Neighbourhood access to open spaces and the physical activity of residents: A national study
Introduction
Physical inactivity is a major determinant of obesity and chronic conditions such as diabetes, stroke and cardiovascular disease (Task Force on Community Prevention Services, 2002, US Department of Health and Human Services, 1996) hence the high priority placed on increasing levels of exercise in New Zealand and elsewhere (Ministry of Health, 2003). Evidence that place of residence may influence physical activity, independently of the individual characteristics of residents, has generated interest in identifying aspects of neighbourhood environments that may increase levels of physical activity (Ellaway et al., 2005, Kavanagh et al., 2005, King et al., 2005, Li et al., 2005).
A striking feature of the literature examining neighbourhood effects on physical activity is the extensive list of potential built environmental variables investigated (Li et al., 2005, Ewing et al., 2003, Frank et al., 2006, Giles-Corti et al., 2005a, Lopez-Zetina et al., 2006, Sallis et al., 2002, van Lenthe et al., 2005). To aid the systematic investigation of relationships between physical environments and physical activity, Pikora et al. (2003) proposed a framework that groups environmental factors into four categories: functional factors (traffic speed, street and path design), safety factors, neighbourhood aesthetics, and destinations (access to desired locations and amenities). The framework provides a useful starting point for exploring specific pathways through which neighbourhood context may impact on variation in levels of physical activity. Access to destinations that provide opportunities for physical activity is the built environment variable examined in this paper.
The accessibility of utilitarian destinations, such as shops, schools, and recreational amenities, as a determinant of physical activity has been indicated in a number of studies (Li et al., 2005, van Lenthe et al., 2005). There is some evidence that individual physical activity levels increase as the number or density of accessible exercise amenities increases (Diez Roux et al., 2007, Parks et al., 2003), and that the use of active modes of transport such as walking and cycling increases as distances to neighbourhood amenities decreases (Social Exclusion Unit, 2003). Self selection into neighbourhoods by people with a preference for active transport and proximate amenity access may explain some of this association (Frank et al., 2007).
Related work has shown that utilisation of recreational facilities increases as the distance between home and facilities decreases (King et al., 2005, Giles-Corti et al., 2005a, Tinsley et al., 2002) and that use of public open spaces is more sensitive to distance than other types of sporting and recreational venues (Giles-Corti and Donovan, 2002). While higher rates of physical activity and reduced levels of obesity have been associated with better access to leisure facilities, including open green space and beaches (Bauman and Smith, 1999, Ellaway et al., 2005, Sallis et al., 1997), recreational amenity access has been relatively weak as a predictor of physical activity compared to individual and social environmental factors (Giles-Corti and Donovan, 2002, Wendel-Vos et al., 2007).
Common weaknesses of studies in this field have been a reliance on self reported measures of neighbourhood accessibility, physical activity, and height and weight measurements used for body mass index (BMI) calculations. This is beginning to change with the increasing use of Geographic Information Systems (GIS) to provide objective measures of locational access to specific destination types and other neighbourhood access measures such as street connectivity, diversity of land use and dwelling density. There is some evidence of increasing precision in the measurement of physical activity for example through the use of accelerometers. A number of the multilevel studies noted earlier have used direct measurement of neighbourhood context but relied on self reported physical activity data (Ellaway et al., 2005, Li et al., 2005, Diez Roux et al., 2007, Ross et al., 2007). We are unaware of research reporting the use of objective measures of physical activity and neighbourhood accessibility within a multilevel framework.
Evidence that physical inactivity levels vary by socio-economic position (Parks et al., 2003, SPARC,) and findings from a US study that neighbourhood differences in physical activity remain after adjusting for individual socio-demographic and health characteristics (Yen and Kaplan, 1999), have raised the question of whether a social gradient exists in access to recreational resources. Studies that have investigated this question have generated mixed findings. These have ranged from evidence of an inequitable distribution in recreational facilities in favour of high income neighbourhoods (Estabrooks et al., 2003, Macintyre et al., 1993, Powell et al., 2006), to no association (Lee et al., 2005, Timperio et al., 2007), to results indicating better access to recreational amenities in more deprived neighbourhoods (Giles-Corti and Donovan, 2002, Craddock et al., 2005, Ellaway et al., 2007). Our New Zealand research has shown access to parks and recreational centres improves as area-level deprivation increases, with no association between deprivation and beach access (Pearce et al., 2007a), although this relationship is not consistent in rural areas and in some regions (Pearce et al., in press). Thus it is important to take social deprivation into account in a study of accessibility and obesity related outcomes.
