Walkability measures to predict the likelihood of walking in a place: A classification and regression tree analysis.

Walkability is a popular and ubiquitous term at the intersection of urban planning and public health. As the number of potential walkability measures grows in the literature, there is a need to compare their relative importance for specific research objectives. This study demonstrates a classification and regression tree (CART) model to compare five familiar measures of walkability from the literature for their relative ability to predict whether or not walking occurs in a dataset of objectively measured locations. When analyzed together, the measures had moderate-to-high accuracy (87.8% agreement: 65.6% of true walking GPS-measured points classified as walking and 93.4% of non-walking points as non-walking). On its own, the most well-known composite measure, Walk Score, performed only slightly better than measures of the built environment composed of a single variable (transit ridership, employment density, and residential density).Thus there may be contexts where transparent and longitudinally available measures of urban form are worth a marginal tradeoff in prediction accuracy. This comparison of walkability measures using CART highlights the importance for public health and urban design researchers to think carefully about how and why particular walkability measures are used.

A cross-sectional analysis of physical activity and weight misreporting in diverse populations: The Seattle Obesity Study III.

BACKGROUND: In-person assessments of physical activity (PA) and body weight can be burdensome for participants and cost prohibitive for researchers. This study examined self-reported PA and weight accuracy and identified patterns of misreporting in a diverse sample. METHODS: King, Pierce and Yakima county residents, aged 21-59 years (n = 728), self-reported their moderate-to-vigorous PA (MVPA) and weight, in kilograms. Self-reports were compared with minutes of bout-level MVPA, from 3 days of accelerometer data, and measured weights. Regression models examined characteristics associated with underreporting and overreporting of MVPA and weight, the potential bias introduced using each measure and the relation between perceived and measured PA and weight. RESULTS: MVPA underreporting was higher among males and college educated participants; however, there was no differential MVPA overreporting. Weight underreporting was higher among males, those age 40-49 years and persons with obesity. Weight overreporting was higher among Hispanic participants and those reporting stress, unhappiness and fair or poor health. The estimated PA-obesity relation was similar using measured and self-reported PA but not self-reported weight. Perceived PA and weight predicted measured values. CONCLUSION: Self-reported PA and weight may be useful should objective measurement be infeasible; however, though population-specific adjustment for differential reporting should be considered.

Higher residential and employment densities are associated with more objectively measured walking in the home neighborhood.

Understanding where people walk and how the built environment influences walking is a priority in active living research. Most previous studies were limited by self-reported data on walking. In the present study, walking bouts were determined by integrating one week of accelerometry, GPS, and a travel log data among 675 adult participants in the baseline sample of the Travel Assessment and Community study. Home neighborhood was defined as being within 0.5 mi of each participants' residence (a 10-minute walk), with home neighborhood walking defined as walking bout lines with at least one GPS point within the home neighborhood. Home neighborhood walkability was constructed with seven built environment variables derived from spatially continuous objective values (SmartMaps). A Zero Inflated Negative Binomial (ZINB) served to estimate associations between home neighborhood environment characteristics and home neighborhood walking frequency. Higher residential density and job density were the two neighborhood walkability measures related to higher likelihood and more time walking in the home neighborhood, highest tertile residential density (22.44 - 62.63 unit/acre) (coefficient=1.434; 95th CI of 1.003, 2.049) and highest tertile job density (12.4 - 272.3 jobs/acre) (coefficient=1.616; 95th CI of 1.102, 2.370). The large proportion of walking that takes place in the home neighborhood highlights the importance of continuing to examine the impact of the home neighborhood environment on walking. Potential interventions to increase walking behavior may benefit from increasing residential and employment density within residential areas.

Beyond the bus stop: where transit users walk.

