3.1 Reliability and validity of the SF-36 scale
3.1.1 Reliability of the scale
We obtained Spearman-Brown correction and overall Cronbach's α coefficients of 0.834 and 0.840, respectively for the scale, comfortably more than the thresholds for acceptable reliability of 0.7 and 0.6, respectively, strongly indicating that it has sufficient internal consistency and reliable ability to detect between-area differences in health parameters.
3.1.2 Validity of the SF-36 Health Scale
Content validity: The correlations between scores for each dimension of the SF-36 Health Scale and total scores, and correlations between the dimensions, were highly significant (p<0.001; Table 2), strongly indicating that it has sufficient content validity
Table 2 Correlation coefficients between subscales and between subscales and total scores.
|
BP
|
PF
|
RP
|
GH
|
VT
|
SF
|
RE
|
MH
|
HT
|
PF
|
.570**
|
|
|
|
|
|
|
|
|
RP
|
.544**
|
.490**
|
|
|
|
|
|
|
|
GH
|
.525**
|
.496**
|
.472**
|
|
|
|
|
|
|
VT
|
.415**
|
.484**
|
.474**
|
.492**
|
|
|
|
|
|
SF
|
.559**
|
.608**
|
.522**
|
.583**
|
.573**
|
|
|
|
|
RE
|
.309**
|
.222**
|
.545**
|
.321**
|
.371**
|
.468**
|
|
|
|
MH
|
.233**
|
.195**
|
.275**
|
.336**
|
.594**
|
.427**
|
.387**
|
|
|
HT
|
.351**
|
.300**
|
.327**
|
.403**
|
.208**
|
.293**
|
.222**
|
.130**
|
|
SFTOTAL
|
.680**
|
.722**
|
.674**
|
.788**
|
.796**
|
.794**
|
.529**
|
.652**
|
.426**
|
Abbreviations: BP, PF, RP, GH, VT, SF, RE, MH, HT—Body Pain, Physical Functioning, Role-Physical, General Health, Vitality, Social Functioning, Role-Emotional, Mental Health, and Reported Health Transition dimensions, respectively; **, P < 0.01 (indicating highly significant correlation).
Construct validity:PCA was performed for each dimension of the scale except HT, and two common factors with root >1 were extracted, which jointly explained 65.2% of the variance. Rotation using the Kaiser normalized maximum variance method to obtain loadings of each dimension on the two communal factors (Table 3) indicated that the PF, BP, GH, RP, and SF dimensions were correlated with the communal factor F1, and the MH, RE, and VT dimensions were correlated with the communal factor F2. The actual PF, BP, GH, MH, and VT models are completely consistent with the respective theoretical models, the actual RE and SF models are partially consistent with the respective theoretical models, and the actual RP model is not consistent with the theoretical model, but its loadings on the metric factors F1 and F2 are close to those of the corresponding theoretical model. In summary, the measurements are highly consistent with the overall conceptual model.
Table 3 Actual and theoretical factor loadings for models of dimensions of the SF-36 scale
Dimensions
|
Actual model
|
Theoretical model
|
F1
|
F2
|
Physiological factors
|
Psychological factors
|
PF
|
0.841
|
0.084
|
≥0.70
|
r≤0.30
|
BP
|
0.820
|
0.131
|
≥0.70
|
r≤0.30
|
GH
|
0.694
|
0.468
|
0.30<r<0.70
|
0.30<r<0.70
|
RP
|
0.657
|
0.385
|
≥0.70
|
r≤0.30
|
MH
|
0.052
|
0.881
|
r≤0.30
|
≥0.70
|
RE
|
0.256
|
0.673
|
r≤0.30
|
≥0.70
|
VT
|
0.442
|
0.685
|
0.30<r<0.70
|
0.30<r<0.70
|
SF
|
0.694
|
0.648
|
0.30<r<0.70
|
≥0.70
|
Abbreviations: BP, PF, RP, GH, VT, SF, RE, MH, HT—Body Pain, Physical Functioning, Role-Physical, General Health, Vitality, Social Functioning, Role-Emotional, Mental Health, and Reported Health Transition dimensions, respectively. Common factors were extracted by Principal Component Analysis and subjected to varimax rotation with Kaiser normalization.
