Content uploaded by Dan WANG
Author content
All content in this area was uploaded by Dan WANG on Oct 29, 2020
Content may be subject to copyright.
Original Study
Is Neighborhood Green Space Associated With Less Frailty? Evidence
From the Mr. and Ms. Os (Hong Kong) Study
Ruby Yu PhD
a
,
b
,
*, Dan Wang MSc
b
,
c
, Jason Leung MSc
c
, Kevin Lau PhD
b
,
d
,
e
,
Timothy Kwok MD
a
,
b
, Jean Woo MD
a
,
b
a
Department of Medicine and Therapeutics, Faculty of Medicine, Chinese University of Hong Kong, Hong Kong
b
CUHK Jockey Club Institute of Ageing, Chinese University of Hong Kong, Hong Kong
c
The Jockey Club Centre for Osteoporosis Care and Control, Chinese University of Hong Kong, Hong Kong
d
Institute of Future Cities, Chinese University of Hong Kong, Hong Kong
e
Institute of Environment, Energy and Sustainability, Chinese University of Hong Kong, Hong Kong
Keywords:
Neighborhood
green space
normalized difference vegetation index
frailty
transitions
physical activity
older adults
path analysis
abstract
Objectives: To examine whether neighborhood green space was related to frailty risk longitudinally and
to examine the relative contributions of green space, physical activity, and individual health conditions to
the frailty transitions.
Design, setting, and participants: Four thousand community-dwelling Chinese adults aged 65 years
participating in the Mr. and Ms. Os (Hong Kong) study in 2001-2003 were followed up for 2 years.
Methods: The percentage of green space within a 30 0-meter radial buffer around the participants’place
of residence was derived for each participant at baseline based on the normalized difference vegetation
index. Frailty status was classified according to the Fried criteria at baseline and after 2 years. Ordinal
logistic regression and path analysis were used to examine associations between green space and the
frailty transitions, adjusting for demographics, socioeconomic status, lifestyle factors, health conditions,
and baseline frailty status.
Results: At baseline, 53.5% of the participants met the criterion for robust, 41.5% were classified as pre-
frailty, and 5.0% were frail. After 2 years, 3240 participants completed all the measurements. Among
these, 18.6% of prefrail or frail participants improved, 66% remained in their frailty state, and 26.8% of
robust or prefrail participants progressed in frailty status. In multivariable models, the frailty status of
participants living in neighborhoods with more than 34.1% green space (the highest quartile) at baseline
was more likely to improve at the 2-year follow-up than it was for those living in neighborhoods with
0 to 4.5% (the lowest quartile) [odds ratio (OR): 1.29, 95% confidence interval (CI): 1.04-1.60; Pfor trend:
0.022]. When men and women were analyzed separately, the association between green space and frailty
remained significant in men (OR: 1.40, 95% CI: 1.03-1.90) but not in women. Path analysis showed that
green space directly affects frailty transitions (
b
¼0.041, P<.05) and also exerts an effect through
physical activity (
b
¼0.034, P<.05). Physical activity directly affects frailty (
b
¼0.134, P<.05), and also
indirectly affects frailty through health conditions including number of diseases (
b
¼0.057, P<.05) and
cognitive functions (
b
¼0.041, P<.05). The magnitude of the direct effect of green space on the 2-year
frailty transitions is comparable to those of the indirect effect through physical activity.
Conclusion: Older people living in neighborhoods with a higher percentage of green space were asso-
ciated with improvement in frailty status, independent of a wide range of individual characteristics.
Ó2018 AMDA eThe Society for Post-Acute and Long-Term Care Medicine.
Frailty represents a state of decline in functional reserves, which
increases the risk of adverse health outcomes such as morbidity,
disability, and institutionalization after a stressor event.
1
It can be
preceded by, but also occurs in the absence of, chronic disease
2
and
has been suggested as a better predictor of health and well-being than
the presence or absence of disease, representing an intermediate stage
The authors declare no conflicts of interest.
* Address correspondence to Ruby Yu, PhD, CUHK Jockey Club Institute of Ageing,
Suite 602, 6/F, Yasumoto International Academic Park, the Chinese University of
Hong Kong, Hong Kong.
E-mail address: rubyyu@cuhk.edu.hk (R. Yu).
JAMDA
journal homepage: www.jamda.com
https://doi.org/10.1016/j.jamda.2017.12.015
1525-8610/Ó2018 AMDA eThe Society for Post-Acute and Long-Term Care Medicine.