OBJECTIVES: Extending the health benefits of public transit requires understanding how transit use affects pedestrian activity, including pedestrian activity not directly temporally or spatially related to transit use. In this study, we identified where transit users walked on transit days compared with non-transit days within and beyond 400m and 800m buffers surrounding their home and work addresses. METHODS: We used data collected from 2008-2013 in King County, Washington, from 221 non-physically-disabled adult transit users, who were equipped with an accelerometer, global positioning system (GPS), and travel diary. We assigned walking activity to the following buffer locations: less than and at least 400m or 800m from home, work, or home/work (the home and work buffers comprised the latter buffer). We used Poisson generalized estimating equations to estimate differences in minutes per day of total walking and minutes per day of non-transit-related walking on transit days compared with non-transit days in each location. RESULTS: We found that durations of total walking and non-transit-related walking were greater on transit days than on non-transit days in all locations studied. When considering the home neighborhood in isolation, most of the greater duration of walking occurred beyond the home neighborhood at both 400m and 800m; results were similar when considering the work neighborhood in isolation. When considering the neighborhoods jointly (i.e., by using the home/work buffer), at 400m, most of the greater duration of walking occurred beyond the home/work neighborhood. However, at 800m, most of the greater duration of walking occurred within the home/work neighborhood. CONCLUSIONS: Transit days were associated with greater durations of total walking and non-transit related walking within and beyond the home and work neighborhoods. Accordingly, research, design, and policy strategies focused on transit use and pedestrian activity should consider locations outside the home and work neighborhoods, in addition to locations within them.

Capturing fine-scale travel behaviors: a comparative analysis between personal activity location measurement system (PALMS) and travel diary.

BACKGROUND: Device-collected data from GPS and accelerometers for identifying active travel behaviors have dramatically changed research methods in transportation planning and public health. Automated algorithms have helped researchers to process large datasets with likely fewer errors than found in other collection methods (e.g., self-report travel diary). In this study, we compared travel modes identified by a commonly used automated algorithm (PALMS) that integrates GPS and accelerometer data with those obtained from travel diary estimates. METHODS: Sixty participants, who made 2100 trips during seven consecutive days of data collection, were selected from among the baseline sample of a project examining the travel behavior impact of a new light rail system in the greater Seattle, WA (USA) area. GPS point level analyses were first conducted to compare trip/place and travel mode detection results using contingency tables. Trip level analyses were then performed to investigate the effect of proportions of time overlap between travel logs and device-collected data on agreement rates. Global performance (with all subjects' data combined) and subject-level performance of the algorithm were compared at the trip level. RESULTS: At the GPS point level, the overall agreement rate of travel mode detection was 77.4% between PALMS and the travel diary. The agreement rate for vehicular trip detection (84.5%) was higher than for bicycling (53.5%) and walking (58.2%). At the trip level, the global performance and subject-level performance of the PALMS algorithm were 46.4% and 42.4%, respectively. Vehicular trip detection showed highest agreement rates in all analyses. Study participants' primary travel mode and car ownership were significantly related to the subject-level mode agreement rates. CONCLUSIONS: The PALMS algorithm showed moderate identification power at the GPS point level. However, trip level analyses found lower agreement rates between PALMS and travel diary data, especially for active transportation. Testing different PALMS parameter settings may serve to improve the detection of active travel and help expand PALMS's applicability in geographically different urbanized areas with a variety of travel modes.

Increased Walking's Additive and No Substitution Effect on Total Physical Activity.

PURPOSE: We assessed the associations between a change in time spent walking and a change in total physical activity (PA) time within an urban living adult sample to test for additive or substitution effects. METHODS: Participants living in the greater Seattle area were assessed in 2008-2009 and again 1-2 yr later (2010-2011). At each time point, they wore accelerometers and GPS units and recorded trips and locations in a travel diary for seven consecutive days. These data streams were combined to derive a more objective estimate of walking and total PA. Participants also completed the International Physical Activity Questionnaire to provide self-reported estimates of walking and total PA. Regression analyses assessed the associations between within-participant changes in objective and self-reported walking and total PA. RESULTS: Data came from 437 participants. On average, a 1-min increase in total walking was associated with an increase in total PA of 1 min, measured by objective data, and 1.2-min, measured by self-reported data. A similar additive effect was consistently found with utilitarian, transportation, or job-related walking, measured by both objective and self-reported data. For recreational walking, the effect of change was mixed between objective and self-reported results. CONCLUSION: Both objective and self-reported data confirmed an additive effect of utilitarian and total walking on PA.

