• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Soc Sci Med. Author manuscript; available in PMC Nov 18, 2012.
Published in final edited form as:
PMCID: PMC3500640
NIHMSID: NIHMS405260

Neighborhood Built Environment and Income: Examining Multiple Health Outcomes

Abstract

There is growing interest in the relation of built environments to physical activity, obesity, and other health outcomes. The purpose of the present study was to test associations of neighborhood built environment and median income to multiple health outcomes and examine whether associations are similar for low- and high-income groups. This was a cross-sectional study of 32 neighborhoods in Seattle, WA and Baltimore, MD regions, stratified by income and walkability, and conducted between 2001–2005. Participants were adults aged 20–65 years (n=2199; 26% ethnic minority). The main outcomes were daily minutes of moderate-to-vigorous physical activity (MVPA) from accelerometer monitoring, body mass index (BMI) based on self-report, and mental and physical quality of life (QoL) assessed with the SF-12.

We found that MVPA was higher in high- versus low-walkability neighborhoods but did not differ by neighborhood income. Overweight/obesity (BMI≥25) was lower in high-walkability neighborhoods. Physical QoL was higher in high-income neighborhoods but unrelated to walkability. Adjustment for neighborhood self-selection produced minor changes. We concluded that living in walkable neighborhoods was associated with more physical activity and lower overweight/obesity but not with other benefits. Lower- and higher-income groups benefited similarly from living in high-walkability neighborhoods. Adults in higher-income neighborhoods had lower BMI and higher physical QoL.

Keywords: obesity, physical activity, built environment, health disparities, USA, quality of life (QoL), neighborhood, walkability

Introduction

Physical inactivity and obesity are prevalent and serious health challenges, contributing to cardiovascular diseases, certain cancers, diabetes, and mental disorders (Andersen, 2003; Dishman, Washburn, & Heath, 2004). Physical activity and obesity have been linked with physical attributes of neighborhoods. Neighborhoods considered walkable have non-residential destinations (e.g., shops) close to residences and well-connected streets. Low-walkability areas separate residences from destinations and have poorly connected street networks, so walking to destinations is difficult. People walk and bicycle more for transportation in high-walkability than low-walkability neighborhoods, as indicated by multiple reviews (Gebel, Bauman, & Petticrew, 2007; Heath, Brownson, Kruger, et al., 2006; Transportation Research Board & Institute of Medicine, 2005). There is a need to confirm whether more walkable neighborhoods are associated with higher total physical activity, particularly using objective measures of environment and activity (Frank, Andresen, & Schmid, 2004), because total physical activity should be most closely related to health benefits. A few studies indicate adults living in high-walkability neighborhoods or regions are less likely to be overweight or obese than those living in low-walkability areas (Papas, Alberg, Ewing, Helzlsouer, et al., 2007), but further studies are needed.

Because disparities in health outcomes (Centers for Disease Control and Prevention, 2004) and physical activity are well documented across socioeconomic groups (Crespo, Smit, Andersen, et al., 2000), an important question is whether favorable built environments could reduce health disparities. Findings that walkability was related to physical activity and obesity among whites but not blacks (Frank, Andresen, & Schmid, 2004; Frank, Sallis, Chapman, & Saelens, 2005) raise the possibility that not all groups benefit from walkable built environments. Because a primary health objective of the United States is to eliminate health disparities (United States Department of Health and Human Services, 2000), it is important to determine whether walkability has similar associations with health outcomes in lower- and higher-income groups.

Advocates of walkable communities propose additional health benefits that have not been examined empirically (Duany, Plater-Zyberk, & Speck, 2000; Frank, Engelke, & Schmid, 2003; Frumkin, Frank, & Jackson, 2004). One hypothesis is that suburban residents who drive everywhere have fewer chances to form bonds with neighbors, negatively impacting social cohesion (Wood, Shannon, Bulsara, et al., 2008). Inadequate social networks are a risk factor for depression (Kawachi & Berkman, 2001), so residents of low-walkability neighborhoods might have more depressive symptoms. Some claim overall quality of life is higher for people living in walkable communities (Duany et al., 2000; Frumkin et al., 2004).

The present study investigated how living in high- vs. low-walkability and high- vs. low-income neighborhoods was related to adults’ biological, behavioral, social, and mental health outcomes. Because self-selection to neighborhood has been identified as a potential confounder of associations with walkability (Transportation Research Board & Institute of Medicine, 2005; Handy, Cao, & Mokhtarian, 2006; Frank, Saelens, Powell, & Chapman, 2007; Eid, Overman, Puga, & Turner, 2007), analyses were conducted with and without adjusting for participants’ reasons for moving to their current neighborhoods.

Method

Study Design

The Neighborhood Quality of Life Study (NQLS) is an observational epidemiologic study designed to compare multiple health outcomes among residents of neighborhoods stratified on “walkability” characteristics and median household income. Data were collected from 2001 to 2005 in two metropolitan areas in the United States that were chosen based on availability of parcel-level land use information, and variability in walkability. The King County-Seattle, WA and Baltimore-Washington DC regions met these criteria.

Participants were recruited from 32 neighborhoods; 16 from Seattle-King County and 16 from Baltimore-Washington DC regions. Table 1 defines quadrants formed by low versus high levels of walkability and low versus high levels of income, an indicator of socioeconomic status (SES). The study was approved by Institutional Review Boards at participating academic institutions, and participants gave written informed consent.

