Abstract
Background
Patients are often advised to initiate a physical activity program by walking for transportation or leisure. This study explored whether neighborhood factors beyond the individual might affect compliance.
Objective
We examined the associations between total walking and neighborhood factors in a multi-ethnic population-based sample in California and the roles race/ethnicity plays in these associations.
Design
Cross-sectional study
Participants
Individual-level data were obtained from the 2003 California Health Interview Survey. Participants’ census tracts were linked to Census 2000 data to capture neighborhood SES.
Measurements and Main Results
The dependent variable was self-reported walking at recommended levels. Neighborhood SES was measured by a scale of 4 Census-based variables (alpha = 0.83). Social cohesion was measured by a scale tapping the extent of perceived social connectedness, trust, and solidarity among neighbors (alpha = 0.70). Neighborhood access to a park, playground, or open space was measured by a single item. Safety was measured by a scale of three items (alpha = 0.66). We performed a series of multiple logit models with robust variance estimates while taking complex survey design into account. Neighborhood social cohesion (odds ratio [OR] = 1.09, 95% CI = 1.04, 1.14) and access to a park, playground, or open space (OR = 1.26, 95% CI = 1.16, 1.36) were significant environmental correlates of walking at recommended levels, independent of individual socio-demographics. Subgroup analysis showed that neighborhood effects were different by race/ethnicity.
Conclusions
Neighborhood physical and social environmental factors are significantly associated with walking at recommended levels. Being aware of the ways that the environment could affect a patient’s compliance with PA recommendations may help physicians tailor recommendations to circumstances.
KEY WORDS: physical activity, walking, neighborhood social cohesion, trust, neighborhood safety, neighborhood SES
INTRODUCTION
Regular physical activity (PA) reduces the risk of many chronic diseases (e.g., cardiovascular conditions, diabetes, and cancer), promotes mental well-being, and improves overall quality of life.1 Increasing PA is a national health priority in the United States.2 However, only about 25% of Americans engage in the recommended amount of PA.3
Few interventions have been effective at increasing PA level long term.4,5 Patients and the public are often advised to initiate a PA program by walking because it is perceived to be relatively accessible.6,7 The Surgeon General’s Call to Action to Prevent and Decrease Overweight and Obesity endorsed walking as the “first step” in increasing PA: “You don’t need special skills or training to be physically active. Walking is a great way to be active”.8 The American Heart Association’s Task Force on Risk Reduction similarly recommended that an exercise program starts with walking: “Walking is the recommended mode of activity unless the individual can attend supervised classes where other activities can be provided.”9 In population-based health surveys, walking is the most commonly reported PA.10
Prior research has shown that most people walk in their neighborhood, and that perceived neighborhood “walkability” is significantly associated with a higher level of PA.6,11–13 Evidence also suggests that neighborhood environment influences individual PA and may thus hold the promise of facilitating long-term increase in PA.14–19 However, the literature is not consistent on which aspects of the neighborhood environment affect PA. Some studies report that neighborhood sociodemographic factors have little independent effect on PA after controlling for individual characteristics.20,21 Evidence for contextual effects of the built environment, defined by the Centers for Disease Control and Prevention as human-formed, developed, or structured areas,22 is stronger than evidence for the effects of the neighborhood social environment, which has been studied less often.23,24
Few studies have simultaneously assessed both physical and social dimensions of neighborhood environment and their associations with walking.16 It is also unclear if the neighborhood-walking associations vary by racial/ethnic group25 and if neighborhood factors contribute to racial/ethnic disparities in PA.26 There is some evidence that the individual and environmental factors associated with PA do differ across racial/ethnic groups of U.S. women.27 In this study, we examine the association between neighborhood and walking in a large population-based, multi-ethnic sample. We also examine whether these factors account for racial/ethnic differences in walking, and if these factors operate differently across racial/ethnic groups. Our data are derived from California, a state with significant racial/ethnic diversity, and also one where season and weather are relatively less likely to be key determinants of outdoor activity.
METHODS
Data
We used cross-sectional data from the 2003 California Health Interview Survey (CHIS). CHIS is a population-based telephone survey of civilian households, selected through random digit dialing. CHIS is designed to provide population-based estimates for California’s overall population and its major racial/ethnic groups. For the CHIS adult sample, the adult interview response rate was 60%,28 comparable to telephone surveys carried out by the National Center for Health Statistics. The final CHIS 2003 estimates are consistent with the 2003 California Department of Finance Population Projections.28 The sample for this analysis was restricted to adults, 18 years and older.
This study has been approved by the Institutional Review Boards (IRB)s at the University of Utah and the University of Chicago.
Dependent Variable
CHIS asked respondents about the frequency and total duration of walking for transport and leisure in the past 1 week. Based on national PA recommendations,3 we constructed the dependent variable for this study as a dichotomous variable indicating “walking at recommended levels”. It was defined as 5 or more sessions of walking (for transportation or leisure) in the previous week totaling at least 150 minutes.