JAMDA 19 (2018) 528e534
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
between robust health and end of life.
1
Depending on the used defi-
nition, the prevalence of frailty at age 65 years varies from 4.0% to
59.1% across various studies.
3e5
Although frailty is common in older
populations, it is a dynamic process that is neither inevitable nor
irreversible as people age.
6,7
Therefore, attempts have been made to
reverse the syndrome of frailty, with physical exercise being the most
widely studied, and the findings were comparatively promising.
8
Neighborhood environments, particularly green space, are being
increasingly recognized as factors that influence physical activity.
Several studies in various countries have demonstrated that access to
and use of green space were associated with increased levels of
physical activity in older people.
9,10
It is therefore plausible that older
people living in neighborhoods with more green space are associated
with better health. Various studies have examined the relationship
between green space and health outcomes as well as their possible
mediators, with an aim of understanding the role of green space and
the underlying mechanisms of the associations. For example, in a
study carried out in Adelaide, Australia, perceived neighborhood
greenness was associated with physical and mental health in adults
aged 20-65 years.
11
In the Survey of the Health of Wisconsin, higher
levels of neighborhood green space were associated with lower levels
of symptomology for depression, anxiety, and stress in a probability
sample of Wisconsin residents aged 21-74 years.
12
In a cohort of in-
dividuals aged 45-72 carried out in Lithuania, distance to green spaces
was positively associated with a higher risk of the incidence of total
cardiovascular disease over 4 years.
13
Recent reviews and studies have
also indicated a relationship between a green environment and
obesity,
14,15
psychosocial well-being,
16
and mortality.
17,18
The potential
mechanisms for the relationship between a green living environment
and perceived general health have also been suggested,
19
including
that (1) the proportion of green space in the living environment could
stimulate physical activity and social contacts, both of which can in-
fluence a variety of health-related outcomes and (2) exposure to green
space can be psychologically restorative by promoting mental health.
However, the relationship between green space and health has
remained understudied in older populations. The potential effects of
neighborhood green space on frailty have not been studied.
Hong Kong, a special administrative region of China situated on the
Southern coast, is one of the most densely populated cities in the
world. More than 7 million people occupy an area of 1,070 km
2
,in
which built-up urban areas account for less than 300 km
2
. Within this
densely populated city variations in the quantity of green space exist,
which offer a natural setting in which to examine the impact of green
space on frailty in older people. Using a sample of 4,000 community-
dwelling Chinese men and women aged 65 years and older living in all
regions of Hong Kong, we examined whether neighborhood green
space was related to frailty risk longitudinally, and examined the
relative contributions of green space, physical activity, and individual
health conditions to the frailty transitions.
Methods and Methods
Study Design and Participants
A total of 4000 community-dwelling Chinese men and women
aged 65 and older were recruited for a cohort study on osteoporosis
and general health (Mr. and Ms. Os) in Hong Kong between August
2001 and December 2003 by placing recruitment notices in commu-
nity centers for older adults and housing estates. The aim was to re-
cruit a stratified sample so that approximately 33% each would be
aged 65 to 69, 70 to 74, and 75 and older. Those who were unable to
walk independently, had had bilateral hip replacement, or were not
competent to give informed consent were excluded. A team of trained
research assistants administered the study questionnaires and phys-
ical measurements for each individual on the same day. The cohort
was invited to re-attend for repeat questionnaire interviews and
physical measurements after 2 years. Details of the survey population
have been reported elsewhere.
20
In the present analysis, participants
without a valid address at baseline (n ¼56), were not living in their
baseline residence at the 2-year follow-up (n ¼125), were living in
Baseline (August 2001-December 2003)
(n=4000)
Exclusion
•Without a valid address at baseline (n=56)
•Were not living in the baseline residence at the 2-year
follow-up(n=125)
•The measured coverage of green space=100% (n=4)
•Did not have data on frailty at baseline (n=19)
Participants included in the analysis
(n=3796)
Participants completed the 2-year follow-up assessment with data
on re-assessment of frailty (August 2003-December 2005)
(n=3240)
Exclusion
•
•
•
Loss to follow-up after 2 years (n=359)
Known to have died during the 2-year follow-up (n=72)
Did not have data on re-assessment of frailty during the 2-year
follow-up (n=125)
Fig. 1. Study flow chart of sample selection.
