Background
A consistent body of evidence shows both at the individual and neighborhood levels that deprivation is associated with poorer access to spaces for physical activity and to healthy food stores. This leads to lower than average physical activity and less healthful food consumption for those residing in deprived areas. To reduce these health disparities, interventions have sought to increase the availability of spaces for physical activity and healthy food stores in deprived areas. However, people residing in deprived areas need more than physical activity and healthy food in their neighborhood. They need access to social networks, employment, and education, which may be located outside their residential neighborhood. A complementary intervention to reduce health disparities would thus increase access to transportation systems including road networks, public transportation, and cycling. Limited research has examined disparities in access to transportation systems. If disparities in transportation systems exist, people in deprived areas may not be able to access necessary services, including those that support active living. As well, public transit and cycling both contribute directly to the accumulation of recommended daily physical activity.

Objectives
To examine individual and neighborhood level disparities in access to the road network, public transit, and a public bicycle share program in Montreal, Canada.

Methods
A cross-sectional design was used. Individual level data drawn from three population-based samples of adults recruited via telephone survey were pooled. Surveys were conducted between May 4^th - June 10^th 2009, October 8^th - December 12^th 2009, and November 8^th - December 12^th 2010 respectively. Response rates for the surveys were 36.9%, 34.6%, and 35.7%. The pooled sample included 7012 respondents (mean age=48.6 years, 59.1% female) nested in 33 neighborhoods. Neighborhood level data were drawn from the 2006 Canadian census. Data analysis consisted of three separate multilevel linear regressions examining each of the three following transportation access indicators: 1) road network access, operationalized by applying a kernel density value of the Montreal road network to each participant based on home address, 2) public transit access, operationalized by the road network distance between the participants home to the nearest subway station, and 3) public bicycle program access, operationalized by the road network distance between the participants home to the nearest public bicycle station. Individual level variables predicting disparities included age, sex, education, income, employment status, country of birth, and having a driver’s license. Neighborhood level variables included a binary education variable (>25% of population without high school vs. <25% of population without high school), a binary income variable (>25% of population low income vs. <25% of population low income), density of destinations, and street connectivity. Analyses were weighted by neighborhood population density.

Results
For the road network, no differences in access were observed between people of different income and employment categories. Neighborhoods with a high percentage of low income resident's had greater road network density in their neighborhood (β=0.5; p<0.01). For public transit, people with less than a high school education (β=138; p<0.01) or employed part time (β=122; p<0.01) lived further from a subway station. Neighborhoods with a high percentage of low income resident's (β=-2354; p<0.05) were closer to the subway system. Density of destinations (β=-49; p<0.05) was positively associated while street connectivity (β=16; p<0.01) was negatively associated with neighborhood access to the subway. For the public bicycle share program people with household incomes less than $20000 (β=-372; p<0.05), live closer while people employed part time (β=100; p<0.05) live significantly further. Neighborhoods with a high percentage of low income resident's (β=-3340; p<0.05) were, on average, closer to the bicycle share. Density of destinations (β=-76; p<0.05) was associated with living nearer the bike share while street connectivity was associated with living further (β=21; p<0.01).

Conclusions
Results suggest that in Montreal neighborhoods with a high percentage of low income residents have better access to transportation systems including the road network, public transit and a bicycle share program. This finding should be interpreted considering the potential positive contribution of cycling and public transit to health, while acknowledging that poorer neighborhoods are often the first candidates for major highway construction leading to increased exposure to air pollution, noise, and injuries due to collisions with motor vehicles. At the individual level, within neighborhood variation explains the finding that people with less education or who are employed part time live in areas with worse access to public transit within low income neighborhoods. The results of this preliminary study show that in Montreal low income neighborhoods have access to transportation systems with potential positive and negative health consequences. Researchers should continue to study disparities in access to transportation in other North American cities and the potential positive and negative health consequences of these disparities.

Support/Funding
Canadian Institutes for Health Research grant for Intervention Research and by LG’s CIHR/CRPO Applied Public Health Chair on Neighborhoods, Lifestyle, and Healthy Body Weights.