Independent Variables
Neighborhood variables The four main independent variables were neighborhood socioeconomic status (SES), social cohesion, physical environment, and safety. Neighborhood-level SES was constructed from Census data including concentrated affluence, concentrated poverty, percent college educated residents, and percent home ownership; coefficient of alpha was 0.83. Higher scores indicated higher neighborhood SES. The remaining three neighborhood variables reflect respondents’ perceptions of their neighborhood. A scale of social cohesion was constructed from 5 items measuring the respondent’s level of agreement (on a 4-point scale) with the following statements: 1) “People in my neighborhood are willing to help each other”; 2) “People in this neighborhood generally do not get along with each other”; 3) “People in this neighborhood can be trusted”; 4) “People in this neighborhood do not share the same values”; and 5) “Most people in this neighborhood know each other.” The coefficient of alpha was 0.70, with higher scores indicating higher levels of social cohesion in the neighborhood. Physical environment was tapped by a single item asking whether the neighborhood had a park, playground, or open space within walking distance of home. A scale of neighborhood safety was constructed from 3 items measuring the respondents’ level of agreement (on a 4-point scale) with the following statements: 1) “Many people in this neighborhood are afraid to go out at night”; 2) “The park or playground closest to where I live is safe during the day”; 3) “The park or playground closest to where I live is safe at night.” The coefficient of alpha was 0.66; with higher scores indicating higher levels of perceived safety. For neighborhoods that have no parks or playgrounds nearby, this scale is missing. Missing values were imputed using the five social cohesion items and an item indicating whether the respondent’s house had ever been broken into (this item did not fit well with the 3 items used to construct the safety scale). The imputation did not change the results on neighborhood safety. Principal component factor analysis with orthogonal rotation was used to construct measures of neighborhood SES, social cohesion, and safety.
Sociodemographic variables Demographic variables included race/ethnicity, marital status (married vs unmarried), sex, age (18–29, 30–39, 40–49, 50–64, 65+), SES, and immigration status.
Individuals were classified by self-report as non-Hispanic White, Hispanic, Black/African American, Asian, or Other. The Other category includes American Indians, Native Hawaiians, Pacific Islanders, those who indicated “other race” or multiple races (although the tables include the “Other” race group, they will not be discussed in detail because of heterogeneity).
Individual-level SES was measured by poverty income ratio (PIR), education, and employment status (employed vs unemployed). PIR is a ratio where the numerator is household income and the denominator is the appropriate federal poverty level (FPL) given the family’s size and composition. Poverty thresholds are revised each year by the Census Bureau. Thus, a PIR less than 100% indicates that the household is living below the poverty threshold. PIR was categorized as: “0–99% of FPL,” “100–199% FPL,” “200–299% FPL” and “300% FPL and above.” Education was categorized as: “less than high school,” “high school graduate,” “some college,” and “college graduate.” Education and PIR were each entered as single ordinal variables in the regression models.
Percent of life years in the United States was employed as an indicator of immigrant status and acculturation.
Because social and environmental factors may affect body mass index (BMI), we also adjusted for BMI (weight[kg]/height[m2]) in the analysis. BMI was categorized as underweight (BMI <18.5), normal weight (BMI = 18.5–24.99), overweight (BMI = 25–24.99), and obese (BMI ≥30).
Statistical Methods
All estimates and analyses were weighted using replicate weights, provided by CHIS, to adjust for non-response and the complex survey design. We performed a series of logistic regression models with robust standard errors, where the outcome was walking at recommended levels. Model 1 included all individual-level variables except percent of life in the US and served as the baseline model. In model 2, we added percent of life in the US to test the idea that acculturation mediates the association between race/ethnicity and PA, especially for Asians and Hispanics. Then, we added each of the 4 neighborhood variables, one by one, to examine the effects of different aspects of neighborhood environment and explore what specific dimensions of neighborhood environment contributes to the race/ethnicity effects on PA. Next, we tested a full model including all 4 neighborhood variables and individual controls to see if the observed neighborhood effects were independent of each other. Lastly, we tested interaction effects between the key neighborhood factors and race/ethnicity and refit the full model for each group. For the Asian group, we further tested whether using separate Asian subgroups (i.e., Chinese, Japanese, Korean, Filipino, South Asian, Vietnamese, and Other Asian) revealed subgroup differences in the neighborhood effects (data not shown). Because there were little significant interactions between Asian subgroups, we used the aggregate race category in the analyses presented. We did not employ random effects models in this research because 3 of our 4 neighborhood variables were based on individual survey responses and therefore were not group-level variables. All neighborhood variables were standardized when entered in the models.
RESULTS
The sample included 41,545 adult respondents. Table 1 presents individual and neighborhood characteristics of the CHIS 2003 sample by race/ethnicity. Twenty-two percent of all subjects reported walking at recommended levels. This prevalence is close to the 25% national estimate.3 In this sample, Blacks had the lowest rate of walking at recommended levels, but racial/ethnic differences were small.
Table 1.