Light Rail Leads to More Walking Around Station Areas.

Areas around Light Rail Transit (LRT) stations offer ideal conditions for Transit-Oriented Development (TOD). Relatively dense, mixed-use neighborhoods can have positive impacts on mobility, health, and perceptions of neighborhood safety among nearby residents, primarily through walking activity for both transit and other purposes. To examine how station areas may attract new activity, this study analyzed changes in walking around station areas among people living close to an LRT station before and after the opening of a new transit system. This study examined walking behavior among the subset of 214 participants living within one mile of one of 13 LRT stations from among a sample of residents living close or further away from a new LRT line in Seattle. They completed a survey and a travel log and wore an accelerometer and a GPS for 7 days both before (2008) and after the opening of the Seattle area LRT (2010). Walking bouts were derived using a previously developed algorithm. The main outcome was the individual-level change in the proportion of daily walking within one quarter Euclidean mile of an LRT station. Overall walking decreased from before to after the LRT opening while station area walking did not change significantly, indicating a shift in walking activity to the station areas after the introduction of LRT. Increases in the proportion of station area walking were negatively related to participants' distance between home and the nearest LRT station, peaking at <0.25 mile and decaying beyond >0.75 mile. Male gender, college education, normal weight status, less access to cars, and frequent LRT use were also significantly associated with greater positive changes in the proportion of station area walking. The shift in walking to station areas after the completion of light rail provides evidence that the local proximate population is attracted to station areas, which may potentially benefit both transit use and TOD area economic activity. The residential catchment area for the shift in LRT area walking was < 0.75 mile of the LRT stations.

The Relationship Between Objectively Measured Walking and Risk of Pedestrian-Motor Vehicle Collision.

Safe urban walking environments may improve health by encouraging physical activity, but the relationship between an individual's location and walking pattern and the risk of pedestrian-motor vehicle collision is unknown. We examined associations between individuals' walking bouts and walking risk, measured as mean exposure to the risk of pedestrian-vehicle collision. Walking bouts were ascertained through integrated accelerometry and global positioning system data and from individual travel-diary data obtained from adults in the Travel Assessment and Community Study (King County, Washington) in 2008-2009. Walking patterns were superimposed onto maps of the historical probabilities of pedestrian-vehicle collisions for intersections and midblock segments within Seattle, Washington. Mean risk of pedestrian-vehicle collision in specific walking locations was assessed according to walking exposure (duration, distance, and intensity) and participant demographic characteristics in linear mixed models. Participants typically walked in areas with low pedestrian collision risk when walking for recreation, walking at a faster pace, or taking longer-duration walks. Mean daily walking duration and distance were not associated with collision risk. Males walked in areas with higher collision risk compared with females, while vehicle owners, residents of single-family homes, and parents of young children walked in areas with lower collision risk. These findings may suggest that pedestrians moderate collision risk by using lower-risk routes.

Differences in Behavior, Time, Location, and Built Environment between Objectively Measured Utilitarian and Recreational Walking.

OBJECTIVES: Utilitarian and recreational walking both contribute to physical activity. Yet walking for these two purposes may be different behaviors. We sought to provide operational definitions of utilitarian and recreational walking and to objectively measure their behavioral, spatial, and temporal differences in order to inform transportation and public health policies and interventions. METHODS: Data were collected 2008-2009 from 651 Seattle-King County residents, wearing an accelerometer and a GPS unit, and filling-in a travel diary for 7 days. Walking activity bouts were classified as utilitarian or recreational based on whether walking had a destination or not. Differences between the two walking purposes were analyzed, adjusting for the nested structure of walking activity within participants. RESULTS: Of the 4,905 observed walking bouts, 87.4% were utilitarian and 12.6% recreational walking. Utilitarian walking bouts were 45% shorter in duration (-12.1 min) and 9% faster in speed (+0.3km/h) than recreational walking bouts. Recreational walking occurred more frequently in the home neighborhood and was not associated with recreational land uses. Utilitarian walking occurred in areas having higher residential, employment, and street density, lower residential property value, higher area percentage of mixed-use neighborhood destinations, lower percentage of parks/trails, and lower average topographic slope than recreational walking. CONCLUSION: Utilitarian and recreational walking are substantially different in terms of frequency, speed, duration, location, and related built environment. Policies that promote walking should adopt type-specific strategies. The high occurrence of recreational walking near home highlights the importance of the home neighborhood for this activity.