This national study investigates the association between travel time access to recreational amenities – parks and beaches – and the physical activity patterns and BMI of residents in New Zealand. The selection of parks and beaches was based on the availability of national datasets, the ubiquitous nature of beaches as sites of recreational activity in New Zealand, given its island nation status, and the work of Giles-Corti and Donovan (2002) that identified parks and beaches as the most frequently used venues for recreational activity, after streets, in the coastal city of Perth.
Section snippets
Methods
In 2005 locational information was obtained from Land Information New Zealand (LINZ) and Department of Conservation for all parks in New Zealand (Pearce et al., 2006). Data on all beaches were also obtained from LINZ. To account for the large surface areas often represented by parks and beaches, each park and beach was converted in a GIS to represent multiple access points 100m apart. In total there were 46,274 access points for parks and 13,313 for beaches. Neighbourhood was defined as the
Results
Odds ratios and 95% confidence intervals were calculated for each outcome for access to beach and access to park quartiles (Table 1, Table 2). The best access quartile was compared to the other three quartiles. For the BMI model the best access quartile took the value of 0. We hypothesised that those with best access to parks and beaches would have lower BMI so we would expect positive B values for the other quartiles.
For the physical activity logistic regression the best access quartile has
Discussion
To our knowledge, this represents the first nationally representative study of the association of access to public open spaces with BMI and physical activity. The analysis showed very little association between access to parks and the outcome variables (and if anything there were indications of a negative association) and a weak association only between beach access, BMI and physical activity.
The strengths of the study are its national coverage of park and beach access, the use of physical
Conclusions
This study found little evidence of an association between locational access to open spaces and physical activity. To substantially advance understanding on the topic, greater specificity is required in both access and outcome measures. Access measures are needed that incorporate dimensions such as amenity attractiveness and safety, as well as travel time access and greater differentiation of physical activity outcome measures in terms of leisure and transport-related activity and activity
Acknowledgments
We thank Maria Turley and Kylie Mason of Public Health Intelligence, Ministry of Health for preparing the Health Survey data. The 2002/03 New Zealand Health Survey was funded by the Ministry of Health and the Crown is the owner of the copyright and the data.
Competing interests: None
Funding: This research was funded by the New Zealand Health Research Council, as part of the Neighbourhoods and Health project within the Health Inequalities Research Programme.
References (46)
- et al.
Geographical influences upon physical activity participation: evidence of a ‘coastal effect’
Aust. N. Z. J. Public Health
(1999) - et al.
Nowhere to play? The relationship between the location of outdoor play areas and deprivation in Glasgow
Health and Place
(2007) - et al.
Stepping towards causation: do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity?
Soc. Sci. Med.
(2007) - et al.
Socioeconomic status differences in recreational physical activity levels and real and perceived access to a supportive physical environment
Prev. Med.
(2002) - et al.
Increasing walking: how important is distance to, attractiveness, and size of public open space?
Am. J. Prev. Med.
(2005) - et al.
Environmental factors associated with participation in physical activity
Am. J. Prev. Med.
(2002) - et al.
Objective measures of neighbourhood environment and physical activity in older women
Am. J. Prev. Med.
(2005) - et al.
The link between obesity and the built environment. Evidence from an ecological analysis of obesity and vehicle miles of travel in California
Health & Place
(2006) - et al.
Developing a framework for assessment of the environmental determinants of walking and cycling
Soc. Sci. Med.
(2003) - et al.
Is availability of public open space equitable across areas?
Health and Place
(2007)