Sample Statistics* (Data from the 2003 California Health Interview Survey)
| Total | White | Black | Hispanic | Asian | Other | |
|---|---|---|---|---|---|---|
| Sample size† | 41,545 | 26,139 | 2,653 | 7,083 | 3,856 | 1,814 |
| Total walking for at least 150 minutes 5 times per week (yes) | 22% | 21% | 19% | 23% | 22% | 24% |
| BMI height/weight | ||||||
| Underweight (0, 18.49) | 2% | 2% | 1% | 1% | 6% | 1% |
| Normal (18.5, 24.99) | 42% | 44% | 33% | 33% | 62% | 38% |
| Overweight (25.0, 29.99) | 35% | 35% | 35% | 39% | 26% | 37% |
| Obese (30.0+) | 20% | 18% | 30% | 27% | 7% | 24% |
| Age (mean) | 44 | 48 | 44 | 37 | 43 | 42 |
| Male | 49% | 49% | 46% | 51% | 47% | 51% |
| Poverty income ratio % | ||||||
| 0–99% FPL Federal Poverty Level | 15% | 6% | 17% | 33% | 15% | 16% |
| 100–99% FPL | 19% | 12% | 20% | 32% | 17% | 23% |
| 200–299% FPL | 14% | 14% | 16% | 14% | 13% | 16% |
| 300% FPL and above | 52% | 69% | 47% | 22% | 55% | 46% |
| Education | ||||||
| Below high school | 20% | 8% | 12% | 51% | 12% | 19% |
| High school diploma | 24% | 23% | 30% | 24% | 19% | 30% |
| College educated, no degree | 25% | 29% | 34% | 16% | 21% | 30% |
| College degree or above | 31% | 40% | 24% | 8% | 49% | 21% |
| Married | 55% | 58% | 37% | 51% | 62% | 47% |
| Employed | 64% | 64% | 61% | 65% | 63% | 64% |
| Percent of life in the US (mean) | 81 | 96 | 96 | 61 | 52 | 87 |
| Components of the Neighborhood SES Scale mean | ||||||
| Percent families with annual income ≥ $75,000 | 29% | 34% | 22% | 20% | 35% | 27% |
| Percent individuals in poverty | 13% | 10% | 18% | 19% | 11% | 14% |
| Percent college educated residents | 38% | 44% | 32% | 25% | 41% | 36% |
| Percent home ownership | 57% | 62% | 50% | 49% | 57% | 57% |
| Components of the Neighborhood Social Cohesion Scale | ||||||
| Neighborly helpfulness (yes)‡ | 86% | 90% | 80% | 80% | 87% | 84% |
| People do not get along with each other (yes)§ | 16% | 9% | 14% | 30% | 17% | 17% |
| Neighborly trust (yes)|| | 83% | 89% | 76% | 74% | 84% | 79% |
| People do not share the same values (yes) | 41% | 34% | 45% | 51% | 51% | 44% |
| People know each other (yes)# | 68% | 67% | 68% | 73% | 60% | 68% |
| Neighborhood Access to Park/Playground/Open Space (yes) | 78% | 80% | 78% | 77% | 75% | 79% |
| Components of the Neighborhood Safety Scale | ||||||
| People are afraid to go out at night (yes) ** | 21% | 14% | 23% | 34% | 24% | 22% |
| Neighborhood park or playground safe during the day (yes) †† | 96% | 98% | 94% | 93% | 96% | 96% |
| Neighborhood park of playground safe at night (yes) †† | 65% | 72% | 58% | 52% | 66% | 61% |
*For categorical variables, the percent for each item (%) is presented. For age and percent of life in the US, means are presented.
†Except for the two neighborhood park or playground safety variables (see note †† below), all sample statistics were calculated based on sample sizes shown in this row.
‡ The item was based on the respondent’s level of agreement with the statement “People in my neighborhood are willing to help each other.” The item presented here was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
§The item was based on the respondent’s level of agreement with the statement “People in this neighborhood generally do not get along with each other.” The item presented here was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
||The item was based on the respondent’s level of agreement with the statement “People in this neighborhood can be trusted.” The item presented here was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
¶The item was based on the respondent’s level of agreement with the statement “People in this neighborhood do not share the same values.” The item was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
#The item was based on the respondent’s level of agreement with the statement “Most people in this neighborhood know each other.” The item presented here was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
**The item was based on the respondent’s level of agreement with the statement “Many people in this neighborhood are afraid to go out at night.” The item was dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no).
††The two items were based on the respondent’s level of agreement with the statement "The park or playground closest to where I live is safe during the day or at night." The items were dichotomized into two levels: strongly agree/agree (yes) versus strongly disagree/disagree (no). These two variables have significant yet legitimate missing values (22%).
Fifty-five percent of all participants were overweight or obese, also comparable to national estimates.2 Blacks and Hispanics had the highest proportions of overweight and obese, whereas Asians were much less likely to be overweight or obese and more likely to be underweight.
Sociodemographic variables varied substantially by race/ethnicity (Table 1). Whites and Asians were more likely to be married and had higher SES than other groups. Asian respondents had spent an average of 52% of life in the United States, whereas Whites and Blacks had spent almost all of their life in the US.
Whites and Asians had comparable SES. Neighborhood SES was much lower for other groups. Levels of perceived social cohesion and perceived safety were the highest among Whites. Perceived safety was the lowest for Hispanics, followed by Blacks. The likelihood of having a nearby park, playground, or open space was also the highest for Whites, although group differences were small.
Table 2 presents our main analytical results. After adjusting for individual factors, compared to Whites, Blacks were significantly less likely to report walking at recommended levels; Hispanics had significantly higher odds of reporting walking at recommended levels; and Asians were not much different. After adding percent of life in the US to the baseline model, Asians became significantly less likely to report walking at recommended levels, and the Hispanic ethnicity effect became insignificant (Table 2, Model 2).
Table 2.
