ARTICLE IN PRESS

Health & Place 12 (2006) 656664 www.elsevier.com/locate/healthplace

The link between obesity and the built environment. Evidence from an ecological analysis of obesity and vehicle miles of travel in California
Javier Lopez-Zetinaa,, Howard Leeb, Robert Friisa
a

Health Science Department, California State University, 1250 Bellower Blvd., Long Beach, California, CA 90840-4902, USA b University of California, Los Angeles, CA, USA Received 5 December 2004

Abstract Aims: Obesity and physical inactivity are known to be risk factors for many chronic diseases including hypertension, coronary artery disease, diabetes, and cancer. We sought to explore the association between an indicator of transportation data (Vehicle Miles of Travel, VMT) at the county level as it relates to obesity and physical inactivity in California. Methods: Data from the California Health Interview Survey 2001 (CHIS 2001), the US 2000 Census, and the California Department of Transportation were merged to examine ecological correlations between vehicle miles of travel, population density, commute time, and county indicators of obesity and physical inactivity. Obesity was measured by body mass index (BMI). Physical inactivity was based on self-reported behaviors including walking, bicycling, and moderate to vigorous activity. The unit of analysis was the county. Thirty-three counties in California with population size greater than 100,000 persons per county were retained in the analyses. Results: CHIS 2001 statewide obesity prevalence ranged from 11.2% to 28.5% by county. Physical inactivity ranged from 13.4% to 35.7%. Daily vehicle miles of travel ranged from 3.3 million to 183.8 million per county. By rank bivariate correlation, obesity and physical inactivity were signicantly associated (po0:01). Furthermore, by rank analysis of variance, the highest mean rank obesity was associated with the highest rank of VMT (po0:01). Similar rank patterns were observed between obesity and physical inactivity and commute time. Associations between VMT and physical inactivity were examined but failed to reach statistical signicance. Conclusion: This analysis adds to the growing evidence supporting the association between VMT (a measure of automobile transportation) and obesity. An urban design characterized by over dependence on motorized transportation may be related to adverse health effects. r 2005 Elsevier Ltd. All rights reserved.
Keywords: Obesity; Physical activity; Environmental health; Urban ecology; Land use United States; Social environment; Transportation

Introduction
Corresponding author. Tel.: +1 562 985 1977;

fax: +1 562 985 2384. E-mail address: jlopezze@csulb.edu (J. Lopez-Zetina). 1353-8292/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.healthplace.2005.09.001

Obesity in the US has reached epidemic proportions in the past several decades (United States Public Health Service. Ofce of the Surgeon

ARTICLE IN PRESS
J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664 657

General, 2001; Centers for Disease Control and Prevention, 2000a, b). National estimates of obesity prevalence in the adult population range from 21% by self-reported assessment to 31% by direct measurement (Centers for Disease Control and Prevention, 2001; Flegal et al., 2002). In California, the number of persons considered obese has increased from 10% in 1990 to 18.4% in 2000. The national median of obesity has increased from 11.6% to 20.1% over the same time period (Centers for Disease Control and Prevention, 2001). In absolute numbers, approximately 1.5 million people have been identied as obese, and two million are considered at risk for obesity in California (Diamant et al., 2003). Levels of physical inactivity, a correlate of obesity, are also high in California. The state of California ranks 17th in level of physical inactivity, with approximately one in four residents not engaging in any physical activity in the past month (Centers for Disease Control and Prevention, 2001). Considered a risk factor for hypertension, diabetes, coronary heart disease, osteoarthritis, cancer, and other diseases, (United States Public Health Service. Ofce of the Surgeon General, 2001) obesity is responsible for escalating medical costs. The economic cost of treating obesity and inactivity is estimated to be 70 billion dollars annually, accounting for almost 10% of the health care expenditures in the country (Colditz, 1999). In California, the annual medical cost attributed to obesity is estimated to be 14.2 billion dollars (California Center of Public Health Policy, 2002). At the physiological level, a critical factor in obesity is an imbalance between dietary intake and energy expenditure through physical activity (Flegal et al., 2002). Furthermore, the social environment is seen as contributing to the problem of increased caloric intake with readily available food and larger portions with high fat and calorie content (Young and Nestle, 2002). In addition, sprawling urban areas exacerbate the tendency toward a sedentary lifestyle (Ewing et al., 2002). Although walking has long been proposed as a sensible public measure to promote physical activity, reliance on motorized transportation has increased. An urban space characterized by over dependence on automobile transportation for work, school, shopping and leisure activities, is considered a crucial ecological factor of physical inactivity and obesity (Sherwood and Jeffery, 2000; Frank, 2000; Berrigan and Troiano, 2002). In fact, the growing availability of personal motorized transportation has been found

