Model-estimated‡ multivariable-adjusted§ predicted probabilities (95% CI) of residing in low, average and high neighbourhood cluster types¶ by income status and levels (categories or ±1SD of mean) of covariates using multivariate multinomial logistic regression, CARDIA examination years 0–25
| Low connectivity/amenities (n=545 (23.2%)) | Average connectivity/amenities (409 (17.4%)) | High connectivity/amenities (1398 (59.4%)) | |
|---|---|---|---|
| Low income (first tertile) | |||
| Black | 0.28 (0.23 to 0.32) | 0.52 (0.47 to 0.58)* | 0.20 (0.16 to 0.23) |
| White | 0.25 (0.20 to 0.29) | 0.57 (0.51 to 0.62)* | 0.19 (0.14 to 0.23) |
| Male | 0.27 (0.23 to 0.30)* | 0.55 (0.50 to 0.59) | 0.19 (0.16 to 0.22) |
| Female | 0.27 (0.23 to 0.31)* | 0.53 (0.48 to 0.58) | 0.20 (0.16 to 0.24) |
| Education<high school | 0.27 (0.23 to 0.31) | 0.55 (0.50 to 0.60) | 0.18 (0.15 to 0.21) |
| Education≥high school | 0.26 (0.22 to 0.30) | 0.53 (0.48 to 0.58) | 0.21 (0.18 to 0.24) |
| Age (year) (−1SD) | 0.24 (0.18 to 0.30) | 0.55 (0.48 to 0.63) | 0.21 (0.15 to 0.27) |
| Age (year) (+1SD) | 0.29 (0.25 to 0.34) | 0.52 (0.47 to 0.58) | 0.18 (0.15 to 0.22) |
| Neighbourhood deprivation score (−1SD) | 0.29 (0.23 to 0.35) | 0.53 (0.45 to 0.61)* | 0.18 (0.13 to 0.24) |
| Neighbourhood deprivation score (+1SD) | 0.24 (0.20 to 0.29) | 0.55 (0.49 to 0.62)* | 0.20 (0.15 to 0.25) |
| Percentage white population (−1SD) | 0.24 (0.16 to 0.31) | 0.53 (0.43 to 0.63)* | 0.23 (0.17 to 0.29) |
| Percentage white population (+1SD) | 0.29 (0.25 to 0.34) | 0.54 (0.49 to 0.60)* | 0.16 (0.13 to 0.20) |
| Percentage population ≤18 years (−1SD) | 0.22 (0.18 to 0.25) | 0.51 (0.45 to 0.56) | 0.28 (0.22 to 0.33) |
| Percentage population ≤18 years (+1SD) | 0.33 (0.26 to 0.39) | 0.57 (0.49 to 0.64) | 0.10 (0.08 to 0.13) |
| Percentage rental properties (−1SD) | 0.27 (0.22 to 0.33) | 0.60 (0.53 to 0.66)* | 0.13 (0.10 to 0.16) |
| Percentage rental properties (+1SD) | 0.26 (0.22 to 0.30) | 0.48 (0.43 to 0.54)* | 0.26 (0.21 to 0.31) |
| HPI (−1SD) | 0.29 (0.21 to 0.38) | 0.50 (0.41 to 0.59) | 0.21 (0.14 to 0.27) |
| HPI (+1SD) | 0.25 (0.20 to, 0.30) | 0.56 (0.50 to 0.62) | 0.19 (0.14 to 0.23) |
| Medium income (second tertile) | |||
| Black | 0.26 (0.21 to 0.31) | 0.55 (0.50 to 0.61)† | 0.19 (0.16 to 0.22) |
| White | 0.23 (0.18 to 0.9) | 0.61 (0.55 to 0.66)† | 0.16 (0.12 to 0.20) |
| Male | 0.25 (0.21 to 0.29)† | 0.58 (0.53 to 0.63) | 0.17 (0.14 to 0.21) |
| Female | 0.24 (0.21 to 0.28)† | 0.58 (0.54 to 0.63) | 0.17 (0.14 to 0.21) |
| Education<high school | 0.26 (0.21 to 0.30) | 0.59 (0.54 to 0.65) | 0.15 (0.11 to 0.18) |
| Education≥high school | 0.24 (0.21 to 0.27) | 0.58 (0.54 to 0.62) | 0.18 (0.15 to 0.21) |
| Age (year) (−1SD) | 0.25 (0.18 to 0.31) | 0.56 (0.48 to 0.64) | 0.19 (0.13 to 0.25) |
| Age (year) (+1SD) | 0.24 (0.20 to 0.29) | 0.59 (0.53 to 0.65) | 0.16 (0.13 to 0.20) |
| Neighbourhood deprivation score (−1SD) | 0.28 (0.22 to 0.34) | 0.54 (0.46 to 0.62)† | 0.18 (0.12 to 0.23) |
| Neighbourhood deprivation score (+1SD) | 0.20 (0.15 to 0.25) | 0.64 (0.56 to 0.72)† | 0.16 (0.11 to 0.21) |
| Percentage white population (−1SD) | 0.26 (0.17 to 0.34) | 0.53 (0.42 to 0.65)† | 0.21 (0.14 to 0.28) |
| Percentage white population (+1SD) | 0.24 (0.19 to 0.29) | 0.61 (0.55 to 0.67)† | 0.15 (0.11 to 0.19) |
| Percentage population ≤18 years (−1SD) | 0.19 (0.14 to 0.24) | 0.56 (0.50 to 0.61) | 0.26 (0.20 to 0.31) |
| Percentage population ≤18 years (+1SD) | 0.30 (0.24 to 0.37) | 0.61 (0.54 to 0.68) | 0.09 (0.06 to 0.11) |
| Percentage rental properties (−1SD) | 0.27 (0.22 to 0.31) | 0.65 (0.59 to 0.70) | 0.09 (0.06 to 0.11) |
| Percentage rental properties (+1SD) | 0.21 (0.16 to 0.26) | 0.52 (0.45 to 0.58) | 0.28 (0.22 to 0.33) |