Seven Logistic Regression Models Predicting Walking at Recommended Levels* (Data from the 2003 California Health Interview Survey)
| 1. Baseline Model | 2. Acculturation | 3. Neighborhood SES | 4. Social Cohesion | 5. Open Space | 6. Safety | 7. Full Model | |
|---|---|---|---|---|---|---|---|
| White (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Black | 0.86* [0.75, 0.98] | 0.87* [0.77, 0.99] | 0.87* [0.77, 0.99] | 0.89 [0.78, 1.01] | 0.87* [0.77, 0.99] | 0.88 [0.77, 1.01] | 0.88* [0.77, 1.00] |
| Hispanic | 1.19*** [1.08, 1.31] | 0.99 [0.89, 1.10] | 0.99 [0.89, 1.10] | 1.00 [0.90, 1.12] | 0.99 [0.89, 1.10] | 1.00 [0.90, 1.11] | 0.99 [0.89, 1.10] |
| Asian | 0.98 [0.87, 1.11] | 0.75*** [0.66, 0.86] | 0.75*** [0.66, 0.85] | 0.77*** [0.67, 0.87] | 0.76*** [0.67, 0.87] | 0.76*** [0.66, 0.86] | 0.77*** [0.68, 0.88] |
| Other | 1.22* [1.03, 1.43] | 1.17* [1.00, 1.37] | 1.17* [1.00, 1.37] | 1.18* [1.01, 1.39] | 1.17* [1.00, 1.37] | 1.18* [1.00, 1.38] | 1.18* [1.00, 1.38] |
| Normal weight (Ref) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 |
| Underweight | 0.91 [0.73, 1.14] | 0.90 [0.72, 1.12] | 0.90 [0.72, 1.12] | 0.90 [0.72, 1.12] | 0.90 [0.73, 1.12] | 0.90 [0.72, 1.12] | 0.91 [0.73, 1.13] |
| Overweight | 0.80*** [0.74, 0.87] | 0.81*** [0.75, 0.86] | 0.81*** [0.75, 0.86] | 0.81*** [0.75, 0.88] | 0.81*** [0.75, 0.88] | 0.81*** [0.75, 0.88] | 0.81*** [0.75, 0.88] |
| Obese | 0.68*** [0.62, 0.75] | 0.70*** [0.63, 0.77] | 0.70*** [0.63, 0.77] | 0.70*** [0.64, 0.78] | 0.70*** [0.63, 0.77] | 0.70*** [0.63, 0.77] | 0.70*** [0.63, 0.77] |
| Age† | 1.06*** [1.04, 1.09] | 1.07*** [1.04, 1.10] | 1.07*** [1.04, 1.10] | 1.06*** [1.03, 1.09] | 1.07*** [1.05, 1.10] | 1.07*** [1.04, 1.10] | 1.07*** [1.04, 1.10] |
| Male gender | 1.04 [0.97, 1.11] | 1.03 [0.96, 1.10] | 1.03 [0.96, 1.10] | 1.04 [0.97, 1.12] | 1.02 [0.96, 1.10] | 1.03 [0.96, 1.10] | 1.03 [0.96, 1.11] |
| Married | 0.94 [0.87, 1.01] | 0.89** [0.83, 0.96] | 0.89** [0.83, 0.96] | 0.87*** [0.82, 0.94] | 0.89*** [0.83, 0.95] | 0.89*** [0.83, 0.95] | 0.88*** [0.82, 0.94] |
| Poverty Income Ratio %‡ | 0.92*** [0.88, 0.97] | 0.96 [0.92, 1.01] | 0.96 [0.92, 1.01] | 0.95* [0.91, 0.99] | 0.96 [0.92, 1.00] | 0.96 [0.92, 1.00] | 0.95* [0.91, 1.00] |
| Education § | 1.07***[1.03, 1.11] | 1.08***[1.04, 1.12] | 1.08***[1.04, 1.12] | 1.07***[1.03, 1.11] | 1.07***[1.03, 1.11] | 1.07***[1.03, 1.11] | 1.07***[1.03, 1.11] |
| Employed | 0.82***[0.76, 0.89] | 0.81***[0.75, 0.88] | 0.81***[0.75, 0.88] | 0.81***[0.75, 0.88] | 0.81***[0.75, 0.88] | 0.81***[0.75, 0.88] | 0.81***[0.74, 0.87] |
| Percent of life In the US || | 0.90***[0.92, 0.95] | 0.90***[0.92, 0.95] | 0.90***[0.92, 0.95] | 0.90***[0.92, 0.95] | 0.90***[0.92, 0.95] | 0.90***[0.92, 0.95] | |
| Neighborhood SES ¶ | 1.00[0.96, 1.04] | 0.98[0.94, 1.01] | |||||
| Neighborhood social cohesion # | 1.10***[1.06, 1.14] | 1.09***[1.04, 1.14] | |||||
| Neighborhood access to open space** | 1.27***[1.17, 1.37] | 1.26***[1.16, 1.36] | |||||
| Neighborhood safety † | 1.05**[1.01, 1.08] | 1.01[0.97, 1.05] |
*N = 41,545 individuals. 95% confidence intervals of the odds ratios are presented in the parentheses. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001
†Age is treated as a continuous variable in the models. It has five levels: 18–29 years, 30–39 years, 40–49 years, 50–64 years, 65+ years.
‡Poverty Income Ratio (%) is treated as a continuous variable in the models. It has four levels: “0–99% of FPL,” “100–199% FPL,” “200–299% FPL,” and “300% FPL and above.”