Table 1
Neighborhood Quality of Life Study design: Neighborhood walkability and median household income by quadrant *

Neighborhood Selection

Land use variables were used to compute a “walkability index” based on conceptual (Frank & Engelke, 2001) and empirical literature (Cervero & Kockelman, 1997; Saelens, Sallis, & Frank, 2003) that identify residential density, mixed land use, and street connectivity as key components of walkability. Building setbacks from the street or sidewalk are also important aspects of pedestrian-oriented design (Cervero & Kockelman, 1997). Thus, retail floor area ratio (retail building square footage divided by retail land square footage) was included in the index, with a higher ratio indicating a more pedestrian-oriented design and lower ratios suggesting more land area devoted to parking. Although other environmental variables have been related to active transport, such as sidewalks, traffic calming, and intersection characteristics (Cervero & Kockelman, 1997; Saelens, Sallis, & Frank, 2003; Handy, Boarnet, Ewing, & Killingsworth, 2002), these variables are not widely available.

The census block group was chosen as the most appropriate geographical scale to develop walkability measures for neighborhood selection. For each block group, the walkability index was derived as a function of four variables: (a) net residential density (ratio of residential units to the land area devoted to residential use); (b) retail floor area ratio (FAR; described above, indicating pedestrian-oriented design); (c) land use mix (diversity of land use types per block group; normalized scores ranged from 0 to 1, with 0 being single use and 1 indicating an even distribution of floor area across 5 uses--residential, retail, entertainment, office, institutional); and (d) intersection density (connectivity of street network measured as the ratio of number of intersections with 3 or more legs to land area of the block group in acres). Though this intersection density measure undercounts intersections on roads that form the edge of blockgroups, this particular metric was one of the best predictors of active transportation in an examination of multiple connectivity measures (Dill, 2004). The absolute count of intersections may not be entirely accurate, but the metric should be more than adequate for the present purpose of ranking blockgroups and neighborhoods.

Standardized scores for each measure were calculated separately for each region, so variables were normalized for the distributions in each region. The walkability index was a weighted sum of z-scores of the four normalized urban form measures as stated in the following expression:

Walkability=[(2×z-intersectiondensity)+(z-netresidentialdensity)+(z-retailfloorarearatio)+(z-landusemix)]

The walkability index is described in more detail elsewhere (Frank, Sallis, Saelens, et al., in press). This walkability index was compared against census journey to work data from 2000 in Seattle and Baltimore regions. Higher walkability was significantly associated with less driving and more walking to work (Frank, Sallis, Saelens, et al., in press). More importantly, use of the walkability index is supported by at least 12 published papers showing the same or similar indexes have been significant positive correlates of walking and physical activity (Frank, Andresen, & Schmid, 2004; Frank, Schmid, Sallis, et al., 2005; Frank, Sallis, Conway, et al., 2006; Frank, Bradley, Kavage, et al., 2007; Frank, Kerr, & Sallis, 2007; Frank, Saelens, Powell, & Chapman, 2007), including studies in Australia (Cerin, Leslie, DuToit, et al., 2007; Leslie, Frank, Owen, et al., 2007; Owen, Cerin, Leslie, et al., 2007) and studies of youth (Kerr, Rosenberg, Sallis, et al., 2006; Kerr, Frank, Sallis, & Chapman, 2007; Kligerman, Sallis, Ryan, et al., 2007).

Correlations among walkability component scores, with data pooled across both regions, were modest, ranging from .04 (land use mix—intersection density) to .31 (retail FAR—intersection density). Correlations of the individual components with the walkability index ranged from .46 (net residential density) to .80 (intersection density). Thus, each component contributed substantial independent variance to the walkability index, and all correlations were positive, as expected. Walkability index values ranged from −1.29 to 8.28 in the Seattle region and from −1.57 to 8.17 in the Baltimore-Washington, DC region, demonstrating substantial and similar variation in both regions.

The walkability index and census-based demographic data were used to select neighborhoods. Block groups are smaller units of geography than tracts and were selected in contiguous clusters that approximated neighborhoods. Because U.S. cities have among the lowest walkability in the world (Newman & Kenworthy, 1991), it is essential to systematically select neighborhoods to produce wide variation. Block groups in King County, WA and five counties in the Baltimore-Washington, DC region were ranked and divided into deciles based on the normalized walkability index within each region. Block groups were categorized into “high income” and “low income” based on 2000 Census median household income data. Block groups with median household incomes less than $15,000 or greater than $150,000 were excluded, to avoid outliers in neighborhood incomes. The 2nd, 3rd, and 4th deciles constituted the “low income” category; the 7th, 8th, and 9th deciles made up the “high income” category; the 5th and 6th deciles were omitted to create separation between the categories.

The “walkability” and income characteristics of each block group were crossed with each other (low/high walkability X low/high income) to produce a list of block groups that fit into one of four quadrants. Clusters of contiguous block groups approximated neighborhoods and were flagged for potential selection. A geographic distribution of neighborhoods was desired in each region to enhance diversity of racial/ethnic composition, access to transit, housing types, and access to employment. Each of the 32 neighborhoods was composed of 2 to 13 census block groups. The goal was to define “neighborhoods” with similar numbers of households, understanding that in all cases low-walkability neighborhoods would be geographically larger than high-walkability neighborhoods. Though adjacent block groups varied in walkability, overall variation was much greater between neighborhoods than across block groups within neighborhoods. To avoid “boundary” problems with very different walkability characteristics just outside the defined neighborhoods, investigators personally inspected all the candidate blockgroups (“ground-truthing”), so final selections were made based on both GIS walkability data and inspection. For example, if a low-walkability blockgroup had a substantial shopping area close by but in another block group, that candidate block group was excluded from the study. Table 1 shows for each region the average values and standard deviations in the walkability index and median household income for the block groups in each quadrant. Additional characteristics of study quadrants and neighborhoods are available in Frank et al. (in press).

