This study investigated the association between neighbourhood green space and participation in physical activity among elderly men with different levels of lower extremity physical function. Our first main finding is that the amount of neighbourhood green space was significantly associated with regular physical activity, regardless of lower extremity physical function level. The likelihood of regular physical activity was higher for those with good lower extremity physical function than for those with poor lower extremity physical function.
One possible underlying mechanism for this association is the opportunity that green space gives for increased physical activity. For example, people are more likely to walk, exercise and tend the garden in a greener environment. Similar observations were obtained among younger adults (aged 26-58 years) in England: people living in the greenest quintile Middle Layer Super Output Area (MSOA) were more likely to achieve the recommended amount of physical activity through different forms of physical activity, including gardening, DIY and occupational physical activity, compared to those living in the least green quintile .
Second, we observed a statistically significant interaction between variation in neighbourhood vegetation and lower extremity physical function, as they related to regular physical activity. Living in a neighbourhood with high variation in neighbourhood vegetation was significantly and negatively associated with participation in regular physical activity, but this adverse effect of high variation neighbourhoods was significantly reduced for subjects reporting good lower extremity physical function, compared to those reporting poor lower extremity physical function.
The positive and negative effect (for good and poor lower extremity physical function, respectively) of variation in neighbourhood vegetation on participation in physical activities was surprising. Low variation in neighbourhood vegetation suggests that most vegetation within a neighbourhood has a similar NDVI value. For example, the majority of vegetation would either have low NDVI values (e.g. grass field, shrub) or high NDVI values (e.g. dense trees, forests, and woodland). In contrast, high variation indicates that mixed vegetation (both high and low NDVI values) exists within the neighbourhood. It is possible that living in a neighbourhood with more homogeneous vegetation (i.e. low variation) is associated with fewer obstacles and challenges for those reporting poor lower extremity physical function to take part in physical activity, whereas high variation in vegetation might be more attractive for those reporting good lower extremity physical function.
Our results are consistent with the findings from three other studies which found that the effect of the neighbourhood environment on physical activity levels among older adults varies according to their level of lower extremity physical function [31–33]. If people are more likely to walk, exercise, garden and DIY in a greener environment, neighbourhood green space may have a role in supporting and maintaining active ageing for those individuals with limited physical function, as the provision of green space is modifiable through the planning process. This has implications for planning policy to design, preserve, facilitate and encourage the use of green space near home. More importantly, green space provision could have a more significant equity impact on those with different levels of lower extremity physical function. In addition, our results showed that relative to general physical and mental health, lower extremity physical function is an important factor associated with regular participation in physical activity. This is consistent with previous finding that better functional status was associated with greater physical activity in older age [52–57].
The strengths of this study include the use of a well-characterized established cohort with careful face-to-face measurement of all the variables. We used an objective measurement of neighbourhood green space at a fine spatial scale (3 meters). The green space in our analysis included both those formal and informal public green spaces (i.e. parks, playing fields, woods, road-side trees and roundabouts), and also those in private domains (i.e. private gardens). This is because the factor analysis of 22 activities suggested that gardening was in the same factor with the other physical activities and we recognised that gardening is an activity which could take place in a private space.
Limitations of this study include the cross-sectional study design which cannot establish causality. We used a self-reported frequency of physical activity and were limited by treating physical activity participation as a binary variable. All three activities have been treated equally although the intensity of activities may have varied from light-intensity at short duration to high-intensity at long duration. This could be further improved by applying pedometers and accelerometers to measure physical activity more accurately in future studies. The postcode centroids rather than the actual address were used to present participants’ home locations in the analysis, due to the availability of the data. The 400 meters radius buffer around postcode centroids might also not match exactly with the definitions of a neighbourhood as judged by an individual. Future studies could combine objective and perceived definitions of individual’s neighbourhood in their analyses. The study only focused on men and the associations might be different in women. There was a two-year mismatch on the data period of neighbourhood green space (2006) and the survey data (2004) due to green space data availability. However, it is highly unlikely that green space in Caerphilly changed significantly between those two years. First, the population of Caerphilly borough was estimated at 170,800 in mid-2004, only 300 fewer than in mid-2006 . Such little variation suggested Caerphilly has been a relatively stable town characterized by low-density. Second, the studied area has not been subjected to any large-scale greening policy during those two years.
Finally, we lacked information on traffic and the quality of neighbourhood green space. It is possible that areas with less green space have more traffic, which may hinder older adults from participation in physical activity through visiting parks and open space. This needs to be investigated in future studies. Previous studies suggested that the quality of green space (e.g. aesthetic features) may play a role in the association , and people living in the most deprived areas in the UK are more likely to experience the poorest quality [60, 61]. Arguably, we used urban/non-urban classification and area deprivation as a rather crude proxy for the quality of green space as with previous analyses and adjusted for these in the models.