Multilevel models for evaluating the risk of pedestrian-motor vehicle collisions at intersections and mid-blocks.

Walking is a popular form of physical activity associated with clear health benefits. Promoting safe walking for pedestrians requires evaluating the risk of pedestrian-motor vehicle collisions at specific roadway locations in order to identify where road improvements and other interventions may be needed. The objective of this analysis was to estimate the risk of pedestrian collisions at intersections and mid-blocks in Seattle, WA. The study used 2007-2013 pedestrian-motor vehicle collision data from police reports and detailed characteristics of the microenvironment and macroenvironment at intersection and mid-block locations. The primary outcome was the number of pedestrian-motor vehicle collisions over time at each location (incident rate ratio [IRR] and 95% confidence interval [95% CI]). Multilevel mixed effects Poisson models accounted for correlation within and between locations and census blocks over time. Analysis accounted for pedestrian and vehicle activity (e.g., residential density and road classification). In the final multivariable model, intersections with 4 segments or 5 or more segments had higher pedestrian collision rates compared to mid-blocks. Non-residential roads had significantly higher rates than residential roads, with principal arterials having the highest collision rate. The pedestrian collision rate was higher by 9% per 10 feet of street width. Locations with traffic signals had twice the collision rate of locations without a signal and those with marked crosswalks also had a higher rate. Locations with a marked crosswalk also had higher risk of collision. Locations with a one-way road or those with signs encouraging motorists to cede the right-of-way to pedestrians had fewer pedestrian collisions. Collision rates were higher in locations that encourage greater pedestrian activity (more bus use, more fast food restaurants, higher employment, residential, and population densities). Locations with higher intersection density had a lower rate of collisions as did those in areas with higher residential property values. The novel spatiotemporal approach used that integrates road/crossing characteristics with surrounding neighborhood characteristics should help city agencies better identify high-risk locations for further study and analysis. Improving roads and making them safer for pedestrians achieves the public health goals of reducing pedestrian collisions and promoting physical activity.

Associations Between Fast-Food Consumption and Body Mass Index: A Cross-Sectional Study in Adult Twins.

Obesity is a substantial health problem in the United States, and is associated with many chronic diseases. Previous studies have linked poor dietary habits to obesity. This cross-sectional study aimed to identify the association between body mass index (BMI) and fast-food consumption among 669 same-sex adult twin pairs residing in the Puget Sound region around Seattle, Washington. We calculated twin-pair correlations for BMI and fast-food consumption. We next regressed BMI on fast-food consumption using generalized estimating equations (GEE), and finally estimated the within-pair difference in BMI associated with a difference in fast-food consumption, which controls for all potential genetic and environment characteristics shared between twins within a pair. Twin-pair correlations for fast-food consumption were similar for identical (monozygotic; MZ) and fraternal (dizygotic; DZ) twins, but were substantially higher in MZ than DZ twins for BMI. In the unadjusted GEE model, greater fast-food consumption was associated with larger BMI. For twin pairs overall, and for MZ twins, there was no association between within-pair differences in fast-food consumption and BMI in any model. In contrast, there was a significant association between within-pair differences in fast-food consumption and BMI among DZ twins, suggesting that genetic factors play a role in the observed association. Thus, although variance in fast-food consumption itself is largely driven by environmental factors, the overall association between this specific eating behavior and BMI is largely due to genetic factors.