§Education is treated as a continuous variable in the models. It has four levels: below high school, high school diploma, college educated (no degree), and college degree or above.
||Odds ratios presented in this row correspond to 20% increase in percent of life in the US.
¶Neighborhood SES is measured by a factor score of concentrated affluence, concentrated poverty, percent college educated residents, and percent home ownership. Higher scores indicate higher stock of socioeconomic resources in the neighborhood.
#Neighborhood social cohesion is measured by reported perceptions of the extent of social connectedness, trust, and solidarity among neighbors. Higher scores indicate higher levels of social cohesion in the neighborhood.
**Neighborhood access to open space, as a proxy for neighborhood physical environment, is tapped by a single item asking whether the neighborhood has a park, playground, or open space within walking distance of home (yes/no).
††Neighborhood safety is measured by a factor score based on perceived safety of the neighborhood park or playground at day time, at night time, plus an item asking respondents’ agreement to the statement "Most people in this neighborhood are afraid to go out at night.” Higher scores indicate higher levels of perceived neighborhood safety.
Neighborhood SES was not a significant correlate of walking at recommended levels. Whereas safety was a positive correlate when none of the other neighborhood variables were examined (Model 6), it was no longer significant in the full model (Model 7). Neighborhood social cohesion (OR = 1.09; 95% CI = 1.04, 1.14) and access to a park, playground, or open space (OR = 1.26; 95% CI = 1.16, 1.36) were both significantly associated with walking at recommended levels, after adjusting for individual sociodemographic factors and neighborhood SES and safety. Neighborhood factors did not seem to mediate racial/ethnic differences in walking at recommended levels. The magnitude of the race/ethnicity coefficients did not change when we added neighborhood factors to the regression models (Table 2, Models 3–7).
Table 3 presents the full model (Table 2, Model 7) stratified by 5 major racial/ethnic groups in the United States. All models were adjusted for age, gender, race/ethnicity, BMI, PIR, education, marital status, employment status, and percent of life in the United States. We focus on the point estimates, as the confidence intervals and significance also reflect differences in the sample sizes. Neighborhood SES was negatively correlated with walking at recommended levels among Blacks but had no effect on Whites, Hispanics or Asians. Social cohesion was positively associated with walking at recommended levels among Whites and Hispanics. Access to a park, playground, or open space was positively correlated with walking at recommended levels among all groups, except for Asians. Neighborhood safety was not significantly associated with walking at recommended levels in any subgroup analysis. Formal interaction tests further showed that neighborhood environment seemed to have stronger impacts on Whites than for minorities.
Table 3.
Ethnic-Group Specific Logistic Regression Models Predicting Walking at Recommended Levels* (Data from the 2003 California Health Interview Survey)
| White | Black | Hispanic | Asian | Other | |
|---|---|---|---|---|---|
| Neighborhood SES† | 0.99[0.95, 1.03] | 0.86*[0.75, 0.98] | 0.96[0.88, 1.05] | 1.04[0.93, 1.17] | 0.81**[0.68, 0.95] |
| Neighborhood Social Cohesion ‡ | 1.06***[1.01, 1.12] | 1.10[0.92, 1.33] | 1.14*[1.02, 1.27] | 1.11[0.97, 1.27] | 1.07[0.90, 1.27] |
| Neighborhood Access to Open Space§ | 1.29***[1.15, 1.45] | 1.64***[1.16, 2.32] | 1.21*[1.02, 1.44] | 1.03[0.81, 1.30] | 1.14[0.80, 1.62] |
| Neighborhood Safety|| | 1.02[0.97, 1.08] | 1.14[0.91, 1.42] | 0.98[0.90, 1.06] | 0.94[0.82, 1.08] | 0.96[0.81, 1.15] |
| Sample size | 26,139 | 2,653 | 7,083 | 3,856 | 1,814 |
*95% confidence intervals of the odds ratios are presented in the parentheses. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. The full model in Table 2 (Model 7) is refit for each of the 5 racial/ethnic groups. All models are adjusted for age, gender, race/ethnicity, BMI, PIR, education, marital status, employment status, and percent of life in the US. 95% confidence intervals of the odds ratios are presented in the parentheses.
†Neighborhood SES is measured by a factor score of concentrated affluence, concentrated poverty, percent college educated residents, and percent home ownership. Higher scores indicate higher stock of socioeconomic resources in the neighborhood.
‡Neighborhood social cohesion is measured by reported perceptions of the extent of social connectedness, trust, and solidarity among neighbors. Higher scores indicate higher levels of social cohesion in the neighborhood.
§Neighborhood access to open space, as a proxy for neighborhood physical environment, is tapped by a single item asking whether the neighborhood has a park, playground, or open space within walking distance of home (yes/no).
||Neighborhood safety is measured by a factor score based on perceived safety of the neighborhood park or playground at day time, at night time, plus an item asking respondents’ agreement to the statement "Most people in this neighborhood are afraid to go out at night.” Higher scores indicate higher levels of perceived neighborhood safety.
DISCUSSION
Using a multi-ethnic, population-based sample from California, we found that a neighborhood’s social and built environment was significantly associated with walking at recommended levels, independent of individual sociodemographics. Surprisingly, we did not find that neighborhood SES and safety were important determinants in the full sample. Among the 4 neighborhood factors examined, both social cohesion and access to a park, playground, or open space were positively associated with walking at recommended levels. The promoting effect of access to facilities and destinations on PA has been confirmed elsewhere,29–31 supporting the importance of the built environment for PA.32,33 In this study, we found that the social environment, which was measured as perceived social cohesion, was also important.