R. Yu et al. / JAMDA 19 (2018) 528e534 529
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
neighborhoods with green space coverage of 100% (n ¼4), and/or did
not have data on frailty (n ¼19) at baseline were excluded. In addition,
participants known to have died (n ¼72), loss to follow-up (n ¼359),
and/or did not have data on re-assessment of frailty (n ¼125) during
the 2-year follow-up were also excluded. Therefore, 3240 participants
remained in the analysis (Figure 1). All participants gave written
consent and the study was approved by the Clinical Research Ethics
Committee of the Chinese University of Hong Kong.
Green Space
Green space was quantified based on the Normalized Difference
Vegetation Index (NDVI), which is an indicator to live vegetation on
the land surface.
21
This was estimated from spectral information ob-
tained from IKONOS satellite images, and the proportion of vegetation
within a 300-meter (m) radial buffer (approximately 5 minutes’
walking distance for older Chinese people)
22e25
was calculated for
each participant according to the geocoded information for partici-
pants’place of residence. NDVI value ranges between 1 and þ1.
Negative values of NDVI indicate water. Values below 0.1 but above
0 correspond to barren areas of rock, sand or snow. Values between
0.2 and 0.3 represent shrub and grassland, whereas higher values
indicate denser green leaves (eg, temperate and tropical rainforests).
21
In this study, NDVI 0.1 corresponded to green spaces. Using NDVI as a
measure of neighborhood greenness has been validated against ex-
perts’perception of greenness
26
and previously applied as a proxy for
greenness in health and behavior research.
10,12
Frailty and Frailty Transitions
Frailty was assessed using the 5-item Cardiovascular Health Study
(CHS) frailty phenotype, with the total score ranging from 0 to 5.
1
The
5 items are unintentional weight loss, self-rated exhaustion, weakness
(grip strength), slow walking speed, and low physical activity. The
equivalent variables used in this study for the construction of the CHS
score were body mass index less than 18.5, having no energy, grip
strength measurement in the lowest quartile, walking speed mea-
surement (6-meter usual pace test) in the lowest quartile, and the
Physical Activity Scale of the Elderly (PASE) score
27
in the lowest
quartile. The total scores were used to categorize participants as
robust (score ¼0), prefrail (score ¼1-2), and frail (score ¼3-5) at
baseline (2001-2003) and subsequently at the 2-year follow-up
(2003-2005). By comparing the states of frailty at baseline and after
the 2-year interval, frailty transitions were determined to be deteri-
orated, stable, or improved frailty status (Figure 2).
Covariates
Multiple questionnaires were used to collect information on (1)
demographic characteristics, including age, sex, and marital status; (2)
self-reported socioeconomic status on a ladder consisting of 10 rungs,
with the top rung indicating people with the most satisfied income,
education level, and job respect
28,29
; (3) lifestyle, including smoking,
alcohol intake, diet quality measured by Diet Quality Index (DQI),
30
and
physical activity (as measured by the PASE); and (4) health conditions
indicated by depression [as assessed by the Geriatric Depression Scale
(GDS), with a score below 8 indicating depressive symptoms],
31
cognitive function [as assessed by the Mini-Mental State Examination
(MMSE), with a score below 24 indicating cognitive impairment],
32,33
and the number of reported diseases (maximum 11 diseases).
Statistical Analysis
For participants whose frailty states were completely assessed at
baseline and at the 2-year follow-up, characteristics were compared
across transitions in frailty status using analysis of variance (ANOVA)
Fig. 2. Transitions in frailty status over 2 years. *Loss to follow-up or incomplete frailty data.
R. Yu et al. / JAMDA 19 (2018) 528e534530
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
for continuous variables and chi-square tests for categorical variables.
Tests for trend were performed, adjusting for age, sex, and baseline
frailty status using ANCOVA for continuous variables and logistic
regression for categorical variables as appropriate. Ordinal logistic
regression was used to evaluate the effect of green space in quartile
with odds ratios (ORs) and 95% confidential intervals (CIs) reported,
controlling for demographics, socioeconomic status, lifestyle, and
health conditions. Marital status was categorized into 4 levels: mar-
ried; widowed; separated or divorced; and single. Smoking was
categorized into 2 levels: current smoker and noncurrent smoker.
Alcohol intake was categorized in to 2 levels: at least 12 alcoholic
drinks in the past 12 months and less than 12 drinks in the past
12 months; as analysis of our data found that those who had at least 12
alcoholic drinks in the past 12 months had lower levels of physical
activity, which is an important factor associated with frailty, compared
with those who had less than 12 drinks in the past 12 months.
Depression was categorized in to 2 levels: GDS <8 and GDS 8.