Participant Recruitment and Assessment Procedures

Recruitment and data collection were conducted during two 18-month phases. During Phase 1 (May 2002–November 2003), participants in the Seattle/King County region were recruited and assessed. During Phase 2 (December 2003–June 2005), participants in the Baltimore/Maryland region were recruited and assessed. In each phase, participants were recruited during the first 12 months, and a second assessment of physical activity was conducted six months later to control for season. Within each phase, participants were recruited across all neighborhoods simultaneously to further prevent seasonal bias.

Contact information of people residing within selected neighborhoods was obtained from a marketing company. Records were randomly selected within each neighborhood, and a letter introducing the project was mailed to heads of households, followed by telephone calls. If the initially-targeted adult refused or was ineligible, another adult in the household was invited. Eligibility was defined as being between 20 – 65 years, not residing in a group living establishment (e.g. nursing home, dormitory), ability to complete written surveys in English, and absence of a medical condition that interfered with the ability to walk.

After a participant returned a signed informed consent, they were mailed an accelerometer (with instructions for wearing and mailing back) to obtain an objective assessment of physical activity. A survey was mailed to the participant so he/she received it on the last day they were supposed to be wearing the accelerometer, so survey content would not influence physical activity. Participants were given the option of completing surveys by mail, online, or telephone interview. Six months later, an accelerometer and a different survey were sent for assessment in a different season. Upon receipt of accelerometer and survey data, incentive payments were mailed; $20 for the first assessment and $30 for the second.

Measures

Total physical activity

Actigraph (Actigraph, Inc; Fort Walton Beach, FL) model 7164 or 71256 accelerometers, with established reliability and validity (Welk, 2002) were used to objectively assess moderate-to-vigorous physical activity (MVPA). Participants were instructed to wear the accelerometer snugly around the waist for 7 days on each measurement occasion. The accelerometer was set to record intensity of movement each minute. A valid accelerometer hour was defined as having no more than 30 consecutive ‘zero’ values, and a valid day consisted of 10 valid hours. If there were not at least 5 valid days or a minimum of 66 valid hours across 7 days, the participant was asked to re-wear the accelerometer. On valid days, each minute was scored as meeting or not a criterion of at least moderate intensity physical activity based on published cutpoints (Freedson, Melanson, & Sirard, 1998). Average daily minutes of MVPA was the summary variable used in analyses.

Walking for transportation and leisure

Items from the long version of the International Physical Activity Questionnaire (IPAQ; http://www.ipaq.ki.se), shown to be reliable and valid (Craig, Marshall, Sjostrom, et al., 2003), were used to assess transportation and leisure walking. The transportation walking items queried number of days during the last week spent walking at least 10 minutes from place to place and the typical minutes per day. Similarly structured items queried time in leisure walking. Total minutes per week (days X minutes per day) were calculated.

Body mass index (BMI)

Self-reported weight and height was used to calculate BMI (kg/m2). Overweight was defined as BMI ≥ 25 and obesity as BMI ≥ 30 (National Institutes of Health & National Heart, Lung, and Blood Institute, 1998).

Quality of life and psychosocial variables

The 12-item Short-Form Health Survey (SF-12; www.sf-36.org) was used to assess physical quality of life (QoL) and mental QoL (Ware, Kosinski, & Keller, 1996). The Center for Epidemiologic Studies’ 20-item depression scale (CES-D) assessed depressive symptoms (Radloff, 1977). Perceived neighborhood social cohesion was assessed using a 5-item scale (Sampson, Raudenbusch, & Earls, 1997). Neighborhood satisfaction was defined as the mean of 17 ratings of satisfaction with aspects of walkability and transportation, social interaction, traffic and crime safety, and school quality. Each item was rated using a 5-point scale from strongly dissatisfied (1) to strongly satisfied (5) on a scale developed by the investigators.

Covariates

Demographic covariates assessed by survey were gender, age, education (5 levels from less than high school to graduate degree), ethnicity (re-categorized as non-Hispanic white or non-white), number of motor vehicles/adults in household, marital status (re-categorized as married/living together or other), number of people in household, and years at current address.

To control for walkability-related self-selection of neighborhoods, a scale (internal consistency alpha = .76) of “reasons for moving” to the current home was computed by averaging ratings of importance of three items; “desire for nearby shops and services,” “ease of walking,” and “closeness to recreational facilities” (adapted from Frank et al., 2007).

Statistical analyses

Mixed effects regression models (using SAS PROC MIXED) were fitted for all continuous variables, and generalized linear mixed models (using SAS PROC GLIMMIX) were fitted for the dichotomous overweight/obesity outcomes. For MVPA, the IPAQ variables (natural-log transformed because of skewness), BMI, and weight status, two time points were available for analysis. Therefore, a repeated measures framework was used for these variables. The analyses took neighborhood clustering into account, so three-level multilevel models were fitted to account for repeated measures nested within subjects and subjects nested within neighborhoods. For the remaining variables in which only one time point was available, a two-level data structure was used where subjects were nested within neighborhoods, and mixed effects regression models were fitted. All analyses were carried out using SAS version 9.1.3.