3.2 Analysis of variance
Results of the ANOVA using the general linear model are shown in Table 4. The ANOVA revealed a highly significant association between residents' age and health (total SF-36 score), and significant associations between their health and both education level and neighborhood. Thus, although other factors are important, it clearly indicated that the mechanisms whereby neighborhood landscape and environmental factors influence public health warrant further attention.
Table 4 General linear model of indicated factors’ effects on resident’s heath (SFTOTAL scores).
|
Sum Squares
|
Degree of Freedom
|
Mean Square
|
F Value
|
p Value
|
Intercept Distance
|
Assumptions
|
133905.20
|
1
|
133905.20
|
767.41
|
0.000
|
Error
|
71698.22
|
410.903
|
174.49a
|
|
|
Age
|
Assumptions
|
4089.53
|
1
|
4089.53
|
23.21
|
0.000
|
Error
|
73656.60
|
418
|
176.21b
|
|
|
Education
|
Assumptions
|
891.57
|
1
|
891.57
|
5.06
|
0.025
|
Error
|
73656.60
|
418
|
176.21b
|
|
|
Neighborhood
|
Assumptions
|
6898.19
|
8
|
862.27
|
3.47
|
0.049
|
Error
|
1999.26
|
8.039
|
248.70c
|
|
|
Gender
|
Assumptions
|
85.18
|
1
|
85.18
|
0.34
|
0.574
|
Error
|
1998.53
|
8.035
|
248.72d
|
|
|
Neighborhood * Gender
|
Assumptions
|
1991.59
|
8
|
248.95
|
1.41
|
0.189
|
Error
|
73656.60
|
418
|
176.21b
|
|
|
Abbreviations: a, .019 MS(Gender) + .981 MS(Error); b, MS(Error); c, .997 MS (Neighborhood * Gender) + .003 MS(Error); d, .997 MS(Neighborhood * Gender) + .003 MS(Error).
3.3 Correlation analysis
Overall physical health (HI1) and mental health (HI2) scores for the participants were calculated using data shown in Table 3, then correlations between total health (SFTOTAL), HI1 and HI2 scores with environmental factors and landscape elements were analyzed (Tables 5 and 6, respectively). Correlations between environmental factors and landscape elements were also analyzed (Table 7).
Results of the correlation analysis show (Table 5) that five landscape elements (GSR, GCR, PWA, TD, and ATHWR) were highly significantly correlated, and two (PHOS and AIS) were significantly correlated with total health scores. In addition, five (GSA, GCR, PWA, TD, ATHWR, and PHOS) were highly significantly correlated with physiological health scores, while GCR and ATHWR were highly significantly correlated (and both GSR and PWA significantly correlated) with psychological health scores. TGCR, PGA and PSS were not significantly correlated with total, physiological or psychological health scores, indicating that these three indicators were not significantly correlated with public health. Thus, to simplify the mediating pathways they were excluded from subsequent analysis.
The correlations between landscape elements and SFTOTAL decreased in the following order: ATHWR > GCR> GSR > TD > PHOS > PWA > PSS > PGA > TGCR. Their correlations with physiological health scores decreased in the order GCR > GSR> TD > ATHWR >PHOS > PWA > AIS > TGCR > PSS > PGA, and their correlations with psychological health scores decreased in the order ATHWR >GCR > GSR > PWA > PHOS > AIS > TD > PSS > TGCR > PGA.
