility in managing structural repairs, and weather-adaptive practices, should be maintained during scale-up. The findings provide practical evidence for BRAC, the government of Bangladesh, and other partners scaling climate-resilient infrastructure. Additionally, it contributes a replicable three-dimensional fidelity assessment framework applicable to other physical-construction interventions.