The primary exposures of interest were the quadrants constructed by crossing high/low walkability neighborhoods with high/low income neighborhoods. The main effects of walkability and income and their interaction were the main focus of these analyses. All models were adjusted for the demographic covariates and study region (Seattle, Baltimore areas). Results are reported before and after including reasons for moving as a covariate.

Results

Participant Characteristics and Representativeness

Data were collected from 2199 participants from 32 neighborhoods. Demographics of the study sample by quadrant are reported in Table 2. The sample was well balanced by sex, mostly well-educated, most were married, and 26% were non-white.

Table 2
Descriptive statistics for demographic variables stratified by neighborhood income and walkability quadrants

A total of 8504 eligible adults were contacted by phone. The study participation rate (i.e., returned survey 1/eligible contacts) was 26% overall and did not differ by quadrant (range of 23% to 29% by quadrant). The 6-month retention rate was 87% overall (range of 84% to 88% by quadrant), after eliminating those who were no longer eligible (e.g., because they moved out of the region). Comparisons of participant demographics with census data showed the study sample was older (median age, 45.1 vs. 35.7 years, p<.01), had fewer females (48.2% vs. 51.8%, p=.03), more whites (74.0% vs. 65.1%, p<.01), fewer Hispanics (3.7% vs. 5.6%, p<.01), and higher household incomes (median incomes, $60–$69,000 vs. $50–$59,000, p<.01) than residents of the census block groups in which participants lived.

Neighborhood Walkability and Income Effects

Differences among participants living in neighborhoods in the high- vs. low-walkability and high- vs. low-income quadrants are shown in Table 3. Quadrant means were adjusted for covariates. Significance levels for the walkability-by-income interactions and the walkability and income main effects for each outcome are indicated.

Table 3
Primary results showing adjusted means by study quadrant and tests of hypotheses for neighborhood income and walkability and the interaction unadjusted for “reasons for moving here” scale.

Total Moderate and Vigorous Physical Activity (MVPA)

The walkability main effect was highly significant (p=.0002). On average, participants in high-walkability neighborhoods had 5.8 more minutes per day of objectively measured MVPA than those in low-walkability neighborhoods.

Walking for Transport

The walkability-by-income interaction (p=.027) and walkability main effect (p=<.0001) were both significant. Overall, the significant walkability main effect indicated a higher average number of minutes per week of walking for transportation in high-walkability neighborhoods (44.3 minutes per week) compared to low-walkability neighborhoods (12.8 minutes per week). Walking for transportation was significantly higher in high-walkability neighborhoods compared to low-walkability neighborhoods for both high- and low-income neighborhoods; however, the differential was larger in high-income neighborhoods (5.1 minutes) versus low-income neighborhoods (2.3 minutes).

Walking for Leisure

The walkability main effect was significant (p=.012), with people living in high-walkability neighborhoods averaging 18.5 minutes per week of leisure walking compared to 14.2 minutes per week in low-walkability neighborhoods.

Body Mass Index

The income main effect (p=.003) indicated that participants living in lower-income neighborhoods had higher average BMI’s (27.4) than those in higher-income neighborhoods (26.4).

QoL and Depression

The income main effect was significant (p=.006), with participants living in higher-income neighborhoods reporting higher physical QoL scores than those living in lower-income neighborhoods (53.4 vs. 52.3, respectively). There were no significant findings for mental QoL and depression.

Neighborhood Satisfaction

The income main effect was highly significant (p<.0001), with participants living in higher-income neighborhoods reporting higher average neighborhood satisfaction than those living in lower-income areas (3.95 vs. 3.52, respectively). The trend (p=.07) for an income-by-walkability interaction indicated somewhat higher neighborhood satisfaction scores in high-walkability vs. low-walkability areas but only in higher-income neighborhoods; there were negligible differences for lower-income neighborhoods.

Social Cohesion

The income main effect was significant (p<.0001), with participants in higher-income neighborhoods reporting higher perceived social cohesion than those in lower-income areas (3.79 vs. 3.44, respectively).

Percent Overweight or Obese (> 25.0 BMI)

The walkability main effect was significant (p=.007), with the odds of being overweight or obese 35% higher for participants living in low- vs. high-walkability neighborhoods (OR=1.35, 95% CI: 1.09, 1.69).

Percent Obese (>30.0 BMI)

The income main effect was significant (p=.007), with participants living in lower-income neighborhoods having 53% greater odds of being obese than those living in higher-income neighborhoods (OR=1.53, 95% CI: 1.12, 2.07).

Impact of Neighborhood Selection on Neighborhood Walkability and Income Effects

All analyses of outcome measures were repeated adding the “reasons for moving here” score as a covariate to control for preferences related to “activity-friendly” environments. Results in Table 4 show minor effects of the additional covariate on minutes of transport walking, minutes of leisure walking, mental QoL, and depression. For minutes of transport walking, the income-by-walkability interaction was no longer significant (p=.11). However, the walkability main effect was still highly significant (p<.0001). For minutes of leisure walking, the walkability main effect was no longer significant (p=.36). For mental QoL, the walkability main effect became significant (p=.03), with participants living in high-walkability neighborhoods having an average score that was slightly lower (49.7) than those living in low-walkability neighborhoods (50.7). For depression, the walkability main effect became significant (p=.015), with participants living in high-walkability neighborhoods having a higher score (9.88) than those in low-walkability neighborhoods (8.85).