Table 5 Results of Pearson correlation analysis of landscape elements with total health, physical health, and mental health scores
Health Subscales
|
GSR
|
GCR
|
TGCR
|
PWA
|
TD
|
ATHWR
|
AIS
|
PHOS
|
PGA
|
PSS
|
SFTOTAL
|
.227**
|
.251**
|
0.006
|
.168**
|
-.171**
|
-.252**
|
.123*
|
.169**
|
-0.008
|
-0.015
|
HI1
|
.180**
|
.194**
|
-0.060
|
.125**
|
-.170**
|
-.156**
|
0.066
|
.130**
|
0.023
|
0.027
|
HI2
|
.112*
|
.136**
|
0.028
|
.110*
|
-0.041
|
-.165**
|
0.061
|
0.085
|
-0.016
|
-0.035
|
Abbreviations: HI1, Physical Component Summary score; HI2, Mental Component Summary score; GSR, green space ratio; GCR, green coverage ratio; TGCR, tree green cover ratio; PWA, percentage of water area; TD, trail density; ATHWR, average trail height to width ratio; AIS, average intersection spacing; PHOS, percentage of hard open space; PGA, percentage of garden architecture area; PSS, percentage of sports space area; * significant ( P<0.05); ** highly significant (p<0.01).
SFTOTAL scores were also highly significantly correlated with two environmental factors (I and RH) and significantly correlated with NOI, HSI and PM2.5 (Table 6). Two environmental factors (I and NOI) were highly significantly correlated, and another two (RH and PM2.5) significantly correlated with physiological health scores. In addition NOI was highly significantly correlated with psychological health scores. N and AWS were not significantly correlated with total, physiological or psychological health scores, and thus were excluded from subsequent analysis.
The correlations between environmental factors and SFTOTAL declined in the order NOI > I > RH > HSI > PM2.5 > N > AWS. The correlations between environmental factors and physiological health scores declined in the order I> NOI > PM2.5 > RH > HSI > N > AWS. The correlations between environmental factors and psychological health scores declined in the order NOI > RH > AWS > I > HSI > PM2.5 > N.
Table 6 Results of Pearson correlation analysis of relationships between total health (SFTOTAL), physical health and mental health scores and environmental factors
Health Subscales
|
N
|
I
|
AWS
|
RH
|
HSI
|
NOI
|
PM2.5
|
SFTOTAL
|
0.061
|
-.172**
|
0.021
|
.132**
|
-.116*
|
.228**
|
.104*
|
HI1
|
0.062
|
-.165**
|
-0.023
|
.101*
|
-0.088
|
.157**
|
.104*
|
HI2
|
0.004
|
-0.043
|
0.061
|
0.064
|
-0.040
|
.139**
|
0.020
|
Abbreviations: HI1, Physical Component Summary score; HI2, Mental Component Summary score. N, Noisiness; I, Illumination; AWS, Average Wind Speed; RH, Relative Humidity; HSi, Heat Stress Index; NOI, Negative Oxygen Ions: PM2.5, Fine Particulate Matter; * significant ( P<0.05); ** highly significant (p<0.01).
The analysis of correlations between landscape elements and environmental factors (Table 7) showed that TGCR was not significantly correlated with N and I, ATHWR was not significantly correlated with N, PHOS was not significantly correlated with AWS and RH, PGA was not significantly correlated with N, RH, HSI, or NOI, and PSS was not significantly correlated with N. The remaining landscape elements showed significant correlations with environmental factors, indicating that landscape elements are closely related in neighborhoods of high-density cities, and further mediation analysis is needed to clarify the pathways affecting public health.