We also found that the pattern of neighborhood effects was not consistent across racial/ethnic groups. For Blacks, neighborhood SES appeared to be an environmental barrier, consistent with an Illinois study finding that neighborhood poverty was associated with more walking.17 Additional analysis showed that increasing neighborhood SES was positively linked to more leisure walking (data not shown), but the inclusion of transportation walking offset this effect. These discrepancies in the size and direction of the association between neighborhood SES and walking suggest that both transportation and leisure walking should be examined to better understand environmental influences on PA. For Asians, neighborhood effects were generally weaker. To date, most active living research has not included Asians, hence data about the factors influencing PA among Asian Americans are limited.34,35 More research is needed to explore determinants of PA in this group. Overall, the racial/ethnic variation in the neighborhood–PA link suggests that community intervention programs may need to be individually tailored to meet the needs of each particular group.
Although the associations between walking at recommended levels and neighborhood factors varied by race/ethnicity, neighborhood factors did not explain the lower rates of walking at recommended levels among Blacks and Asians compared to non-Hispanic Whites. Cultural factors should be examined to better understand disparities in PA across ethno-cultural groups.
The role of neighborhood safety is noteworthy. We found perceived safety was not an important correlate in the presence of social cohesion and access to open space. Prior evidence on the effect of perceived neighborhood safety on PA is mixed.36–41 Different measures of safety and PA are possible sources of this inconsistency. More sophisticated measures of neighborhood safety should be used in future work to further evaluate the contribution of neighborhood safety in PA.
Many individual factors were significantly associated with walking at recommended levels. Older age and more education were positive factors, whereas marriage, higher BMI, higher income, employment, and greater percent of life in the United States were negative correlates. Some of these results are unexpected. Age has been negatively linked to leisure-time PA,42,43 and a curvilinear relationship between age and PA has been reported for California residents when PA was measured by non-leisure walking and biking.44 However, few studies have directly assessed how age is linked to total amount of walking. It is plausible that walking is a more important form of PA for older adults, both for leisure and as a means of transport.45 Hence, activity-friendly environment may be particularly important for older people.
The longer a respondent has lived in the United States, the less likely he/she is to report walking at recommended levels. Stratified analyses showed that the magnitude of this effect was fairly consistent across racial/ethnic groups (data not presented). Moreover, controlling for acculturation, compared to whites, the Hispanic advantage disappeared and the Asian disadvantage became strengthened (see Model 1 to Model 2 in Table 2). These results suggest that acculturation may be a risk factor for physical inactivity among immigrants.
There is insufficient evidence about the effectiveness of routine PA counseling in the primary care setting. However, several professional health organizations and the US Department of Health and Human Services encourage PA counseling in the primary care setting,46 and there is some evidence that a patient-centered assessment and counseling influences PA.47 In light of this, our findings may have implications for primary care physicians who hope to increase PA levels among their patient. When discussing PA with patients and developing a patient-centered exercise plan, the physician may want to enquire about neighborhood factors and how these affect the patient’s willingness or ability to walk for exercise.
This study has several limitations. First, cross-sectional data limit our ability to make causal inferences about the neighborhood environment and walking behavior. Even longitudinal observational data would not resolve the direction of causality because of the selection bias. Experimental studies are rarely used to examine neighborhood effects on PA, but they are not entirely impossible. For example, were policy experiments, such as the Gautreaux Assisted Housing Program in Chicago and the more recent Moving to Opportunity housing voucher programs in various American cities,48–50 to incorporate behavioral outcomes, then we would have experimental or quasi-experimental data to more rigorously investigate whether there is a causal link between the environment and PA. Second, the study is based in California and these results are not necessarily generalizable to other places. Third, our measure of walking was based on self-report and was subject to response bias.
This study suggests that neighborhood environment is associated with walking at recommended levels, and that the strength of these associations varies across racial/ethnic groups. Neighborhood factors do not explain the lower rates of walking at recommended levels among Asians and Blacks in California. Interventions that address both neighborhood and individual-level factors may be more fruitful in increasing PA; however, tailoring interventions to the needs of individuals and groups is likely to be more effective, and more research is needed to understand if such interventions can address racial/ethnic disparities in PA.
Acknowledgments
This work was supported by a Research Scholars Grant GPHPS 107922 from the American Cancer Society, Atlanta, GA (Lauderdale is the PI; Wen and Kandula are co-investigators) and a grant awarded to Wen (NICHD R03 HD0525370-01). The authors thank UCLA Center for Health Policy Research for providing access to the confidential data of the 2003 California Health Interview Survey.
Conflict of Interest This research was supported by a research grant from the American Cancer Society to all three authors, with Dr. Lauderdale as PI and Dr. Wen and Dr. Kandula as co-investigators. Dr. Wen also received a grant from the NICHD to partly support her work in this research.No other potential conflict of interest is involved in this research.
References
- 1.Centers for Disease Control and Prevention. Perceptions of neighborhood characteristics and leisure-time physical inactivity—Austin/Travis County, Texas, 2004. MMWR Morb Mort Wkly Rep. 2005;54(37):926–8. [PubMed]
- 2.U.S. Department of Health and Human Services. Healthy People 2010. With Understanding and Improving Health and Objectives for Improving Health. Vol 2 volumes, 2nd ed. Washington, DC: U.S. Government Printing Office; 2000.