Table 4
Primary outcome results showing adjusted means by quadrant and tests of hypotheses for neighborhood income and walkability and the interaction adjusted for “reasons for moving here” scale.

Discussion

Four major findings emerged from the present study. First, neighborhood walkability was related to higher levels of physical activity and lower risk of being overweight or obese, but not to social or psychological outcomes. Second, neighborhood income was not related to any measure of physical activity, but lower-income adults had less favorable weight status, physical QoL, neighborhood satisfaction, and social cohesion than higher-income participants. Third, there was only one significant interaction between neighborhood walkability and income, indicating walkability had a stronger positive association with walking for transport in high-income than in low-income participants. Fourth, after adjusting for potential self-selection bias (i.e., “reasons for moving here”), all significant associations of outcomes with walkability and income remained significant, except walking for leisure. However, associations with mental quality of life and depression score became significant, indicating slightly poorer mental health in residents of high-walkability neighborhoods, particularly for those in low income areas.

Adults living in high-walkability neighborhoods had higher objectively measured total physical activity as well as higher self-reported walking for transportation and leisure than did participants from low-walkability neighborhoods. The weekly difference in objectively measured physical activity was about 47 minutes per week for the higher-income group and about 34 minutes for the lower-income group. On average, living in a high-walkability neighborhood was associated with meeting the 30 minute per day physical activity guidelines (Haskell, Lee, Pate, et al., 2007) at least one day more per week than those in low-walkability neighborhoods. Present findings confirm previous results of higher total physical activity in high-walkability neighborhoods (Frank et al., 2005; Saelens, Sallis, Black, & Chen, 2003). These results extend the evidence by demonstrating the effect generalizes to both higher- and lower-income groups, and the walkability effect appears to be stronger for objectively-measured than self-reported physical activity. Walkability associations with physical activity were not explained by self-selection into neighborhoods based on predisposition towards activity-friendly environments, a finding consistent with recent studies (Frank et al., 2007; Handy et al., 2006; Handy, Cao, & Mokhtarian, 2008). Nonsignificant differences in total physical activity by neighborhood income were unexpected, because higher activity levels among higher-income participants have been reported (Crespo et al., 2000; United States Department of Health and Human Services, 2000), but studies reporting SES differences in objectively measured physical activity are rare and generally agreed with present results (Troiano, Berrigan, Dodd, et al., 2008).

It appears walkability differences in walking for both transportation and leisure contributed to observed differences in total physical activity. It is well-established that adults walk more for transportation in walkable neighborhoods (Heath et al., 2006; Transportation Research Board & Institute of Medicine, 2005; Frank et al., 2004), but the few studies that examined leisure walking or total self reported physical activity usually reported no walkability effect (Owen, Humpel, Leslie, Bauman, & Sallis, 2004; Saelens & Handy, 2008). The walkability—leisure walking association was weaker than the relation with transport walking, and after adjustment for self-selection, the walkability—leisure association became nonsignificant. This was expected because the walkability index was designed to explain transport walking.

There were no significant income differences on walking for transport or leisure, but there was an interaction between walkability and income on walking for transportation. The walkability—walking for transport association was weaker for adults living in lower-income than in higher-income neighborhoods. This is an important finding because it suggests lower-income residents may not experience all of the benefits from living in a walkable neighborhood unless other needs are met. Perceived danger from crime, which is higher among lower-income adults (Loukaitou-Sideris, & Eck, 2007), could reduce their willingness to walk for transport even in high-walkability neighborhoods (Doyle, Kelly-Schwartz, Schlossberg, & Stockard, 2006). After adjusting for self-selection, the walkability by income interaction became nonsignificant. Self-selection may not apply equally to lower- and higher-income groups, since higher-income groups may be able to satisfy more personal criteria when selecting neighborhoods (Levine & Frank, 2007).

Previous studies found walkable neighborhoods protected against overweight and obesity (Papas et al., 2007), but the present study extends previous work. There was a highly significant walkability effect for percent overweight or obese. Though the walkability by income interactions were not significant, living in low-walkability neighborhoods was associated with about a 50% increased risk of being overweight or obese in the higher-income group (OR=1.53), and the odds ratio was somewhat lower in the lower-income group (OR=1.20). Adjusting for self-selection had virtually no effect on the odds ratios, raising questions about claims that the walkability—obesity association is due to self-selection (Handy et al., 2006; Frank et al., 2007; Eid et al., 2007).

Hypothesized QoL, social, and psychological benefits of living in walkable neighborhoods received no empirical support. Despite using high-quality measures and examining a variety of outcomes, there was no evidence residents of walkable neighborhoods had benefits beyond physical activity and weight status.

The negative finding of walkability in relation to mental health, after adjusting for neighborhood selection factors, is consistent with evidence linking high residential densities with psychological stress (Evans, 2003). However, scores on the present mental health measures were well within the normal range, so the practical impact of these small differences is unclear. A recent review reported some studies found built environment factors were related to depressive symptomatology, but the evidence base was small for any specific built environment characteristic, such as walkability (Mair et al., 2008; Clark et al., 2006). Results were inconsistent, with one study finding walkability was protective of depressive symptoms among older men, but not women (Berke et al., 2007). Neighborhood population density, a component of walkability, has been found previously to be positively, negatively, or not associated with mental health outcomes (Clark et al., 2006). The presence/absence or magnitude of a built environment attribute may not be as important as its quality. For example, poorer quality of housing (e.g., state of repair) appears related to greater lifetime incidence of depression (Galea et al., 2005) and higher depressive symptomatology (Weich et al., 2002). More and better evidence is needed to improve understanding of built environment effects on mental health.