to correlate with observed secular trends in physical inactivity (James, 1995). Greater reliance on automotive transportation has also been identied as a risk factor for obesity. In one longitudinal study of a Chinese cohort, vehicle ownership was associated with weight gain and greater likelihood of becoming obese (Bell et al., 2002). In the US, recent evidence indicates that the greater the time spent in a car, the greater the likelihood of developing obesity (Frank et al., 2004). Private automobile use continues an upward trend worldwide. Transportation surveys in the US and UK have documented important declines in the amount of walking as a result of overreliance on private automobile transportation (Tudor-Locke et al., 2001). In the US, 75% of all trips less than one mile were taken by car in 2001, while walking as a percentage of all trips taken decreased from 9.3% in 1977, to 7.2% in 1990, and to 5.4% in 1995 (United States. Bureau of Transportation Statistics, 2003). In the UK, the proportion of children walking to school decreased nearly 20% over the period of 1970 to 1991 (Hillman and Policy Studies Institute, 1993). An important indicator to quantify use of motorized transport is vehicle miles of travel (VMT) (Texas Transportation Institute. Texas Department of Transportation, 2002). In California, from 1990 to 2000, total VMT increased nearly 93% to 300 billion. Meanwhile, the states population grew by only 33% over the same period (The Road Information Program, 2001). Although several factors including increased vehicle ownership and an expanding number of eligible drivers account partially for the increase of VMT, the single factor mainly responsible for the increase in VMT in California in the last three decades is the increased number of miles driven per vehicle (Thomas and Deakin, 2001). This upward trend in VMT is inuenced by a number of critical ecological determinants such as population density and urban design; contextual factors which in turn may have an impact on health outcomes at the individual level (Wicker, 1979; Karpati et al., 2002). In California, counties such as San Bernardino and Riverside, with their lack of well-dened, high-density downtown areas have become emblematic of decentered urban forms in the nation (Ewing et al., 2002), while cities like San Francisco are perceived as an alternative to the dominant mode of urban design in the US (Cervero and Gorham, 1995). In the next two decades, San Bernardino and Riverside counties will lead the rest of California counties in urban

ARTICLE IN PRESS
658 J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664

growth rates, which will double those of the next fastest growing counties (Thomas and Deakin, 2001). Given the increase in the VMT required by individuals to cope with extreme urban sprawl in the US, the potential repercussions on human health deserve further attention. We therefore sought to explore the impact of VMT on obesity and physical activity. Specically, we present an ecological analysis of these relationships in select California counties.

Methods Data from the California Health Interview Survey 2001 (CHIS 2001) (University of California, 2003), the California Department of Transportation (Caltrans) (California Department of Transportation, 2000), and the 2000 US Census (Census of Population and Housing, 2000, 2003) were used for this analysis. The CHIS 2001 was implemented to determine the prevalence of and trends in healthrelated behaviors among the general adult, adolescent and child populations; however, the present analysis reports ndings related to the population 18 years of age or older. The CHIS 2001 has a complex sampling design, and survey data were collected by telephone from individuals who met various age participation criteria. Interviewees were randomly selected from 55,428 households in the state. CHIS 2001 was designed for counties with populations of 100,000 or more. Of the 58 California counties surveyed in CHIS, we retained for our analyses 33 urban counties with populations greater than 100,000 persons. Two counties with greater than 100,000 people were excluded because data surveyed from these counties were reported together as a block with two other counties with less than 100,000 individuals. Using these exclusion criteria (counties reported as blocks and counties with population less than 100,000) a total of a total of 25 counties were excluded from the analysis. Caltrans provides various transportation data including county estimates of the average number of vehicle miles traveled each day for the state highway system. The US Census includes demographic and commuting data such as population density, the total working population, and aggregate travel time to work (Census of Population and Housing, 2000, 2003; The Regents of the University of California, 2003). US census data for California were obtained