Table 7 Results of Pearson correlation analysis between environmental factors and landscape elements
|
GSR
|
GCR
|
TGCR
|
PWA
|
TD
|
ATHWR
|
AIS
|
PHOS
|
PGA
|
PSS
|
N
|
.097*
|
.101*
|
0.025
|
-.353**
|
-.418**
|
0.019
|
.374**
|
.240**
|
0.005
|
0.067
|
I
|
-.383**
|
-.387**
|
0.045
|
-.283**
|
.904**
|
.543**
|
-.477**
|
-.133**
|
.344**
|
-.112*
|
AWS
|
.275**
|
.225**
|
.114*
|
.616**
|
.551**
|
-.231**
|
-.224**
|
0.011
|
.343**
|
-.662**
|
RH
|
.274**
|
.415**
|
.280**
|
.630**
|
-.599**
|
-.524**
|
.116*
|
0.084
|
0.068
|
.121*
|
HSI
|
-.631**
|
-.447**
|
-.284**
|
-.322**
|
.191**
|
.536**
|
-.378**
|
-.288**
|
0.087
|
.505**
|
NOI
|
.855**
|
.844**
|
.232**
|
.507**
|
-.253**
|
-.810**
|
.450**
|
.670**
|
0.004
|
-.226**
|
PM2.5
|
.180**
|
.314**
|
.117*
|
.190**
|
-.836**
|
-.341**
|
.490**
|
.210**
|
-.300**
|
.571**
|
Abbreviations: GSR, green space ratio; GCR, green coverage ratio; TGCR, tree green cover ratio; PWA, percentage of water area; TD, trail density; ATHWR, average trail height to width ratio; AIS, average intersection spacing; PHOS, percentage of hard open space; PGA, percentage of garden architecture area; PSS, percentage of sports space area; N, Noisiness; I, Illumination; AWS, Average Wind Speed; RH, Relative Humidity; HIS, Heat Stress Index; NOI, Negative Oxygen Ions: PM2.5, Fine Particulate Matter. *, p<0.05; **, p<0.01.
Table 8 Multiple mediation models linking adjusted landscape elements to public health (Outcome: SFTOTAL)
Landscape
|
Total Effect
|
Total Indirect Effect
|
Direct effect
|
Specific indirect effects (mediators tested one-at-a-time)
|
I
|
RH
|
HSI
|
NOI
|
PM2.5
|
GSR
|
0.22(0.15,0.29)*
|
0.18(0.01,0.35)*
|
0.04(-0.14,0.23)
|
0.07(0.01,0.14)*
|
0.01(-0.02,0.04)
|
-0.13(-0.23,-0.03)*
|
0.25(0.07,0.42)*
|
-0.02(-0.05,0.01)
|
GCR
|
0.25(0.16,0.32)*
|
0.2(-0.01,0.40)
|
0.05(-0.17,0.26)
|
0.07(0.01,0.14)*
|
0.01(-0.04,0.05)
|
-0.09(-0.17,-0.01)*
|
0.23(0.01,0.46)*
|
-0.03(-0.08,0.03)
|
PWA
|
0.15(0.06,0.23)*
|
0.15(0.05,0.25)*
|
-0.01(-0.14,0.12)
|
0.05(0.01,0.10)*
|
0.02(-0.07,0.10)
|
-0.07(-0.13,-0.01)*
|
0.17(0.09,0.26)*
|
-0.02(-0.05,0.02)
|
TD
|
-0.15(-0.24,-0.06)*
|
-0.23(-0.47,0.01)
|
0.08(-0.17,0.33)
|
-0.22(-0.42,-0.02)*
|
-0.01(-0.08,0.05)
|
0.04(0.01,0.09)*
|
-0.09(-0.14,-0.04)*
|
0.04(-0.12,0.20)
|
ATHWR
|
-0.25(-0.33,-0.17)*
|
-0.17(-0.39,0.02)
|
-0.08(-0.29,0.15)
|
-0.09(-0.19,0.01)
|
0.01(-0.06,0.06)
|
0.10(0.01,0.19)*
|
-0.21(-0.41,-0.02)*
|
0.02(-0.03,0.08)
|
AIS
|
0.15(0.08,0.22)*
|
0.05(-0.09,0.18)
|
0.11(-0.06,0.27)
|
0.11(0.02,0.19)*
|
0.01(-0.01,0.03)
|
-0.11(-0.18,-0.03)*
|
0.14(0.07,0.21)*
|
-0.10(-0.23,0.03)
|
PHOS
|
0.18(0.10,0.26)*
|
0.14(0.03,0.26)*
|
0.04(-0.10,0.18)
|
0.03(0.01,0.06)*
|
0.01(-0.01,0.01)
|
-0.06(-0.11,-0.01)*
|
0.20(0.07,0.32)*
|
-0.02(-0.06,0.02)
|
Abbreviations: GSR, green space ratio; GCR, green coverage ratio; TGCR, tree green cover ratio; PWA, percentage of water area; TD, trail density; ATHWR, average trail height to width ratio; AIS, average intersection spacing; PHOS, percentage of hard open space; I, Illumination; RH, Relative Humidity; HSI, Heat Stress Index; NOI, Negative Oxygen Ions; PM2.5, Fine Particulate Matter. A separate mediation model was fitted for each greenspace measure and each potential mediator. Independent indirect paths were tested. Standardized regression coefficients and 95% CIs are reported. a, adjusted for age and education. *, p<0.05.