The present study confirmed the negative effects of low SES on multiple health outcomes. Lower-income participants had less favorable physical QoL, social cohesion, and neighborhood satisfaction. Unfortunately, there was little evidence that living in walkable neighborhoods alleviated these disadvantages, so efforts to improve social and physical environments, enhance health and social services, and empower vulnerable populations need to be strengthened. A recent study found walkable low-income, mostly-minority neighborhoods had lower levels of maintenance, aesthetic, and safety qualities than higher-income neighborhoods (Zhu & Lee, 2008), so neighborhood built environment attributes beyond walkability should be examined to determine their relation to health outcomes.

A strength of the present study was the design to recruit participants from two regions of the United States that differed in demographic composition, climate, geography, and era of development. Results generalized across the two regions. Other strengths included use of accelerometers to objectively assess physical activity, assessment of walking for multiple purposes, control for seasonal effects, selection of neighborhoods that varied widely on walkability defined by GIS and income, and use of validated measures. The present study is one of the few to statistically adjust for potential self-selection bias (Handy et al., 2006; Frank et al., 2007; Handy et al., 2008; Bagley & Mokhtarian, 2002).

An important limitation was the modest recruitment rate and the under-representation of racial-ethnic minority groups and very low SES participants. Thus, present findings should not be generalized to the most disadvantaged populations, and studies of very low income and specific racial-ethnic populations are needed. The cross-sectional design is an important limitation, so prospective designs that follow people who move are needed to determine the relative contributions of personal and environmental influences on physical activity and weight status. Though the validity of the walkability index was supported in this study and several others, it has limitations related to the completeness and accuracy of the multiple data sets required for its computation. In addition, the intersection density variable, based on census block group geography, misses intersections at the boundaries of the blockgroup.

Physical inactivity and obesity are two of the most significant health problems in the United States and globally (Andersen, 2003; Dishman et al., 2004; World Health Organization, 2004), and both outcomes were related to neighborhood attributes which are directly controlled by public policies. Policies to encourage development of more walkable neighborhoods and enhancements to existing neighborhoods could provide health benefits to large proportions of the population, both low- and high-income, on a relatively permanent basis. Policies that favor walkable neighborhood designs have also been related to reductions in driving, greenhouse gases, and air pollution; conservation of open space; and reduced spending on public infrastructure (Frank et al., 2003; Frumkin et al., 2004; Frank et al., 2006; Ewing, Bartholomew, Winkelman, Walters, & Chen, 2007). Some negative effects have been identified, such as local traffic congestion and concentration of air pollution (Frumkin et al., 2004). Thus, walkable neighborhoods are not a panacea, and policies promoting walkable development patterns should be combined with other policies to avoid negative outcomes, especially among low-income populations. The potential to produce widespread and long-lasting favorable impacts on physical activity and overweight/obesity should make the creation and improvement of walkable neighborhoods a high priority on the public health agenda. An important next step in research is to identify the shape of the relation of neighborhood environment characteristics to physical activity and overweight/obesity outcomes so recommended levels of walkability attributes can be developed. Other studies are needed to strengthen evidence of causality through prospective and quasi-experimental studies.

Acknowledgments

This study was supported by grant HL67350 from the National Heart, Lung, and Blood Institute. The funder was not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. We acknowledge the important contributions of the office of the County Executive in King County, WA, and multiple state and county agencies in Maryland.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Contributor Information

James F. Sallis, San Diego State University, San Diego, CA UNITED STATES.

Brian E Saelens, University of Washington and Children’s Hospital and Regional Medical Center, Seattle, WA.

Lawrence D Frank, School of Community and Regional Planning, University of British Columbia, Vancouver, BC, Canada; Lawrence Frank & Company, Point Robert, WA.

Terry L Conway, Graduate School of Public Health, San Diego State University, San Diego, CA.

Donald J Slymen, Graduate School of Public Health, San Diego State University, San Diego, CA.

Kelli L Cain, Department of Psychology, San Diego State University, San Diego, CA.

James E Chapman, Lawrence Frank & Company, Point Robert, WA.

Jacqueline Kerr, San Diego State University; University of California, San Diego.