in aggregate form from Counting California (The Regents of the University of California, 2003). Counties were the unit of analysis examined. Aggregate daily travel time to work (hereafter referred to as commute time) for those traveling 60 minutes or more by non-public means of transportation was used to further explore the relationship between urban transportation and obesity and physical activity. Aggregate indicators of VMT and commute time showed a distribution with various outliers suggesting an extremely skewed, non-normal distribution of the driving experience in metropolitan and urbanized counties in California. Unweighted VMT for the counties ranged from 3.3 million in Madera county to 183.8 million in Los Angeles county. These two counties also represented the lower and upper extremes for unweighted commute time in minutes with a range of 0.18 million to 2.5 million (a complete table with the distribution of commute time by California counties is available at the Counting California website). In order to account for county population size, we weighted VMT by the county working population 16 years of age not working at home. Commute time was weighted by the county working population 16 years of age or older driving a private vehicle to work. These weighted estimates of VMT and commute time as well as population density, BMI and physical inactivity were tested for nonnormality using the KolmogoronovSmirnov test (KS test) (SPSS, Chicago, Ill., 2001) (SPSS Inc., 2001). All variables in this analysis, with the exception of population density, showed evidence of non-normality (p40:10). Health variables from the CHIS 2001 survey include county-specic estimates of the population considered obese by body mass index (BMI), which is calculated from self-reported weight in kilograms divided by the square of height in meters. By CHIS 2001 criteria, a BMI of 30 or higher is considered obese (NHLBI, 1998). The variable of physical inactivity was derived from three CHIS survey items: walking or riding a bicycle to/from work in the past 30 days, and any moderate or vigorous activity during the same time frame. CHIS interviewees were asked to report whether or not they have engaged in this type of physical activity. Those individuals who did not report any of these activities were considered physically inactive. Both, BMI of 30 or greater and physical inactivity are reported as proportion of persons falling under these categories.

ARTICLE IN PRESS
J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664 659

Statistical analysis We examined ecological correlates of obesity and physical inactivity using quartiles of county population density, weighted average VMT, and commute time. Non-parametric statistical methods were used to account for the extreme non-normality of the transportation data, and the relatively small sample size represented by the 33 counties retained in the analysis. Bivariate Spearmans r-rank-correlation test, the MannWhitney test for pairwise comparison of median, and the KruskalWallis analysis of variance test were used to determine associations between VMT, commute time, obesity and physical inactivity. Presentation of statistical results from the

KruskalWallis analysis of variance test follows the approach suggested by Chan and Walsmley (Chan and Walmsley, 1997). The statistical package SPSS (SPSS, Chicago, Ill., 2001) was used to obtain these statistics (SPSS Inc., 2001). Results Crude county obesity estimates from CHIS 2001 ranged from 11.2% in San Francisco county to 28.5% in Merced county, with a statewide obesity prevalence of 18.4%. Physical inactivity county estimates ranged from 13.4% in Marin county to 35.7% in Tulare county (complete list of obesity and physical activity estimates by county are available

Table 1 Obesity and physical inactivity by demographic characteristics Obesity Median Sex Male (1) Female (2) Age Group 1824 (1) 2544 (2) 45+ (3) Race/Ethnicity Latino (1) Native-American() Asian (2) African-American(3) White (4) Other () Subgroup comparison Sex S12 Age Group A12 A13 A23 Race/ethnicity R12 R13 R14 R23 R24 R34 W p Physical inactivity Median W p