3.4 Mediation analysis
3.4.1 Mediation effects of greenness on public health by physical environmental factors
Results of modeling with GSR and GCR as independent variables, I, RH, HSI, NOI, and PM2.5 as mediating variables, total health score (SFTOTAL) as the dependent variable, and age and education as covariates, with bootstrapping tests of mediation effects, are shown in Table 7, Table 8, and Figure 6.
Regression analysis showed that GSR had a positive effect on SFTOTAL (β=0.23, t=4.88, p<0.001; Table 8). Wen environmental factors were added as mediating variables, the direct effect of GSR on SFTOTAL was not significant (β=0.04, t=0.44, p=0.6615). However, effects of GSR on SFTOTAL through I (β=-0.19, t=-2.24, p< 0.05), HSI (β=0.21, t=2.64, p<0.05), and NOI (β=0.29, t=2.96, p<0.01) were significant, and none of the corresponding 95% confidence intervals contained 0, indicating significant mediating effects, with effect values of 0.07, -0.13, and 0.25, respectively. The environmental factors’ mediation of GSR’s effects on SFTOTAL decrease in the order NOI>HSI>I>PM2.5>RH, indicating that GSR’s effects on public health are mainly mediated through effects on the acoustic, thermal, and light environments.
GCR had a positive effect on SFTOTAL (β=0.25, t=5.77, p<0.001), but when environmental factors were added as mediating variables the direct effect of GCR on SFTOTAL was not significant (β=0.05, t=0.40, p=0.6886). Its overall mediating effect though physical environmental factors was also insignificant. However, GCR had significant effects through I (β=-0.19, t=-2.26, p<0.05), HSI (β=0.19, t=2.19, p<0.05), and NOI (β=0.28, t=2.11, p<0.05), with effect values of 0.07, -0.09, and 0.23, respectively. This suggests that GCR mainly affects public health through I, HSI, and NOI, probably with contributions from other mediators that require elucidation.
Taken together, although GCR did not play a significant role in the total mediating effect of the considered environmental factors, I, HSI, and NOI were significant mediators of effects of both greenness indicators, indicating that greenness affects public health mainly through its effects on the light, thermal, and atmospheric environments of the focal high-density urban neighborhoods.
3.4.2 Mediation effects of blue space on public health by physical environmental factors
Results of modeling with PWA as the independent variable, I, RH, HSI, NOI, and PM2.5 as mediating variables, total health score (SFTOTAL) as the dependent variable, and age and education as covariates, with bootstrapping tests of mediation effects, are also shown in Table 7, Table 8, and Figure 6. They show that PWA had a positive effect on SFTOTAL (β=0.15, t=3.38, p<0.001). When environmental factors were added as mediating variables, the direct effect of PWA on public health was not significant (β=-0.01, t=-0.11, p=0.915). However, it had significant effects through I (β=-0.20, t=-2.33, p<0.05), HSI (β=0.21, t=2.51, p<0.05), and NOI (β=0.33, t=4.54, p<0.001), with effect values of 0.05, -0.07, and 0.17, respectively. The environmental factors’ mediation of PWA’s effects on SFTOTAL decrease in the order NOI>HSI>I>RH>PM2.5, indicating that NOI, I and HSI are the main mediators of blue spaces’ effects on public health. These findings indicate that in neighborhoods of high-density cities, blue spaces mainly affect public health through atmospheric, light and thermal environments.
