Hunter, Ruth F., Adlakha, Deepti, Cardwell, Christopher, Cupples, Margaret E., Donnelly, Michael, Ellis, Geraint, Gough, Aisling, Hutchinson, George, Kearney, Therese, Longo, Alberto, Prior, Lindsay, McAneney, Helen, Ferguson, Sara, Johnston, Brian, Stevenson, Michael, Kee, Frank, and Tully, Mark A.
Background: Evidence for the health benefits of urban green space tends to stem from small, short-term quasi-experimental or cross-sectional observational research, whilst evidence from intervention studies is sparse. The development of an urban greenway (9 km running along 3 rivers) in Northern Ireland provided the opportunity to conduct a natural experiment. This study investigated the public health impact of the urban greenway on a range of physical activity, health, wellbeing, social, and perceptions of the environment outcomes. Methods: A repeated cross-sectional household survey of adult residents (aged ≥16 years) who lived ≤1-mile radius of the greenway (intervention sample) and > 1-mile radius of the greenway (control sample) was conducted pre (2010/2011) and 6-months post implementation (2016/2017). We assessed changes in outcomes pre- and post-intervention follow-up including physical activity behaviour (primary outcome measure: Global Physical Activity Questionnaire), quality of life, mental wellbeing, social capital and perceptions of the built environment. Linear regression was used to calculate the mean difference between post-intervention and baseline measures adjusting for age, season, education, car ownership and deprivation. Multi-level models were fitted using a random intercept at the super output area (smallest geographical unit) to account for clustering within areas. The analyses were stratified by distance from the greenway and deprivation. We assessed change in the social patterning of outcomes over time using an ordered logit to make model-based outcome predictions across strata. Results: The mean ages of intervention samples were 50.3 (SD 18.9) years at baseline (n = 1037) and 51.7 (SD 19.1) years at follow-up (n = 968). Post-intervention, 65% (adjusted OR 0.60, 95% CI 0.35 to 1.00) of residents who lived closest to the greenway (i.e., ≤400 m) and 60% (adjusted OR, 0.64 95% CI 0.41 to 0.99) who lived furthest from the greenway (i.e.,≥1200 m) met the physical activity guidelines - 68% of the intervention sample met the physical activity guidelines before the intervention. Residents in the most deprived quintiles had a similar reduction in physical activity behaviour as residents in less deprived quintiles. Quality of life at follow-up compared to baseline declined and this decline was significantly less than in the control area (adjusted differences in mean EQ5D: -11.0 (95% CI − 14.5 to − 7.4); − 30.5 (95% CI − 37.9 to − 23.2). Significant change in mental wellbeing was not observed despite improvements in some indicators of social capital. Positive perceptions of the local environment in relation to its attractiveness, traffic and safety increased. Conclusions: Our findings illustrate the major challenge of evaluating complex urban interventions and the difficulty of capturing and measuring the network of potential variables that influence or hinder meaningful outcomes. The results indicate at this stage no intervention effect for improvements in population-level physical activity behaviour or mental wellbeing. However, they show some modest improvements for secondary outcomes including positive perceptions of the environment and social capital constructs. The public health impact of urban greenways may take a longer period of time to be realised and there is a need to improve evaluation methodology that captures the complex systems nature of urban regeneration. [ABSTRACT FROM AUTHOR]