20 18 12 21 22 27 29 6 33 18 24

22 33 15 24 33 41 27 34 34 19 24

1203.5 665.0 586.5 985.5 1125.0 713.0 1390.5 191.0 203.0 1135.0

0.211 o0.001 o0.001 0.012 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001

690.0 696.0 541.0 747.5 1217.5 1186.5 1638.5 890.5 1380.0 1014.0

o0.001 o0.001 o0.001 o0.001 0.026 o0.001 o0.001 0.398 o0.001 o0.001

Results of a MannWhitney comparison test of demographic subgroups. Thirty-three California counties, 2001. Note: S12, A12A23, R12R34 Pairwise comparison of demographic subgroups indicated with numerals in parentheses. W MannWhitney statistics of pairwise comparison. p p-value of pairwise comparison. Due to small sample size, obesity and physical inactivity median estimates for respondents reporting Native American and Other race/ethnicity are based on statistically unstable gures. Native American and Other race/ethnicity category were excluded from the MannWhitney test of signicance.

ARTICLE IN PRESS
660 J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664

from CHIS 2001). Obesity and physical inactivity levels were examined by demographic characteristics (Table 1). The median obesity was higher for males when compared to females, although this difference was not statistically signicant. The median physical inactivity among females was higher than that of males. Although the median obesity for the groups aged 2544 years and 45 and older did not differ greatly, their median obesity was almost twice the value as those 1824 years. The youngest age bracket 1824 years also had the lowest median physical inactivity level; with the high level of physical inactivity found among older Californians (Table 1). By race/ethnicity, the highest median obesity was found among AfricanAmericans, followed by Native-Americans, Latinos, people of Other race/ethnicity, the whites. The

Table 2 Rank bivariate correlation of obesity and physical inactivity and urban ecological variables Inactivity Obesity Density Inactivity 1 Obesity Density VMT Commute 0.635 1 VMT Commute

0.186 0.061 0.516 0.342 0.792 0.545 0.081 1 0.756 1 0.299 1

Thirty-three California counties, 20002001. VMT: vehicle miles of travel. Commute: Commute time. po0:01.

lowest median obesity was found among AsianAmericans. By race/ethnicity, the highest median physical inactivity was found among Latinos and the lowest among whites. Table 2 shows bivariate rank correlation coefcients between obesity, physical inactivity, and the urban ecological variables of population density, VMT, and commute time. The largest correlation was found between VMT and proportion of obesity. VMT was also strongly negatively associated with population density. Physical inactivity was highly correlated with obesity, and population density was negatively correlated with increased obesity, though this last association was not as strong. Rank oneway analysis of variance with the KruskalWallis test was performed to further characterize the relationships between these variables (Table 3). The association between VMT and obesity was found to be signicant, and a clear gradient between VMT and obesity was observed in this analysis. Four of six pairwise quartiles comparisons were found to be statistically signicant. A gradient between obesity and commuting was observed but only the pairwise comparison between the rst and third quartiles was found to be statistically signicant. Physical inactivity and commuting generally showed the same gradient, except for the third quartile, which exhibited the highest mean rank. When pairwise comparisons of quartiles of commuting and physical inactivity were analyzed, two of six quartiles showed evidence of statistical signicance.

Table 3 KruskalWallis ranks and mean rank differences between obesity and physical inactivity and quartiles of vehicle miles of travel and commuting Mean rank and vehicle miles of travel Q1 Obesity 9.50 P. Inactivity 13.67 Q2 14.30 14.0 Q3 18.05 16.95 Q4 27.08 25.42 H 11.0 6.2 df 3 3 p 0.01 0.10 Mean rank differences between quartiles of vehicle miles of travel Q1 vs Q2 2.4 Q1 vs. Q3 4.9 Q1 vs. Q4 5.6 Q2 vs. Q3 2.3 Q2 vs. Q4 6.4 Q4 vs. Q3 6.1

Mean Rank and commuting Obesity P. Inactivity 9.13 8.94 15.19 12.88 21.33 25.39 21.81 19.75 9.3 14.4 3 3 0.02 0.01

Mean rank differences between quartiles of commuting 3.2 1.7 6.4 8.0 3.4 6.6 3.4 6.6 3.2 3 0.59 3.1

Thirty-three California counties, 20002001. Note. Q1 First Quartile, Q2 Second Quartile, Q3 Third Quartile, Q4 Fourth Quartile. P. Inactivity Physical Inactivity. po0.01. po0.05.