Baseline frailty status was also adjusted in the models because par-
ticipants with frailty at baseline may be less likely to use green space.
The analysis was repeated separately for men and women because of
the sex differences in the usage of green space and the pattern of
frailty. In addition, path analysis was carried out to determine the
direct and indirect effects of green space, physical activity, and indi-
vidual health conditions on the 2-year frailty transitions. Participants
with missing data affecting the assessment of the 2-year transitions in
frailty status were excluded from the analyses. All analyses were
performed using SPSS Statistics v24.0 (IBM Corp. Released 2013.
Armonk, NY), with the 2-tailed significance level of P<.05.
Results
Table 1 presents baseline characteristics of participants. The mean
age of the participants was 72.2 years, the proportion of women was
49.2%, and most participants were married (71.7%). The percentage of
green space within a 300-m radial buffer around each participant’s
place of residence ranged from 0 to 99.4%, with a mean of 21.2%.
Participants living in neighborboods with a higher percentage of green
space were associated with a higher likelihood of improvement in
frailty status (adjusted P-trend ¼0.014). In addition, higher physical
activity and lower number of diseases were associated with a higher
likelihood of remaining in or improvement in frailty status, whereas
depression was associated with higher likelihood of deterioration in
frailty status.
Figure 2 shows the frailty transition between baseline and the
2-year follow-up. At baseline, 53.5% of the participants met the cri-
terion for robust, 41.5% were classified as prefrailty, and 5.0% were
classified as frail. After 2 years, 3240 participants completed all the
measurements. Among these, 18.6% of prefrail or frail participants
improved, 66.0% remained in their frailty state, and 26.8% of robust or
prefrail participants progressed in frailty status.
In multivariable models, frailty status of participants living in
neighborhoods with more than 34.1% green space (the highest
quartile) at baseline was more likely to improve at the 2-year follow-
up than it was for those living in neighborhoods with 0 to 4.5% (the
lowest quartile) (OR: 1.29, 95% CI: 1.04-1.60). A linear trend was also
observed (P¼.022). When men and women were analyzed sepa-
rately, the association between green space and frailty remained
Table 1
Baseline Characteristics of Study Participants
Characteristic All Participants Transition States of Frailty PValue PTrend PTrend
Adjusted*
Deteriorated in
Frailty Status
Stable in
Frailty Status
Improved in
Frailty Status
n¼3240 n ¼835 n ¼2140 n ¼265
Age, mean SD 72.2 5.0 72.8 5.1 72.0 5.0 71.8 4.6 <.001 .003 /
Sex, n (%)
Men 1646 (50.8) 346 (21.0) 1151 (69.9) 149 (9.1) <.001 <.001 /
Women 1594 (49.2) 489 (30.7) 989 (62.0) 116 (7.3)
Marital status, n (%)
Married 2324 (71.7) 559 (24.1) 1579 (67.9) 186 (8.0) .004 / /
Widowed 768 (23.7) 236 (30.7) 470 (61.2) 62 (8.1)
Separated/divorced 78 (2.4) 23 (29.5) 44 (56.4) 11 (14.1)
Single 70 (2.2) 17 (24.3) 47 (67.1) 6 (8.6)
SES ladder, mean SD 4.6 1.8 4.6 1.8 4.6 1.9 4.4 1.7 .173 .118 .639
Smoking status (%)
Nonsmoker 2036 (62.8) 572 (28.1) 1312 (64.4) 152 (7.5) <.001 <.001 .015
Current smoker 1204 (37.2) 263 (21.8) 828 (68.8) 113 (9.4)
Current alcohol use, n (%)
No 2803 (86.5) 747 (26.7) 1835 (65.5) 221 (7.9) .009 .002 .114
Yes 437 (13.5) 88 (20.1) 305 (69.8) 44 (10.1)
DQI score, mean SD 64.7 9.5 65.3 9.3 64.7 9.5 63.4 9.7 .014 .004 .541
Physical activity 92.8 43.0 92.6 37.3 92.7 45.2 94.3 42.2 .844 .600 <.001
Depression (GDS8), n (%)
No 2967 (91.6) 754 (25.4) 1976 (66.6) 237 (8.0) .076 .558 .001
Yes 272 (8.4) 81 (29.8) 163 (59.9) 28 (10.3)
MMSE score, mean SD 25.8 3.5 25.4 3.8 26.0 3.4 25.9 3.5 <.001 .051 .062
No. of diseases 2.0 1.4 2.1 1.4 2.0 1.4 2.0 1.3 .023 .114 .022
Green space quartile, n (%)
Q1 (0.00-4.53) 798 (24.7) 227 (28.5) 519 (65.0) 52 (6.5) .112 .005 .014
Q2 (4.54-13.20) 806 (25.0) 204 (25.3) 540 (67.0) 62 (7.7)
Q3 (13.21-34.12) 805 (25.0) 212 (26.3) 519 (64.5) 74 (9.2)
Q4 (34.13) 818 (25.4) 190 (23.2) 552 (67.5) 76 (9.3)
Baseline frailty status, n (%)
Robust 1818 (56.1) 686 (37.7) 1132 (62.3) 0 (0.0) <.001 <.001 /
Prefrail 1295 (40.0) 149 (11.5) 934 (72.1) 212 (16.4)
Frail 127 (3.9) 0 (0.0) 74 (58.3) 53 (41.7)
DQI, Dietary Quality Index; GDS, Geriatric Depression Scale; MMSE, Mini-Mental State Examination; SD, standard deviation; SES, socioeconomic status.
