Linear regression results for the relationship between city greenness groups (low, medium and high) and cause-specific mortality rates
| Cause of death | City greenness | Male | Female |
| All causes | Low | 1.00† | 1.00† |
| Medium | 84.91 (−12.60 to 182.43)† | 51.39 (−10.28 to 113.06)† | |
| High | 132.90 (18.33 to 247.46)* † | 94.21 (21.76 to 166.66)* † | |
| Heart disease | Low | 1.00 | 1.00 |
| Medium | −19.67 (−78.12 to 38.77) | −7.64 (−44.82 to 29.54) | |
| High | 6.49 (−62.46 to 75.45) | 1.90 (−41.96 to 45.76) | |
| Diabetes | Low | 1.00 | 1.00 |
| Medium | 5.40 (−0.87 to 11.67) | 3.31 (−0.92 to 7.54) | |
| High | 4.34 (−3.06 to 11.73) | 4.18 (−0.81 to 9.17) | |
| Lung cancer | Low | 1.00 | 1.00 |
| Medium | −2.77 (−16.91 to 11.36) | −5.03 (−14.57 to 4.50) | |
| High | 7.93 (−8.75 to 24.60) | 2.47 (−8.78 to 13.72) | |
| Motor accident | Low | 1.00 | 1.00 |
| Medium | −3.97 (−11.32 to 3.38) | −2.38 (−6.70 to 1.95) | |
| High | 0.55 (−8.13 to 9.22) | −3.37 (−8.47 to 1.74) |
Models were adjusted for socioeconomic deprivation (median household income), ethnicity (percentage non-Hispanic white), particulate air pollution (PM10) and car dependency (% households with no car). Regression coefficients (plus 95% CIs) indicate the difference in rate per 100 000 population associated with that group compared with the least green cities. If removal of outliers was necessary to normalise model residuals this is indicated, and the result is included only if this changed the substantive finding.
*0.01≤p<0.05.
↵† Las Vegas and Tucson were excluded.