ARTICLE IN PRESS
J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664 661

Discussion Our analysis found obesity to be associated with urban transportation indicators related to automobile use. Recent research on obesity and urban housing patterns has reported greater levels of obesity in communities where the private automobile is the dominant means of transportation (Brownson et al., 2001; Craig et al., 2002; Eyler et al., 2003; Saelens et al., 2003). Over dependence on the automobile has been in turn linked to urban sprawl where low residential density and lack of mixed land uses are common (Cervero, 1989; Ewing et al., 2003). This study did not identify a correlation between physical inactivity and the transportation variables examined, however, given the statistically signicant ecological correlation between obesity and physical inactivity, further exploration of VMT and physical inactivity may be warranted. Further research will be needed to elucidate the specic ecological mechanisms, mediators and correlates of physical inactivity as interest in this issue grows (Bauman et al., 2002). Although the link between urban design and reduced opportunity for walking continues to be the subject of ongoing reports (Dannenberg et al., 2003; Frumkin, 2002; Frumkin, 2003), there still remain unresolved issues regarding the consistency of such associations at the ecological level (Handy, 1996; Boarnet and Crane, 2001). For example, even though large metropolitan regions may have small neighborhoods with pedestrianfriendly street blocks, high population density, and mixed-use land development, such areas may still be too highly auto-dependent to signicantly promote routine physical activity (Cervero and Gorham, 1995). Despite these inconsistencies, a land use pattern appears to emerge as a dominant feature of the urban landscape in metropolitan areas in California: a network of freeways built to accommodate vehicle transportation, and suburban housing designed to lter transit and pedestrian life out of these neighborhoods (Cervero and Gorham, 1995). The conuence of land use and transportation modalities for the development of reduced opportunities for physical activity may have further public health implications when demographic factors are considered. In the present analysis, Latinos, the fastest growing demographic group in California, and older Californians exhibited the highest level of physical inactivity. These subgroups have been previously identied as populations with signicant

barriers to access to recreational facilities (Carvajal et al., 2002; Perez-Stable et al., 1994; Yancey et al., 2004) where they may encounter opportunities for physical activityopportunities that are otherwise unavailable in their neighborhoods. Conceptual limitations in understanding the relationship between urban form and physical inactivity, other methodological concerns should be considered when performing ecological analyses. Studies linking aggregate estimates of contextual factors to health data collected at the individual level are often criticized because of the potential for ecological fallacies. Furthermore, the associations uncovered are derived from cross-sectional data, and causality cannot be presumed. Ecological studies suggest rather than provide denite answers for the associations among complex factors related to the urban environment. However, contextual studies may be more appropriate than person-based studies when risk factors (e.g. urban design) operate at a population level. Another limitation of this study is our focus on urban counties. Although obesity is prevalent in both urban and rural areas, previous research has suggested that rural areas exhibit different demographic characteristics and ecological and contextual factors in relation to physical activity (Berrigan and Troiano, 2002), and should be examined separately. The transportation data in this analysis also suffer from some limitations. One key variable in this analysis, VMT, reects only vehicle-miles traveled on state highways. Surface street trafc is not represented in the number of vehicle-miles accrued and reported annually in each county. However, VMT data appear better suited to conveying the dynamics of private transportation in states like California where highway driving is common. In fact, the urban transportation experience in California often involves inter-county travel for work and leisure activities. For example, the two-county area of San BernardinoRiverside, which has the highest score on the urban sprawl index (Ewing et al., 2002) and one of the lowest rank in mass transportation modalities (Cervero and Gorham, 1995), has a population which often commutes among the contiguous counties of Orange, Los Angeles, Ventura and San Diego (Thomas and Deakin, 2001). Further research on other urban transportation indicators is needed to rene our understanding of transport modalities and reduced opportunities for physical activity. For example, the San BernardinoRiverside area has