3.4.3 Mediation effects of walking trail systems on public health by physical environmental factors
Results of modeling with TD, ATHWR, and AIS as independent variables, I, RH, HSI, NOI, and PM2.5 as mediating variables, total health score (SFTOTAL) as the dependent variable, and age and education as covariates, with bootstrapping tests of mediation effects, are also shown in Table 7, Table 8, and Figure 6. They show that TD had a negative effect on SFTOTAL (β=-0.15, t=-3.42, p<0.001), but when environmental factors were added as mediating variables, the direct effect was not significant (β=0.08, t=0.66, p=0.5107). , while the direct effect on public health through I (β=-0.24, t=-2.20, p<0.05), HSI (β=0.21, t=2.64, p<0.05), and NOI (β=0.34, t=5.11, p<0.001) had a significant effect on public health.
The overall mediating effect of the considered physical environment factors was not significant either. However, the effects through I (β=-0.24, t=-2.20, p<0.05), HSI (β=0.21, t=2.64, p<0.05), and NOI (β=0.34, t=5.11, p<0.001) were significant, with effect values of -0.22, 0.04, and -0.09, respectively. In summary, TD affects public health through effects on light, thermal and acoustic environments, but the environmental factors had no significant overall mediating effect, and other factors likely contribute.
ATHWR had a negative effect on SFTOTAL (β=-0.25, t=-5.94, p<0.001), but when environmental factors were added as mediating variables, the direct effect was not significant (β=-0.08, t=-0.7, p=0.4863). The overall mediating effect through physical environment factors on public health was not significant either. However, ATHWR had significant effects through HSI (β=0.18, t=2.14, p<0.05), NOI (β=0.26, t=2.24, p<0.05), suggesting that it affects public health through the thermal and acoustic environments, probably with contributions from other mediating factors.
AIS had a positive effect on SFTOTAL (β=0.15, t=3.48, p<0.001). Again, when environmental factors were added as mediating variables the direct effect of AIS on public health was not significant (β=0.11, t=1.33, p=0.1856), and the overall mediating effect of physical environmental factors was not significant either. However, ASI had significant effects through I (β=-0.22, t=-2.60, p<0.05), NOI (β=0.30, t=4.50, p<0.001), and HSI (β=0.27 , t=2.92, p<0.05), with effect values of 0.11, -0.11, and 0.14, respectively. The results indicate that overall effects of AIS through the combined physical environment factors are insignificant, but it has significant effects mediated through light, acoustic and thermal environments, probably with contributions from other mediating factors.
Taken together, none of the three indicators of the walking trail system had significant effects on public health mediated through the combined physical environment factors. However, effects mediated through I, HSI, and NOI were significant. These findings indicate that in neighborhood environments of high-density cities walking trail systems have effects on public health mediated through the light, thermal, and atmospheric environments, probably in combination with other mediating factors and influence pathways that warrant further exploration and confirmation.
3.4.4 Mediation effects of hard open space on public health by physical environmental factors
Results of modeling with PHOS as the independent variable, I, RH, HSI, NOI, and PM2.5 as mediating variables, SFTOTAL as the dependent variable, and age and education as covariates, with mediation effects tested by bootstrapping are also shown in Table 7, Table 8, and Figure 6. They show that PHOS had a positive effect on public health (β=0.18, t=4.16, p<0.001), but again when environmental factors were added as mediating variables, the direct effect of PHOS on public health was not significant (β=0.04, t=0.52, p=0.6065). However, it had significant effects through I (β=-0.21, t=-2.39 , p<0.05), HSI (β=0.20, t=2.58, p<0.05), and NOI (β=0.29, t=3.27, p<0.05), with effect values of 0.03, -0.06, and 0.20, respectively. The mediating effects of PHOS on public health are, in descending order, NOI>HSI>I>PM2.5>RH. These results indicate that effects of hard open spaces on public health are mainly mediated through the atmospheric, thermal and light environments.