Validity of results
This population-based cohort study was based on a random sample of children identified from the roster of Statistics Finland. Additionally, the relatively high rates of participation (80.3% at baseline and 63.2% at follow-ups) and the similar participant characteristics at baseline and follow-ups,18 suggest that any major selection bias was unlikely. We assessed exposure to green spaces at individual level, independently from the assessment of allergic rhinitis. We were able to use both the prenatal and life-time residential addresses in the exposure assessment, thus minimizing the likelihood of misclassification of exposure. Furthermore, we evaluated the effects of cumulative exposure to green spaces separately for spring and summer seasons. Season-specific exposures to allergenic pollen may affect the risk of allergic rhinitis differently from each other,15 which can explain the heterogeneity in associations between green spaces and allergic rhinitis at least partly. The outcome assessment was based on parental and self-administered questionnaire information on the presence and age of onset of allergic rhinitis up to 27 years and thus, this study captured the effect of prenatal and early-life exposure to green spaces across life course. Any misclassification of the outcome was independent from exposure assessment and could introduce non-differential error, potentially leading to underestimation of the studied effects. We were able to adjust for several potential confounders, including individual characteristics, socioeconomic status, and environmental exposures. This is also one of the few studies assessing the mediation and modification effects of air pollution on the association between green spaces and allergic rhinitis.34-36 Finally, sensitivity analyses were conducted to assess the impact of the selected distance buffers (100 m, 500 m, and 1000 m) and quartiles for exposure to NDVI, and their consistent results demonstrate the robustness of the study findings.
The current study had some limitations. NDVI is an objective measure used to assess green vegetation cover and density, but it does not provide qualitative information on vegetation characteristics (e.g., plant species), feature, nor on quality or accessibility to green spaces. We did not measure the pollen counts/concentrations or their quality, nor specific weather characteristics that could affect pollen and air pollutant concentrations and NDVI exposures. These may limit the assessment of potential mechanistic pathways. However, considering that the season may affect the type and the conditions of greenness/vegetation as well as the levels of pollens,37 the season-specific effect estimates indirectly provide information on the role of the quality of vegetation or on changes in the abundance of vegetation on the association between exposure to green spaces and allergic rhinitis. Additionally, similarly with other studies on neighborhood health effects, our study may not capture an individual’s true geographic context (Uncertain Geographic Context Problem), since some of mothers’ and children’s activities may take place outside of their residential environment.38