ARTICLE IN PRESS
662 J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664

seen an increase in not only overall VMT, but also the duration of rush-hour trafc from 3.4 hour in 1982 to 7.6 hour in 2000. In addition, in this area, congested travel as a proportion of peak VMT has increased from 23% in 1982 to 72% in 2000, a net increase of 213%. In contrast, the increase in congested travel in San Francisco during the same period was only 54% (Texas Transportation Institute. Texas Department of Transportation, 2002). Despite these limitations, our report is the rst to provide a preliminary analysis of obesity, physical inactivity and ecological indicators of transportation data in California, using county-specic, population-based estimates. Recent national studies with more complex measures of suburban housing and transportation have also provided evidence of the link between the suburban environment and obesity (Ewing et al., 2003); however, those reports are more difcult to interpret because they are based on obesity and physical activity estimates from survey data not originally designed to provide representative samples at the county level. Furthermore, our conservative analytical statistical approach made no assumptions about the underlying distribution of the data, and therefore provides more condence about the associations observed. Although based on rst-order statistical analyses and a limited number of contextual variables, the present study provides preliminary evidence supporting the notion that the ecological urban design in California, dened by excessive dependence on the automobile, is associated with high levels of obesity and physical inactivity. The predominant US suburban housing form, with its over dependence on the private automobile, developed over six decades of sprawling urban growth (Melosi, 2000), as did the obesity epidemic (Posner et al., 1995). In 2000, the prevalence of overweight among US adult men and women was 67% and 62%, respectively; but as far back as 40 years ago, already half of the adult male population and 40% of the female population were overweight (Flegal et al., 2002). Given the association between obesity and physical inactivityat both the individual as well as the ecological levelefforts to reverse the obesity epidemic will require an interdisciplinary research approach between urban planners, public health researchers, and policy makers with the ultimate goal of identifying strategies and incentives that accommodate walking and other forms of physical activity in daily life. Emerging environmental paradigms for obesity

prevention and research (Centers for Disease Control and Prevention, 2003; The Robert Wood Johnson Foundation, 2000), will be required to complement ve decades worth of obesity-prevention guidelines mostly focused on changing individual behavior (Nestle and Jacobson, 2000).

Acknowledgements We thank Lee Habte, CHIS 2001 Data Access Center Manager, UCLA Center for Health Policy Research for her assistance and collaboration. This paper was presented in part, as a poster presentation at the 2004 Preventive Medicine Conference sponsored by the American College of Preventive Medicine, 1822 February 2004, in Orlando, FL.

References
Bauman, A.E., Sallis, J.F., Dzewaltowski, D.A., Owen, N., 2002. Toward a better understanding of the inuences on physical activity: the role of determinants, correlates, causal variables, mediators, moderators, and confounders. American Journal of Preventive Medecine 23, 514. Bell, A.C., Ge, K., Popkin, B.M., 2002. The road to obesity or the path to prevention: motorized transportation and obesity in China. Obesity Research 10, 277283. Berrigan, D., Troiano, R.P., 2002. The association between urban form and physical activity in US adults. American Journal of Preventive Medecine 23, 7479. Boarnet, M., Crane, R., 2001. The inuence of land use on travel behavior: specication and estimation strategies. Transportation Research Part A, 823845. Brownson, R.C., Baker, E.A., Housemann, R.A., Brennan, L.K., Bacak, S.J., 2001. Environmental and policy determinants of physical activity in the United States. American Journal of Public Health 91, 19952003. California Center of Public Health Policy, 2002. An Epidemic: Overweight and Unt Children in California Assembly Districts. California Department of Transportation, 2000. CALTRANS. Vehicle Miles of Travel by County on the State Highway System (in Millions) Year County 2001, 2000, 1999. Carvajal, S.C., Hanson, C.E., Romero, A.J., Coyle, K.K., 2002. Behavioural risk factors and protective factors in adolescents: a comparison of Latinos and non-Latino whites. Ethnicity & Health 7, 181193. Census of Population and Housing 2000, 2003. Census of Population and Housing, 2000 [California]: Summary File 3 [California] [on-line data le]/prepared by the California Digital Library, Oakland, CA. Centers for Disease Control and Prevention, 2000a. 2000 BRFSS Summary Prevalence Report. US Department of Health and Human Services, Atlanta, GA.