How far from home? The locations of physical activity in an urban U.S. setting.

UNLABELLED: Little is known about where physical activity (PA) occurs, or whether different demographic groups accumulate PA in different locations. METHOD: Objective data on PA and location from 611 adults over 7days were collected in King County, WA in 2008-2009. The relative amounts of time spent in sedentary-to-low and moderate-to-vigorous PA (MVPA) were quantified at three locations: "home" (<125m from geocoded home locations); "near" home (125-1666m, defining the home neighborhood); and "away" from home (>1666m). Differences in MVPA by demographics and location were examined. The percent of daily time in MVPA was estimated using a mixed model adjusted for location, sex, age, race/ethnicity, employment, education, BMI, and income. RESULTS: Most MVPA time occurred in nonhome locations, and disproportionately "near" home; this location was associated with 16.46% greater time in MVPA, compared to at-home activity (p<0.001), whereas more time spent at "away" locations was associated with 3.74% greater time in MVPA (p<0.001). Location was found to be a predictor of MVPA independent of demographic factors. CONCLUSION: A large proportion of MVPA time is spent at "near" locations, corresponding to the home neighborhood studied in previous PA research. "Away" locations also host time spent in MVPA and should be the focus of future research.

The geography of diabetes by census tract in a large sample of insured adults in King County, Washington, 2005-2006.

INTRODUCTION: Identifying areas of high diabetes prevalence can have an impact on public health prevention and intervention programs. Local health practitioners and public health agencies lack small-area data on obesity and diabetes. METHODS: Clinical data from the Group Health Cooperative health care system were used to estimate diabetes prevalence among 59,767 adults by census tract. Area-based measures of socioeconomic status and the Modified Retail Food Environment Index were obtained at the census-tract level in King County, Washington. Spatial analyses and regression models were used to assess the relationship between census tract-level diabetes and area-based socioeconomic status and food environment variables. The mediating effect of obesity on the geographic distribution of diabetes was also examined. RESULTS: In this population of insured adults, diabetes was concentrated in south and southeast King County, with smoothed diabetes prevalence ranging from 6.9% to 21.2%. In spatial regression models, home value and college education were more strongly associated with diabetes than was household income. For each 50% increase in median home value, diabetes prevalence was 1.2 percentage points lower. The Modified Retail Food Environment Index was not related to diabetes at the census-tract level. The observed associations between area-based socioeconomic status and diabetes were largely mediated by obesity (home value, 58%; education, 47%). CONCLUSION: The observed geographic disparities in diabetes among insured adults by census tract point to the importance of area socioeconomic status. Small-area studies can help health professionals design community-based programs for diabetes prevention and control.

Environments perceived as obesogenic have lower residential property values.

BACKGROUND: Studies have tried to link obesity rates and physical activity with multiple aspects of the built environment. PURPOSE: To determine the relation between residential property values and multiple perceived (self-reported) measures of the obesogenic environment. METHODS: The Seattle Obesity Study (SOS) used a telephone survey of a representative, geographically distributed sample of 2,001 King County adults, collected in 2008-2009 and analyzed in 2012-2013. Home addresses were geocoded. Residential property values at the tax parcel level were obtained from the King County tax assessor. Mean residential property values within a 10-minute walk (833-m buffer) were calculated for each respondent. Data on multiple perceived measures of the obesogenic environment were collected by self-report. Correlations and multivariable linear regression analyses, stratified by residential density, were used to examine the associations among perceived environmental measures, property values, and BMI. RESULTS: Perceived measures of the environment such as crime, heavy traffic, and proximity to bars, liquor stores, and fast food were all associated with lower property values. By contrast, living in neighborhoods that were perceived as safe, quiet, clean, and attractive was associated with higher property values. Higher property values were associated, in turn, with lower BMIs among women. The observed associations between perceived environment measures and BMI were largely attenuated after accounting for residential property values. CONCLUSIONS: Environments perceived as obesogenic are associated with lower property values. Studies in additional locations need to explore to what extent other perceived environment measures can be reflected in residential property values.