- 3.Centers for Disease Control and Prevention. Physical activity trends—United States, 1990–1998. MMWR Morb Mort Wkly Rep. 2001;50:166–9. [PubMed]
- 4.Dishman RK, Buckworth J. Increasing physical activity: a quantitative synthesis. Med Sci Sports Exerc. 1996;28(6):706–19. [DOI] [PubMed]
- 5.Kahn EB, Ramsey LT, Brownson RC, et al. The effectiveness of interventions to increase physical activity: a systematic review. Am J Prev Med. 2002;22(4S). [DOI] [PubMed]
- 6.Eyler AA, Brownson RC, Bacak SJ, Housemann RA. The epidemiology of walking for physical activity in the United States. Med Sci Sports Exerc. 2003;35(9):1529–36. [DOI] [PubMed]
- 7.Hillsdon M, Thorogood M, Antiss T, Morris J. Randomized controlled trials of physical activity promotion in a free-living populations: a review. J Epidemiol Community Health. 1995;49:448–53. [DOI] [PMC free article] [PubMed]
- 8.The Surgeon General’s Call to Action Prevent and Decrease Overweight and Obesity. Washington, DC: U.S. Government Printing Office; 2001. [PubMed]
- 9.Fletcher GF. How to implement physical activity in primary and secondary prevention: a statement for healthcare professionals from the task force on risk reduction, American Heart Association. Circulation. 1997;96:355–8. [DOI] [PubMed]
- 10.U.S. Department of Health and Human Services, Centers of Disease Control and Prevention. Physical Activity and Health: A Report of the Surgeon General. National Center for Chronic Disease Prevention and Health Promotion; 1996:87–142.
- 11.Ainsworth BE, Wilcox S, Thompson WW, Richter DL, Henderson KA. Personal, social, and physical environmental correlates of physical activity in African-American Women in South Carolina. Am J Prev Med. 2003;25(3Si):23–9. [DOI] [PubMed]
- 12.Huston SL, Evenson KR, Bors P, Gizlice Z. Neighborhood environment, access to places for activity, and leisure-time physical activity in a diverse North Carolina Population. Am J Health Promot. 2003;18:58–69. [DOI] [PubMed]
- 13.Robinson TN, Sirard JR. Preventing childhood obesity: a solution-oriented research paradigm. Am J Prev Med. 2005;28(suppl 2):194–201. [DOI] [PubMed]
- 14.Duncan SC, Duncan TE, Strycker LA. A multilevel analysis of neighborhood context and youth alcohol and drug problems. Prevention Science 2002;3(2):125–33. [DOI] [PubMed]
- 15.Giles-Corti B, Donovan RJ. Relative influences of individual, social environmental, and physical environmental correlates of walking. Am J Public Health. 2003;93(9):1583–9. [DOI] [PMC free article] [PubMed]
- 16.McNeill LH, Kreuter MW, Subramanian SV. Social environment and physical activity: a review of concepts and evidence. Soc Sci Med. 2006;63:1011–22. [DOI] [PubMed]
- 17.Ross CE. Walking, exercising, and smoking: does neighborhood matter? Soc Sci Med. 2000;51(2):265–74. [DOI] [PubMed]
- 18.Sundquist J, Malmström M, Johansson S-E. Cardiovascular risk factors and the neighborhood environment: a multilevel analysis. Int J Epidemiol. 1999;28(5):841–5. [DOI] [PubMed]
- 19.Yen IH, Kaplan GA. Poverty area residence and changes in physical activity level: Evidence from the Alameda County study. Am J Public Health. 1998;88:1709–12. [DOI] [PMC free article] [PubMed]
- 20.Cubbin C, Hadden WC, Winkleby MA. Neighborhood context and cardiovascular disease risk factors: the contribution of material deprivation. Ethn Dis. 2001;11(4):687–700. [PubMed]
- 21.Lee RE, Cubbin C. Neighborhood context and youth cardiovascular health behaviors. Am J Public Health. 2002;92(3):428–36. [DOI] [PMC free article] [PubMed]
- 22.Centers for Disease Control and Prevention. Healthy places terminology-designing and building healthy places. Available at: http://www.cdc.gov/healthyplaces/terminology.htm.; Accessed on: August 12, 2005.
- 23.Duncan MJ, Spence JC, Mummery WK. Perceived environment and physical activity: a meta-analysis of selected environmental characteristics. Int J Behav Nutr Phys Act. 2005;2(11):1–9. [DOI] [PMC free article] [PubMed]
- 24.Ewing R. Can the physical environment determine physical activity levels? Exerc Sport Sci Rev. 2005;33(2):69–75. [DOI] [PubMed]
- 25.Eyler AA, Baker E, Cromer L, King AC, Brownson RC, Donatelle RJ. Physical activity and minority women: a qualitative study. Health Educ Behav. 1998;25:640–52. [DOI] [PubMed]
- 26.Boslaugh SE, Luke DA, Brownson RC, Naleid KS, Kreuter MW. Perceptions of neighborhood environment for physical activity: Is it “who you are” or “where you live”? J Urban Health 2004;81(4):671–81. [DOI] [PMC free article] [PubMed]
- 27.King AC, Castro C, Wilcox S, Eyler AA, Sallis JF, Brownson RC. Personal and environmental factors associated with physical inactivity among different racial-ethnic groups of U.S. middle-aged and older-aged women. Health Psychol. 2000;19(4):354–64. [DOI] [PubMed]
- 28.California Health Interview Survey. The 2003 California Health Interview Survey Response Rates. Los Angeles, CA: UCLA Center for Health Policy Research; 2003.