| HPI (−1SD) | 0.26 (0.17 to 0.36) | 0.55 (0.44 to 0.66) | 0.19 (0.12 to 0.26) |
| HPI (+1SD) | 0.24 (0.19 to 0.28) | 0.60 (0.53 to 0.67) | 0.16 (0.11 to 0.21) |
| High income (third tertile) | |||
| Black | 0.25 (0.19 to 0.30) | 0.53 (0.46 to 0.60) | 0.22 (0.17 to 0.28) |
| White | 0.22 (0.17 to 0.28) | 0.48 (0.42 to 0.55) | 0.29 (0.24 to 0.35) |
| Male | 0.18 (0.14 to 0.22) | 0.52 (0.46 to 0.58) | 0.30 (0.25 to 0.35) |
| Female | 0.28 (0.23 to 0.33) | 0.48 (0.42 to 0.53) | 0.24 (0.20 to 0.29) |
| Education<high school | 0.21 (0.15 to 0.27) | 0.50 (0.42 to 0.59) | 0.28 (0.21 to 0.36) |
| Education≥high school | 0.24 (0.20 to 0.28) | 0.49 (0.44 to 0.54) | 0.27 (0.23 to 0.31) |
| Age (year) (−1SD) | 0.24 (0.12 to 0.37) | 0.54 (0.41 to 0.67) | 0.22 (0.14 to 0.30) |
| Age (year) (+1SD) | 0.23 (0.17 to 0.29) | 0.47 (0.40 to 0.54) | 0.30 (0.25 to 0.35) |
| Neighbourhood deprivation score (−1SD) | 0.29 (0.21 to 0.37) | 0.37 (0.28 to 0.45) | 0.35 (0.27 to 0.42) |
| Neighbourhood deprivation score (+1SD) | 0.16 (0.12 to 0.20) | 0.67 (0.59 to 0.75) | 0.17 (0.11 to 0.24) |
| Percentage white population (−1SD) | 0.23 (0.13 to 0.33) | 0.37 (0.28 to 0.47) | 0.40 (0.30 to 0.49) |
| Percentage white population (+1SD) | 0.22 (0.17 to 0.28) | 0.56 (0.50 to 0.63) | 0.21 (0.17 to 0.26) |
| Percentage population ≤18 years (−1SD) | 0.14 (0.07 to 0.21) | 0.48 (0.40 to 0.56) | 0.38 (0.28 to 0.48) |
| Percentage population ≤18 years (+1SD) | 0.33 (0.23 to 0.43) | 0.53 (0.40 to 0.66) | 0.14 (0.05 to 0.24) |
| Percentage rental properties (−1SD) | 0.25 (0.20 to 0.30) | 0.61 (0.54 to 0.67) | 0.14 (0.08 to 0.20) |
| Percentage rental properties (+1SD) | 0.10 (−0.001 to 0.21) | 0.19 (0.04 to 0.34) | 0.71 (0.50 to 0.91) |
| HPI (−1SD) | 0.21 (0.10 to 0.33) | 0.41 (0.29 to 0.53) | 0.38 (0.26 to 0.50) |
| HPI (+1SD) | 0.23 (0.18 to 0.27) | 0.48 (0.43 to 0.53) | 0.29 (0.25 to 0.34) |
Referent outcome=high road connectivity and activity-related and diet-related amenities.
*Indicates that the difference in the predicted probabilities in the low income tertile are statistically significantly different than the difference in the predicted probabilities in the high income tertile within each neighborhood cluster type, at the p<0.05 level.
†Indicates that the difference in the predicted probabilities in the medium income tertile are statistically significantly different than the difference in the predicted probabilities in the high income tertile within each neighborhood cluster type, at the p<0.05 level.
‡Predicted probabilities were generated from the model coefficients and depict the probability of residing in neighbourhoods with low (vs high) road connectivity and activity-related and diet-related amenities at fixed levels of the covariates (categories or ±1SD of mean).
§Variables include individual-level race, gender, educational attainment and age; neighbourhood-level deprivation score, percentage white population, percentage population ≤18 years and HPI; examination year (time); and baseline study centre.
¶Neighbourhood clusters created using non-hierarchical cluster analysis of measures related to connectivity and neighbourhood amenities, including intersection density (number of three-way, four-way and higher intersections per km2), links (count), cul-de-sacs (count), β-index (ratio of links to nodes), total road length (km), percentage residential (of total road length, km), distance from participants' residential location to nearest employment centre centroid (km), total park area within each neighbourhood (km2), and total physical activity facilities, supermarkets, co-ops and chain fast food restaurants (separately; count per km2 land).
CARDIA, Coronary Artery Risk Development in Young Adults; HPI, housing price index.