Walking objectively measured: classifying accelerometer data with GPS and travel diaries.

PURPOSE: This study developed and tested an algorithm to classify accelerometer data as walking or nonwalking using either GPS or travel diary data within a large sample of adults under free-living conditions. METHODS: Participants wore an accelerometer and a GPS unit and concurrently completed a travel diary for seven consecutive days. Physical activity (PA) bouts were identified using accelerometry count sequences. PA bouts were then classified as walking or nonwalking based on a decision-tree algorithm consisting of seven classification scenarios. Algorithm reliability was examined relative to two independent analysts' classification of a 100-bout verification sample. The algorithm was then applied to the entire set of PA bouts. RESULTS: The 706 participants' (mean age = 51 yr, 62% female, 80% non-Hispanic white, 70% college graduate or higher) yielded 4702 person-days of data and had a total of 13,971 PA bouts. The algorithm showed a mean agreement of 95% with the independent analysts. It classified PA into 8170 walking bouts (58.5 %) and 5337 nonwalking bouts (38.2%); 464 bouts (3.3%) were not classified for lack of GPS and diary data. Nearly 70% of the walking bouts and 68% of the nonwalking bouts were classified using only the objective accelerometer and GPS data. Travel diary data helped classify 30% of all bouts with no GPS data. The mean ± SD duration of PA bouts classified as walking was 15.2 ± 12.9 min. On average, participants had 1.7 walking bouts and 25.4 total walking minutes per day. CONCLUSIONS: GPS and travel diary information can be helpful in classifying most accelerometer-derived PA bouts into walking or nonwalking behavior.

How to identify food deserts: measuring physical and economic access to supermarkets in King County, Washington.

OBJECTIVES: We explored new ways to identify food deserts. METHODS: We estimated physical and economic access to supermarkets for 5 low-income groups in Seattle-King County, Washington. We used geographic information system data to measure physical access: service areas around each supermarket were delineated by ability to walk, bicycle, ride transit, or drive within 10 minutes. We assessed economic access by stratifying supermarkets into low, medium, and high cost. Combining income and access criteria generated multiple ways to estimate food deserts. RESULTS: The 5 low-income group definitions yielded total vulnerable populations ranging from 4% to 33% of the county's population. Almost all of the vulnerable populations lived within a 10-minute drive or bus ride of a low- or medium-cost supermarket. Yet at most 34% of the vulnerable populations could walk to any supermarket, and as few as 3% could walk to a low-cost supermarket. CONCLUSIONS: The criteria used to define low-income status and access to supermarkets greatly affect estimates of populations living in food deserts. Measures of access to food must include travel duration and mode and supermarket food costs.

Obesity and supermarket access: proximity or price?

OBJECTIVES: We examined whether physical proximity to supermarkets or supermarket price was more strongly associated with obesity risk. METHODS: The Seattle Obesity Study (SOS) collected and geocoded data on home addresses and food shopping destinations for a representative sample of adult residents of King County, Washington. Supermarkets were stratified into 3 price levels based on average cost of the market basket. Sociodemographic and health data were obtained from a telephone survey. Modified Poisson regression was used to test the associations between obesity and supermarket variables. RESULTS: Only 1 in 7 respondents reported shopping at the nearest supermarket. The risk of obesity was not associated with street network distances between home and the nearest supermarket or the supermarket that SOS participants reported as their primary food source. The type of supermarket, by price, was found to be inversely and significantly associated with obesity rates, even after adjusting for individual-level sociodemographic and lifestyle variables, and proximity measures (adjusted relative risk=0.34; 95% confidence interval=0.19, 0.63) CONCLUSIONS: Improving physical access to supermarkets may be one strategy to deal with the obesity epidemic; improving economic access to healthy foods is another.

Residential property values are associated with obesity among women in King County, WA, USA.

