Associations of perceived neighborhood attributes with self-report and objective measures of walking in Hong Kong adults

Background: Walking for different purposes can help accrue health-enhancing levels of physical activity. There is evidence that the neighborhood environment impacts on residents’ walking. Most research on environmental correlates of walking has been conducted in low-density urbanized areas of Western countries. The public transport networks and built and socio-cultural environments of Asian densely-populated urban areas differ from their Western counterparts. An analysis of the levels of walking and environment-walking associations in areas at the high end of the walkability spectrum (commonly defined as high residential density, street connectivity, and access to diverse destinations) can help determine the magnitude of the effects of environmental attributes on residents’ walking. As part of the International Physical Activity and the Environment Network initiative, a cross-sectional study exploring environment-physical activity associations in Chinese-speaking adult residents of Hong Kong is currently being conducted. We report on preliminary findings of the study. Objectives: To examine the relationships of perceived neighborhood characteristics with self-reported walking for different purposes within the neighborhood of residence, accelerometer-based daily minutes of moderate-intensity physical activity (MPA) and step counts. Methods: A stratified two-stage cluster sampling design was used to recruit 195 Chinese-speaking adults, aged 20-65, residents of private dwellings and able to walk without assistance. The study sample was drawn from residential addresses within 16 Tertiary Planning Units (TPU) and classified based on their objective walkability and socio-economic status (SES) into four strata: high walkable/ high SES; low walkable/ high SES; high walkable/ low SES; and low walkable/ low SES. Area SES was operationalized as the median weekly household income for a TPU. Walkability was defined using Geographic Information Systems data on dwelling density and street connectivity. Participants provided information on socio-demographics, transportation- and recreation-related walking within the neighborhood, and perceived attributes of the neighborhood environment (Chinese version of the Neighborhood Environment Walkability Scale). Accelerometry-based (Actigraph GT1M) objective data on weekly patterns of physical activity was collected on 106 participants. For the purpose of this study, average daily minutes of MPA and step counts (markers of walking) were examined. Generalized linear models with appropriate variance and link functions and with standard errors adjusted for clustering effects were used to determine the associations of interest. All models were adjusted for socio-demographic confounders. Models of objective measures of walking were also adjusted for total time of accelerometer wearing and number of weekend/holidays days during which accelerometer data were collected. Results: Participants reported an average of 246 (SD=238) weekly minutes of walking for transport and 114 (SD=195) weekly minutes of walking for recreation. The average daily minutes of MPA were 45 (SD=25), while the average step counts were 10,039 (SD=3,681). Residents of objectively high-walkable areas reported 87 more weekly minutes of transport-related walking within the neighborhood than did those from low-walkable areas (p<.01). No significant differences between residents of areas differing in walkability were found in walking for recreation, step counts, and accelerometry-based MPA. Significant positive associations were found between transport-related walking within the neighborhood and household density (p<.05), street connectivity (p<.001), indoor places for walking (p<.05), traffic safety (p<.01), land use mix - diversity (p<.05), social environment (presence of people; p<.001) and crime (p<.05). Walking for recreation was positively related with indirect access to services (p<.01), building aesthetics (p<.05), green areas (p<.05), and land use mix - diversity (p<.05). Negative associations were observed for crime (p<.05) and household density (p<.05). MPA were positively related to perceived traffic hazards (p<.001), places for pedestrians (p<.05), social environment (p<.05), and fences separating traffic from pedestrians (p<.05). With the exception of social environment, these neighborhood attributes were also positively related with step counts. Perceived indirect access to services (p<.05) and traffic safety (p<.05) were also positive correlates of step counts. Conclusions: Adult residents of Hong Kong reported high levels of walking. However, it appears that some of the walking they undertook was at low intensity. Recent studies indicate that low-intensity physical activity may also be health enhancing, as it replaces sedentary time. As expected, walking for different purposes was associated with different environmental attributes. These findings mirror those observed in Western countries. The sets of environmental predictors varied between objective and self-report measures of walking, which may be due to measurement bias, and different measures capturing different aspects of walking behavior (i.e., any walking, walking at a moderate pace, and walking within the neighborhood). The observed between-area differences in within-neighborhood self-report but not objective measures of overall walking indicate that, unlike residents of many Western countries, Hong Kong residents may be able to overcome deficiencies in their local environment walkability by walking in areas outside of their neighborhood of residence, easily accessible thanks to an efficient public transport network. Support: Research Grants Council - CERG - Hong Kong SAR; US Department of Agriculture, Current Research Information System