References

  • Andersen R, editor. Obesity: Etiology, assessment, treatment, and prevention. Champaign, IL: Human Kinetics; 2003.
  • Bagley MN, Mokhtarian PL. The impact of residential neighborhood type on travel behavior: a structural equation modeling approach. Annals of Regional Science. 2002;36:279–297.
  • Berke EM, Gottlieb LM, Moudon AV, Larson EB. Protective association between neighborhood walkability and depression in older men. Journal of the American Geriatric Society. 2007;55:526–33. [PubMed]
  • Besser LM, Dannenberg AL. Walking to public transit: Steps to help meet physical activity recommendations. American Journal of Preventive Medicine. 2005;29:273–280. [PubMed]
  • Centers for Disease Control and Prevention. REACH 2010 surveillance for health status in minority communities--United States, 2001–2002. MMWR Surveillance Summary. 2004;53(SS06):1–36. [PubMed]
  • Cerin E, Leslie E, DuToit K, Owen N, Frank LD. Destinations that matter: Associations with walking for transport. Health & Place. 2007;13:713–724. [PubMed]
  • Cervero R, Kockelman KM. Travel demand and the 3Ds: Density, diversity, and design. Transportation Research-Part D. 1997;2:199–219.
  • Clark C, Stansfeld SA, Candy B. A systematic review on the effect of the physical environment on mental health. Epidemiology. 2006;17:S527.
  • Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, Pratt M, Ekelund U, Yngve A, Sallis JF, Oja P. International Physical Activity Questionnaire: 12-country reliability and validity. Medicine & Science in Sports & Exercise. 2003;35:1381–1395. [PubMed]
  • Crespo CJ, Smit E, Andersen RE, Carter-Pokras O, Ainsworth BE. Race/ethnicity, social class and their relation to physical inactivity during leisure time: Results from the third National Health and Nutrition Examination Survey, 1988–1994. American Journal of Preventive Medicine. 2000;18:46–53. [PubMed]
  • Dill J. Measuring connectivity for bicycling and walking. Presentation at Active Living Research Conference; 2004. http://www.activelivingresearch.org/files/dill_presentation.pdf.
  • Dishman R, Washburn R, Heath G, editors. Physical activity epidemiology. Champaign, IL: Human Kinetics; 2004.
  • Doyle S, Kelly-Schwartz A, Schlossberg M, Stockard J. Active community environments and health: The relationship of walkable and safe communities to individual health. Journal of the American Planning Association. 2006;72:19–32.
  • Duany A, Plater-Zyberk E, Speck J. Suburban nation: The rise of sprawl and the decline of the American dream. New York: North Point; 2000.
  • Eid J, Overman HG, Puga D, Turner MA. Fat city: Questioning the relationship between urban sprawl and obesity. Department of Economics, University of Toronto; 2007. [Accessed December 20, 2007]. www.chass.utoronto.ca/jeaneid/
  • Evans G. The built environment and mental health. Journal of Urban Health. 2003;80:536–555. [PMC free article] [PubMed]
  • Ewing R, Bartholomew K, Winkelman S, Walters J, Chen D. Growing cooler: The evidence on urban development and climate change. Washington, DC: Urban Land Institute; 2007.
  • Frank LD, Andresen MA, Schmid TL. Obesity relationships with community design, physical activity, and time spent in cars. American Journal of Preventive Medicine. 2004;27:87–96. [PubMed]
  • Frank LD, Bradley M, Kavage S, Chapman J, Lawton K. Urban form, travel time, and cost relationships with work and non-work tour complexity and mode choice. Transportation. 2008;35:37–54.
  • Frank L, Kerr J, Chapman J, Sallis J. Urban form relationships with walk trip frequency and distance among youth. American Journal of Health Promotion. 2007;21(supplement):305–311. [PubMed]
  • Frank LD, Engelke PO. The built environment and human activity patterns: Exploring the impacts of urban form on public health. Journal of Planning Literature. 2001;16:202–218.
  • Frank LD, Engelke PO, Schmid TL. Health and community design: The impact of the built environment on physical activity. Washington: Island; 2003.
  • Frank LD, Saelens BE, Powell KE, Chapman JE. Stepping towards causation: Do built environments or neighborhood and travel preferences explain physical activity, driving, and obesity? Social Science and Medicine. 2007;65:1898–1914. [PubMed]
  • Frank LD, Sallis JF, Conway TL, Chapman JE, Saelens BE, Bachman W. Many pathways from land use to health: Associations between neighborhood walkability and active transportation, body mass index, and air quality. Journal of the American Planning Association. 2006;72:75–87.
  • Frank LD, Sallis JF, Saelens BE, Leary L, Cain K, Conway TL, Hess PM. The development of a walkability index: Application to the Neighborhood Quality of Life Study. British Journal of Sports Medicine (in press) [PubMed]
  • Frank LD, Schmid TL, Sallis JF, Chapman J, Saelens BE. Linking objectively measured physical activity with objectively measured urban form: Findings from SMARTRAQ. American Journal of Preventive Medicine. 2005;28(2, Suppl 2):117–125. [PubMed]
  • Freedson PS, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Medicine & Science in Sports & Exercise. 1998;30:777–81. [PubMed]
  • Frumkin H, Frank L, Jackson R. The public health impacts of sprawl. Washington, DC: Island Press; 2004.
  • Galea S, Ahern J, Rudenstine S, Wallace Z, Vlahov D. Urban built environment and depression: a multilevel analysis. Journal of Epidemiology and Community Health. 2005;59:822–827. [PMC free article] [PubMed]
  • Gebel K, Bauman AE, Petticrew M. The physical environment and physical activity: A critical appraisal of review articles. American Journal of Preventive Medicine. 2007;32:361–369. [PubMed]
  • Handy SL, Boarnet MG, Ewing R, Killingsworth RE. How the built environment affects physical activity: Views from urban planning. American Journal of Preventive Medicine. 2002;23(2S):64–74. [PubMed]
  • Handy S, Cao X, Mokhtarian PL. Self-selection in the relationship between the built environment and walking. Journal of the American Planning Association. 2006;72:55–74.