ARTICLE IN PRESS
J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664 Centers for Disease Control and Prevention, 2000b. Behavioral Risk Factor Surveillance System Survey Data. US Department of Health and Human Services, Atlanta, GA. Centers for Disease Control and Prevention, 2001. Behavioral Risk Factor Surveillance System Survey Data. US Department of Health and Human Services, Atlanta, GA. Centers for Disease Control and Prevention, 2003. ACES: Active Community Environments Initiative. National Center for Chronic Disease Prevention and Health Promotion. Division of Nutrition and Physical Activity. Physical Activity and Health Branch. Cervero, R., 1989. Americas Suburban Centers: The Land Use Transportation Link. Unwin Hyman, Boston. Cervero, R., Gorham, R., 1995. Commuting in transit versus automobile nighborhoods. Journal of American Planning Association, 210225. Chan, Y., Walmsley, R.P., 1997. Learning and understanding the KruskalWallis one-way analysis-of-variance-by-ranks test for differences among three or more independent groups. Physical Therapy 77, 17551762. Colditz, G.A., 1999. Economic costs of obesity and inactivity. Medicine and Science in Sports and Exercise 31, S663S667. Craig, C.L., Brownson, R.C., Cragg, S.E., Dunn, A.L., 2002. Exploring the effect of the environment on physical activity: a study examining walking to work. American Journal of Preventive Medicine 23, 3643. Dannenberg, A.L., Jackson, R.J., Frumkin, H., Schieber, R.A., Pratt, M., Kochtitzky, C., Tilson, H.H., 2003. The impact of community design and land-use choices on public health: a scientic research agenda. American Journal of Public. Health 93, 15001508. Diamant, A.L., Babey, S.H., Brown, E.R., Chawla, N., 2003. Diabetes in California: nearly 1.5 million diagnosed and 2 million more at risk. Policy Brief of UCLA Center for Health Policy Research, 18. Ewing, R., Pendall, R., Chen, D., 2002. Measuring Sprawl and its Impact. Smartgrowth America, Washington, DC. Ewing, R., Schmid, T., Killingsworth, R., Zlot, A., Raudenbush, S., 2003. Relationship between urban sprawl and physical activity, obesity, and morbidity. American Journal of Health Promotion 18, 4757. Eyler, A.A., Brownson, R.C., Bacak, S.J., Housemann, R.A., 2003. The epidemiology of walking for physical activity in the United States. Medicine and Science in Sports and Exercise 35, 15291536. Flegal, K.M., Carroll, M.D., Ogden, C.L., Johnson, C.L., 2002. Prevalence and trends in obesity among US adults, 19992000. JAMAJournal of the American Medical Association 288, 17231727. Frank, L.D., 2000. Land use and transportation interactions: implications in public health and quality of life. Journal of Planning Education Research, 622. Frank, L.D., Andresen, M.A., Schmid, T.L., 2004. Obesity relationships with community design, physical activity, and time spent in cars. American Journal of Preventive Medicine 27, 8796. Frumkin, H., 2002. Urban sprawl and public health. Public Health Report 117, 201217. Frumkin, H., 2003. Healthy places: exploring the evidence. American Journal of Public Health 93, 14511456. Handy, S., 1996. Urban forms and pedestrian choices. Transportation Research Record 1, 135144. 