*Adjusted for age, sex, and baseline frailty status.
R. Yu et al. / JAMDA 19 (2018) 528e534 531
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
significant in men (OR: 1.40, 95% CI: 1.03-1.90) but not in women
(Table 2).
Figure 3 shows the estimates of the direct and indirect effects of
green space, physical activity, and health conditions on the 2-year
frailty transitions. Green space directly affects frailty transitions
(standardized regression coefficient
b
¼0.041, P<.05) and also exerts
an effect through physical activity (
b
¼0.034, P<.05). Physical activity
directly affects frailty (
b
¼0.134, P<.05), and also indirectly affects
frailty through health conditions including number of diseases
(
b
¼0.057, P<.05) and cognitive functions (
b
¼0.041, P<.05), which
also exert direct effects on frailty (
b
¼0.043 and 0.056 respectively,
P<.05). GDS was not related to green space but it directly affects
frailty (
b
¼0.045, P<.05) as well as indirectly through a number of
diseases (
b
¼0.129, P<.05) and cognitive function (
b
¼0.137,
P<.05). MMSE was not related to green space. The magnitude of the
direct effect on green space on the frailty transitions is comparable to
those of the indirect effect though physical activity (Figure 3).
Discussion
Strong evidence exists that frailty can be prevented by an increase
in physical activity, which is more likely to be achieved if the envi-
ronments are supportive of physical activity. Green space is now
viewed as an important part of neighborhoods that support healthy
and active lifestyles; it is therefore plausible that individuals living in
neighborhoods with more green space have a lower risk of frailty. To
our knowledge, this is the first study to explore this hypothesis
longitudinally. Our findings revealed that older people living in
neighborhoods with a higher percentage of green space were associ-
ated with improvement in frailty status, independent of a wide range
of individual characteristics. This is consistent with our previous
analysis, which found the relationship between neighborhood green
space and mortality to be equally strong.
18
The finding showed that
more green space is associated with improvement of the frailty con-
dition is also consistent with previous reports of increased ill health
(eg, cardiovascular diseases, diabetes) with fewer green space in older
adults,
13,34
reinforcing the potential health benefits of green space.
To some extent, the basis of our hypothesis was that green space
promotes physical activity, which then has an influence on frailty risk,
possibly through physical (eg, decreased number of disease), cognitive
(increased MMSE), and psychological (decrease GDS) improvements.
The path analysis supported this hypothesis, showing that physical
activity was a mediator of the relationship between green space and
frailty transitions. This is consistent with previous findings in the
Caerphilly Prospective Study that older men living in neighborhoods
with more green space have higher levels of participation in regular
physical activity.
10
Similar observations were demonstrated in previ-
ous studies, in which proximity to parks and trails was associated with
greater walking in older age.
35e40
While green space may encourage higher levels of physical activity,
this is likely to be only part of the explanation why frailty risk is lower
in neighborhoods with more green space. There is increasing evidence
that green space may influence health by directly promoting cognitive
functions and well-being, which are strongly related to the onset of
frailty.
41,42
Previous studies demonstrated significant associations
between neighborhood green spaces and increased memory span and
mood,
43
increased memory and attention,
44
better mental health,
45
and reduced risks of having psychological distress,
46
possibility due
to stress reduction
47,48
and attention restoration.