  • Handy SL, Cao X, Mokhtarian P. The causal influence of neighborhood design on physical activity within the neighborhood: Evidence from Northern California. American Journal of Health Promotion. 2008;22:350–358. [PubMed]
  • Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, et al. Physical activity and public health: Updated recommendations for adults from the American College of Sports Medicine and the American Heart Association. Medicine & Science in Sports & Exercise. 2007;39:1423–1434. [PubMed]
  • Heath G, Brownson R, Kruger J, Miles R, Powell KE, Ramsey LT. the Task Force on Community Preventive Services. The effectiveness of urban design and land use and transport policies and practices to increase physical activity: A systematic review. Journal of Physical Activity and Health. 2006;3(suppl 1):S55–S76.
  • Kawachi I, Berkman LF. Social ties and mental health. Journal of Urban Health. 2001;78:458–467. [PMC free article] [PubMed]
  • Kerr J, Frank LD, Sallis J, Chapman J. Urban form correlates of pedestrian travel in youth: Differences by gender, race-ethnicity and household attributes. Transportation Research Part D. 2007;12:177–182.
  • Kerr J, Rosenberg D, Sallis JF, Saelens BE, Frank LD, Conway TL. Active commuting to school: Associations with built environment and parental concerns. Medicine and Science in Sports and Exercise. 2006;38:787–794. [PubMed]
  • Kligerman M, Sallis JF, Ryan S, Frank LD, Nader PR. Association of neighborhood design and recreation environment variables with physical activity and body mass index in adolescents. American Journal of Health Promotion. 2007;21:274–277. [PubMed]
  • Levine J, Frank LD. Transportation and land-use preferences and residents’ neighborhood choices: The sufficiency of compact development in the Atlanta region. Transportation: Planning, Policy, Research, & Practice. 2007;34:abstract 01046579.
  • Leslie E, Frank LD, Owen N, Bauman A, Coffee N, Hugo G. Walkability of local communities: Using geographic information systems to objectively assess relevant environmental attributes. Health and Place. 2007;13:111–122. [PubMed]
  • Loukaitou-Sideris A, Eckc JE. Crime prevention and active living. American Journal of Health Promotion. 2007;21(4 suppl):380–389. [PubMed]
  • Mair C, Diez Roux AV, Galea S. Are neighbourhood characteristics associated with depressive symptoms? A review of evidence. Journal of Epidemiology and Community Health. 2008;62:940–946. [PubMed]
  • National Institutes of Health, & National Heart Lung and Blood Institute. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults: The evidence report. Obesity Research. 1998;6:51S–129S. [PubMed]
  • Newman PWG, Kenworthy JR. Transport and urban form in thirty-two of the world’s principal cities. Transportation Reviews. 1991;11:249–272.
  • Owen N, Cerin E, Leslie E, duToit L, Coffee N, Frank LD, Bauman AE, Hugo G, Saelens BE, Sallis JF. Neighborhood walkability and the walking behavior of Australian adults. American Journal of Preventive Medicine. 2007;33:387–395. [PubMed]
  • Owen N, Humpel N, Leslie E, Bauman A, Sallis JF. Understanding environmental influences on walking: Review and research agenda. American Journal of Preventive Medicine. 2004;27:67–76. [PubMed]
  • Papas MA, Alberg AJ, Ewing R, Helzlsouer KJ, Gary TL, Klassen AC. The built environment and obesity. Epidemiologic Reviews. 2007;29:129–43. [PubMed]
  • Radloff LS. The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1:385–401.
  • Saelens BE, Handy SL. Built environment correlates of walking: a review. Medicine & Science in Sports & Exercise. 2008;40(7S):S550–S566. [PMC free article] [PubMed]
  • Saelens BE, Sallis JF, Black J, Chen D. Neighborhood-based differences in physical activity: an environment scale evaluation. American Journal of Public Health. 2003;93:1552–1558. [PMC free article] [PubMed]
  • Saelens BE, Sallis JF, Frank LD. Environmental correlates of walking and cycling: Findings from the transportation, urban design, and planning literatures. Annals of Behavioral Medicine. 2003;25:80–91. [PubMed]
  • Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science. 1997;277:918–924. [PubMed]
  • Transportation Research Board and Institute of Medicine. Special Report. Vol. 282. Washington, DC: National Academies Press; 2005. Does the built environment influence physical activity? Examining the evidence.
  • Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, Macdowell M. Physical activity in the United States measured by accelerometer. Medicine and Science in Sports and Exercise. 2008;40:181–188. [PubMed]
  • United States Department of Health and Human Services. Healthy people 2010. Washington, DC: USDHHS; 2000. (017-001-00547-9)
  • Ware JE, Kosinski M, Keller SD. A 12-item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Medical Care. 1996;34:220–233. [PubMed]
  • Weich S, Blanchard M, Prince M, Burton E, Erens B, Sproston K. Mental health and the built environment: cross-sectional survey of individual and contextual risk factors for depression. British Journal of Psychiatry. 2002;180:428–433. [PubMed]
  • Welk GJ. Use of accelerometry-based activity monitors to assess physical activity. In: Welk GJ, editor. Physical activity assessments for health-related research. Champaign, IL: Human Kinetics; 2002. pp. 125–142.
  • Wood L, Shannon T, Bulsara M, Pikora T, McCormack G, Giles-Corti B. The anatomy of the safe and social suburb: An exploratory study of the built environment, social capital and residents’ perceptions of safety. Health & Place. 2008;14:15–31. [PubMed]
  • World Health Organization. Global strategy on diet, physical activity, and health. Geneva: WHO; 2004. http://www.who.int/dietphysicalactivity/
  • Zhu X, Lee C. Walkability and safety around elementary schools: Economic and ethnic disparities. American Journal of Preventive Medicine. 2008;34:282–290. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

  • PubMed
    PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...