Studies of social determinants of weight and health in the US have typically relied on self-reported education and incomes as the two primary measures of socioeconomic status (SES). The assessed value of one's home, an important component of wealth, may be a better measure of the underlying SES construct and a better predictor of obesity. The Seattle Obesity Study (SOS), conducted in 2008-9, was a cross-sectional random digit dial telephone survey of 2001 adults in King County, Washington State, US. Participants' addresses were geocoded and residential property values for each tax parcel were obtained from the county tax assessor's database. Prevalence ratios of obesity by property values, education, and household income were estimated separately for women and men, after adjusting for age, race/ethnicity, household size, employment status and home ownership. Among women, the inverse association between property values and obesity was very strong and independent of other SES factors. Women in the bottom quartile of property values were 3.4 times more likely to be obese than women in the top quartile. No association between property values and obesity was observed for men. The present data strengthen the evidence for a social gradient in obesity among women. Property values may represent a novel and objective measure of SES at the individual level in the US. Measures based on tax assessment data will provide a valuable resource for future health studies.

Arterial roads and area socioeconomic status are predictors of fast food restaurant density in King County, WA.

BACKGROUND: Fast food restaurants reportedly target specific populations by locating in lower-income and in minority neighborhoods. Physical proximity to fast food restaurants has been associated with higher obesity rates. OBJECTIVE: To examine possible associations, at the census tract level, between area demographics, arterial road density, and fast food restaurant density in King County, WA, USA. METHODS: Data on median household incomes, property values, and race/ethnicity were obtained from King County and from US Census data. Fast food restaurant addresses were obtained from Public Health-Seattle & King County and were geocoded. Fast food density was expressed per tract unit area and per capita. Arterial road density was a measure of vehicular and pedestrian access. Multivariate logistic regression models containing both socioeconomic status and road density were used in data analyses. RESULTS: Over one half (53.1%) of King County census tracts had at least one fast food restaurant. Mean network distance from dwelling units to a fast food restaurant countywide was 1.40 km, and 1.07 km for census tracts containing at least one fast food restaurant. Fast food restaurant density was significantly associated in regression models with low median household income (p < 0.001) and high arterial road density (p < 0.001) but not with percent of residents who were nonwhite. CONCLUSION: No significant association was observed between census tract minority status and fast food density in King County. Although restaurant density was linked to low household incomes, that effect was attenuated by arterial road density. Fast food restaurants in King County are more likely to be located in lower income neighborhoods and higher traffic areas.

Using built environment characteristics to predict walking for exercise.

BACKGROUND: Environments conducive to walking may help people avoid sedentary lifestyles and associated diseases. Recent studies developed walkability models combining several built environment characteristics to optimally predict walking. Developing and testing such models with the same data could lead to overestimating one's ability to predict walking in an independent sample of the population. More accurate estimates of model fit can be obtained by splitting a single study population into training and validation sets (holdout approach) or through developing and evaluating models in different populations. We used these two approaches to test whether built environment characteristics near the home predict walking for exercise. Study participants lived in western Washington State and were adult members of a health maintenance organization. The physical activity data used in this study were collected by telephone interview and were selected for their relevance to cardiovascular disease. In order to limit confounding by prior health conditions, the sample was restricted to participants in good self-reported health and without a documented history of cardiovascular disease. RESULTS: For 1,608 participants meeting the inclusion criteria, the mean age was 64 years, 90 percent were white, 37 percent had a college degree, and 62 percent of participants reported that they walked for exercise. Single built environment characteristics, such as residential density or connectivity, did not significantly predict walking for exercise. Regression models using multiple built environment characteristics to predict walking were not successful at predicting walking for exercise in an independent population sample. In the validation set, none of the logistic models had a C-statistic confidence interval excluding the null value of 0.5, and none of the linear models explained more than one percent of the variance in time spent walking for exercise. We did not detect significant differences in walking for exercise among census areas or postal codes, which were used as proxies for neighborhoods. CONCLUSION: None of the built environment characteristics significantly predicted walking for exercise, nor did combinations of these characteristics predict walking for exercise when tested using a holdout approach. These results reflect a lack of neighborhood-level variation in walking for exercise for the population studied.