- 29.Bedimo-Rung AO, Mowen AJ, Cohen D. The significance of parks to physical activity and public health: a conceptual model. Am J Prev Med. 2005;28(suppl 2):159–68. [DOI] [PubMed]
- 30.Cohen DA, McKenzie TL, Sehgal A, Williamson S, Golinelli D, Lurie N. Contribution of public parks to physical activity. Am J Public Health. 2007;97(3):509–14. [DOI] [PMC free article] [PubMed]
- 31.Godbey GC, Caldwell LL, Floyd M, Payne LL. Contributions of leisure studies and recreation and park management research to the active living agenda. Amerique. 2005;28(suppl 2):150–8. [DOI] [PubMed]
- 32.Sallis JF, Bauman A, Pratt M. Environmental and policy interventions to promote physical activity. Am J Prev Med. 1998;15:379–97. [DOI] [PubMed]
- 33.Sallis JF, Linton LS, Kraft MK. The first Active Living Research conference: growth of a transdisciplinary field. Am J Prev Med. 2005;28(2S2):93–5. [DOI] [PubMed]
- 34.Centers for Disease Control and Prevention. Physical activity among Asians and Native Hawaiian or other Pacific Islanders—50 states and the District of Columbia, 2001–2003. MMWR Morb Mort Wkly Rep. 2004;53(33):756–60. [PubMed]
- 35.Kandula NR, Lauderdale DS. Leisure time, non-leisure time, and occupational physical activity in Asian Americans. Ann Epidemiol. 2005;15(4):257–65. [DOI] [PubMed]
- 36.Grzywacz JG, Marks NF. Social inequalities and exercise during adulthood: toward an ecological perspective. J Health Soc Behav. 2001;42(2):202–20. [DOI] [PubMed]
- 37.Humpel N, Owen N, Iverson D, Leslie E, Bauman A. Perceived environment attributes, residential location, and walking for particular purposes. Am J Prev Med. 2004;26(2):119–25. [DOI] [PubMed]
- 38.Motl RW, Dishman RK, Ward DS, et al. Perceived physical environment and physical activity across one year among adolescent girls: self-efficacy as a possible mediator? J Adolesc Health. 2005;37(5):403–8. [DOI] [PubMed]
- 39.Shenassa ED, Liebhaber A, Ezeamama A. Perceived safety of area of residence and exercise: a pan-European study. Am J Epidemiol. 2006;163(11):1012–7. [DOI] [PubMed]
- 40.Voorhees CC, Young DR. Personal, social, and physical environmental correlates of physical activity levels in urban Latinas. Am J Prev Med. 2003;25(3Si):61–8. [DOI] [PubMed]
- 41.Wen M, Browning CR, Cagney K. A multi-level study of neighborhood environment and its relationship to physical activity in adulthood. Urban Stud. 2007;44(13):1–18.
- 42.Degrance JM, Mouton CP, Lichtenstein MJ, Hazuda HP. Potential mediators of ethnic differences in physical activity in older Mexican Americans and European Americans: Results from the San Antonio Longitudinal Study of Aging. J Am Geriatr Soc. 2005;53:1240–7. [DOI] [PubMed]
- 43.Lindström M, Hanson BS, Wirfalt E, Ostergren PO. Socioeconomic differences in the consumption of vegetables, fruit and fruit juices. The influence of psychosocial factors. Eur J Public Health. 2001;11(1):51–9. [DOI] [PubMed]
- 44.Berrigan D, Troiano RP, McNeel T, DiSogra C, Ballard-Barbash R. Active transportation increases adherence to activity recommendations. Am J Prev Med. 2006;31(3):210–6. [DOI] [PubMed]
- 45.Berke EM, Koepsell TD, Moudon AV, Hoskins RE, Larson EB. Association of the built environment with physical activity and obesity in older persons. Am J Public Health. 2007;97(3):486–92. [DOI] [PMC free article] [PubMed]
- 46.Jacobson D, Strohecker L, Compton M, Katz D. Physical activity counseling in the adult primary care setting position statement of the American College of Preventive Medicine. Am J Prev Med. 2005;29(2):158–62. [DOI] [PubMed]
- 47.Bolognesi M, Nigg CR, Massarini M, Lippke S. Reducing obesity indicators through PACE in Italian Primary Care settings. Annals Behav Med 2006;31:179–85 [DOI] [PubMed]
- 48.Heymann J, Fishcher A. Neighborhoods, Health Research, and Its Relevance to Public Policy. In: Kawachi I, Berkman LF, eds. Neighborhoods and Health. New York: Oxford University Press; 2003:335–47.
- 49.Katx L, Kling J, Liebman J. Moving to opportunity in Boston: early results of a randomized mobility experiment. Q J Econ. 2001;116(2):604–54.
- 50.Rosenbaum JE. Changing the geography of opportunity by expanding residential choice: lessons from the Gautreaux Program. Hous Policy Debate. 1995;6:231–69.