663 Hillman, M., Policy Studies Institute, 1993. Children, Transport and the Quality of Life. Policy Studies Institute, London. James, W.P., 1995. A public health approach to the problem of obesity. International Journal of Obesity Related Metabolic Disorders 19 (Suppl 3), S37S45. Karpati, A., Galea, S., Awerbuch, T., Levins, R., 2002. Variability and vulnerability at the ecological level: implications for understanding the social determinants of health. American Journal of Public Health 92, 17681772. Melosi, M.V., 2000. The Sanitary City: Urban Infrastructure in America from Colonial Times to the Present. Johns Hopkins University Press, Baltimore. Nestle, M., Jacobson, M.F., 2000. Halting the obesity epidemic: a public health policy approach. Public Health Report 115, 1224. NHLBI, N.H., Lung, Blood and Heart Institute, 1998. Clinical guidelines on the identication, evaluation, and treatment of overweight and obesity in adultsthe evidence report. National Institutes of Health. Obesity Research 6 (Suppl 2), 51S209S. Perez-Stable, E.J., Marin, G., Marin, B.V., 1994. Behavioral risk factors: a comparison of Latinos and non-Latino whites in San Francisco. American Journal of Public Health 84, 971976. Posner, B.M., Franz, M.M., Quatromoni, P.A., Gagnon, D.R., Sytkowski, P.A., Dagostino, R.B., Cupples, L.A., 1995. Secular trends in diet and risk factors for cardiovascular disease: the Framingham Study. Journal of American Dieticians Association 95, 171179. Saelens, B.E., Sallis, J.F., Frank, L.D., 2003. Environmental correlates of walking and cycling: ndings from the transportation, urban design, and planning literatures. Annals of Behavioral Medicine 25, 8091. Sherwood, N.E., Jeffery, R.W., 2000. The behavioral determinants of exercise: implications for physical activity interventions. Annual Review of Nutrition 20, 2144. SPSS INC, 2001. SPSS for Windows, release 11.0.1. SPSS, Chicago, IL. Texas Transportation Institute. Texas Department of Transportation, 2002. Urban Mobility Study. Texas Transportation Institute, College Station, TX. The Regents of the University of California, 2003. Counting California. The Road Information Program, 2001. Californias Roads and Highways: Conditions and Travel Trends. Washington, DC. The Robert Wood Johnson Foundation, 2000. Healthy people, healthy places: promoting public health and physical activity through community design. In: Report of an Experts Meeting, 2728 November 2000, Washington, DC. Thomas, J.V., Deakin, E., 2001. California Demographic Trends: Implications for Transportation Planning. University of California Transportation Center, Berkeley, CA. Tudor-Locke, C., Ainsworth, B.E., Popkin, B.M., 2001. Active commuting to school: an overlooked source of childrens physical activity? Sports Medicine 31, 309313. United States Public Health Service. Ofce of the Surgeon General, 2001. The Surgeon Generals Call to Action to Prevent and Decrease Overweight and Obesity. US Department of Health and Human Services, Public Health Service, Rockville, MD.

ARTICLE IN PRESS
664 J. Lopez-Zetina et al. / Health & Place 12 (2006) 656664 Yancey, A.K., Wold, C.M., Mccarthy, W.J., Weber, M.D., Lee, B., Simon, P.A., Fielding, J.E., 2004. Physical inactivity and overweight among Los Angeles County adults. American Journal of Preventive Medicine 27, 146152. Young, L.R., Nestle, M., 2002. The contribution of expanding portion sizes to the US obesity epidemic. American Journal of Public Health 92, 246249. United States. Bureau Of Transportation Statistics, 2003. 2001 Highlights Report. US Department of Transportation, Washington, DC. University of California, C. F. H. P. R., 2003. California Health Interview Survey, 2001. Sample Design Description for CHIS 2001. University of California, Los Angeles, CA. Wicker, A.W., 1979. An Introduction to Ecological Psychology. Brooks/Cole Publication Company, Monterey, CA.