- Click here for explanation of classification scheme.

Using data from a nationally-representative, ethnically-diverse sample of adolescents, the aims of this study were to

1. Identify meaningful patterns of socio-demographic and built features in neighborhood environments that have been identified as potentially important determinants of physical activity
2. Describe the cross-sectional associations between these neighborhood patterns and adolescent resident's physical activity and weight status.

• The Wave-I in-home survey (1994 to 1995) included 20,745 adolescent participants
• Analyses were conducted in 2005 and 2006
• Add Health is a school-based longitudinal survey of youths, Grades Seven through 12. A random sample of 80 high schools and 52 junior high feeder schools was selected.
• The Add Health sample was designed to be nationally-representative of students in Grades Seven through 12 in 1995 in the United States.

• Observations with missing covariate or outcome data were excluded
• Participants who were severely disabled or pregnant were also excluded.

Recruitment

• Add Health is a school-based longitudinal survey of youths, Grades Seven through 12. A random sample of 80 high schools and 52 junior high feeder schools was selected.
• The Add Health sample was designed to be nationally-representative of students in Grades Seven through 12 in 1995 in the United States
• The Wave-I in-home survey (1994 to 1995) included 20,745 adolescent particpants.

Design

• Weight, height, physical activity (PA) and sedentary behavior were self-reported
• Using diverse measures of the participants’ residential neighborhoods (e.g., socio-economic status, crime, road type, street connectivity, physical activitiy recreation facilities), cluster analyses identified homogeneous groups of adolescents sharing neighborhood characteristics
• Poisson regression predicted relative risk (RR) of being physically active (five or more bouts per week of moderate-to-vigorous PA) and overweight (body mass index at or above the 95th percentile, Centers for Disease Control and Prevention/National Center for Health Statistics growth curves).

Statistical Analysis

Cluster analyses were used to identify patterns of environmental characteristics and to specify homogeneous, non-overlapping clusters (or patterns) of neighborhoods sharing various meaningful characteristics. Multiple cluster analyses were conducted partitioning data into different numbers of clusters by Euclidean distances between observations that were weighted for national representation, using SAS FASTCLUS, SAS version 9 (Research Triangle Institute, Research Triangle Park NC, 2004). Representing different constructs of the neighborhood, 19 variables were used. Z-score transformations of variables were used to generate clusters, allowing for the appropriate weighing of variables with different scales.

To identify intitial cluster centers (i.e., seed values), 1,000 iterations of each cluster procedure were conducted. The initail group center for each iteration was randomly generated. The iteration with the largest overall R² value, which allowed for the evaluation of relative heterogeneity between clusters (vs. homogeneity within clusters), was identified. Clusters best fitting the data maximized this inter- to intra-variability ratio, yielding a higher R². [For the six-cluster solution series (i.e., the final cluster solution), the maximum R² value identified through this iterative process was 0.41.] Results of these numerous analyses were assessed to identify common patterns appearing across various procedures. The final presented clusters were those representing the most robust data patterns. 

To demonstrate meaningful variability between patterns and to validate these findings, neighborhood clusters were assessed as independent variables in generalized linear models predicting adolescent PA, sedentary behavior and overweight. As another tool for comparison, broad neighborhood characteristics were examined (e.g., broad urbanicity classifications of urban, suburban and rural; median household income; percentage college-educated population; percentage minority population), which have been used extensively in previous literature.

Timing of Measurements

The measurements were colllected during 1994 and 1995.

Dependent Variables

• Residential location (home street addresses of most participants were idenfitied and geocoded)
• Buffers for respondent locations (a three-km buffer was drawn around each respondent's residental location)
• Plysical activity facilities within three km [detailed characterization from the commercially-available, retrospective (1995) digitized Yellow Pages were obtained]
• Walkability within three km (neighborhood street networks that are continuous, integrated and maximize linkages between starting points and destinations)
• Road type within three km (road networks wre mapped using retrospective US Census line files)
• Census measures
• Crime (reported by the 1995 US Federal Bureau of Investigation Uniform Crime Reporting county-level data from the National Archive of Criminal Justic Data).

Independent Variables

• Physical activity and sedentary behaviors (self-reported using standard epidemilogic seven-day recall methodology)
• Weight status (self-reported height and weight to calculate body mass index).

Control Variables

Age, race or ethnicity and parent education and income were controlled for.

• Initial N: 20,745
• Attrition (final N): 20,745 (50.1% males)
• Age: 15.4±0.12 years
• Ethnicity: 68.5% white, 15.2 % black, 11.4 % Hispanic, 4.0% Asian
• Other relevant demographics: Approximately 14.7% of participants’ parents had less than a high school education, 32.5% had graduated from high school (or had a general equivalency diploma), 27.8% had some college and 25.0% had a college degree or higher
• Anthropometrics (e.g., were groups same or different on important measures): See below
• Location: United States.

Key Findings

Six robust neighborhood patterns were identified:

1. Rural working class
2. Exurban
3. Newer suburban
4. Upper-middle class, older suburban
5. Mixed-race urban
6. Low-socioeconomic-status (SES) inner-city areas.

These neighborhood patterns were distinguished by important differences in the 19 neighborhood attributes used to generate the final cluster soluction, inclusing SES, race or ethnicity, socio-environment, crime, road type, street connectivity and recreation facilities. 

Compared to adolescents living in newer suburbs, those in rural working-class [adjusted RR (ARR), 1.38; 95% confidence interval (CI), 1.13 to 1.69), exurban (ARR, 1.30; CI, 1.04 to 1.64) and mixed-race urban (ARR, 1.31; CI, 1.05 to 1.64) neighborhoods were more likely to be overweight, independent of individual SES, age and race or ethnicity. Adolescents living in older suburban areas were more likely to be physically active than residents of newer suburbs (ARR, 1.1; CI, 1.04 to 1.18). Those living in low-SES inner-city neighborhoods were more likely to be active (though not significantly so), compared to mixed-race urban residents (ARR, 1.09; CI, 1.00 to 1.18).

 

These findings demonstrate disadvantageous associations between specific rural and urban  environments and behavior, illustrating important effects of the neighborhood on health and the inherent complexity of assessing residential landscapes across the United States.Simple classical urban-suburban-rural measures mask these important complexities.

 

Government:	NIH, CDC
Not-for-profit	Robert Wood Johnson Foundation's Active Living Research Program Other non-profit:

Large study examining the association between adolescent overweight, activity and the built environment.