49e52
We did not find
any associations between green space and cognitive function (as
measured by MMSE) or mental health (as measured by GDS), but re-
sults from the path analysis showed that physical activity was asso-
ciated with both MMSE and GDS, suggesting that green space may
offer cognitive and mental benefits through physical activity, which in
turn reduce the risk of frailty. A recent study also revealed that
walking (typical low-intensity physical activity for older people) in an
urban green space environment is likely to trigger changes in the
levels of excitement, engagement and frustration in the brains of the
older people living in urban areas.
53
It has also been suggested that the positive effects of green space
on health could be through mechanisms of reduced air pollution;
54,55
exposure to air pollution has been linked to hospital admissions for
respiratory diseases, and may be contributory to variations in health
outcomes including frailty.
56
Future studies are needed to examine the
role of air pollution in mediating the observed association between
green space and frailty.
Our findings also demonstrated a gender difference in the associ-
ation between green space and frailty, where men, but not women,
Table 2
Multivariable Ordinal Regression Models of the 2-Year Frailty Transitions According to Quartiles of Green Space
Deteriorated in
Frailty Status
Stable in
Frailty Status
Improved in
Frailty Status Crude
OR (95% CI)
Adjusted
y
(Model 2)
Adjusted*(Model 1)
All participants n ¼833 n ¼2130 n ¼264
Green space quartile, n (%)
Q1 (0.00-4.53) 227 (28.5) 519 (65.0) 52 (6.5) 1 1 1
Q2 (4.54-13.20) 204 (25.3) 540 (67.0) 62 (7.7) 1.17 (0.96, 1.44) 1.16 (0.94, 1.44) 1.13 (0.91, 1.40)
Q3 (13.21-34.12) 212 (26.3) 519 (64.5) 74 (9.2) 1.18 (0.97, 1.45) 1.18 (0.95, 1.46) 1.18 (0.95, 1.47)
Q4 (34.13) 190 (23.2) 552 (67.5) 76 (9.3) 1.34 (1.10, 1.65) 1.33 (1.07, 1.64) 1.29 (1.04, 1.60)
Ptrend ddd0.006 0.0135 0.022
Men n ¼346 n ¼1148 n ¼149
Green space quartile, n (%)
Q1 (0.00-4.53) 108 (24.7) 294 (67.3) 35 (8.0) 1 1 1
Q2 (4.54-13.20) 87 (20.9) 294 (70.7) 35 (8.4) 1.19 (0.89, 1.59) 1.16 (0.86, 1.56) 1.11 (0.83, 1.50)
Q3 (13.21-34.12) 90 (22.6) 267 (67.1) 41 (10.3) 1.19 (0.89, 1.59) 1.08 (0.80, 1.46) 1.06 (0.78, 1.43)
Q4 (34.13) 61 (15.6) 293 (74.7) 38 (9.7) 1.56 (1.16, 2.10) 1.47 (1.08, 1.99) 1.40 (1.03, 1.90)
Ptrend ddd0.005 0.0287 0.056
Women n ¼487 n ¼982 n ¼115
Green space quartile, n (%)
Q1 (0.00-4.53) 119 (33.0) 225 (62.3) 17 (4.7) 1 1 1
Q2 (4.54-13.20) 117 (30.0) 246 (63.1) 27 (6.9) 1.19 (0.89, 1.58) 1.18 (0.86, 1.61) 1.18 (0.86, 1.62)
Q3 (13.21-34.12) 122 (30.0) 252 (61.9) 33 (8.1) 1.23 (0.92, 1.64) 1.28 (0.94, 1.75) 1.31 (0.95, 1.79)
Q4 (34.13) 129 (30.3) 259 (60.8) 38 (8.9) 1.24 (0.94, 1.65) 1.23 (0.90, 1.67) 1.23 (0.90, 1.68)
Ptrend ddd0.138 0.1715 0.156
*Model 1: adjusted for age, sex, marital status, socioeconomic status, current smoking status, alcohol intake, diet quality, and baseline frailty status.
y
Model 2: adjusted for covariates in model 1 and also for number of diseases, cognitive function, physical activity, and depression.
R. Yu et al. / JAMDA 19 (2018) 528e534532
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
living in neighborhoods with a higher percentage of green space had a
higher likelihood of improvement in frailty status at the 2-year follow-
up, compared to those living in neighborhoods with a lower per-
centage of green space. These findings are consistent with results of
previous studies in which gender differences were observed in the
associations between urban green space and health outcomes in
terms of cardiovascular disease and respiratory disease mortality rates
as well as mental health.
57,58
Possible explanations for the gender
differences in the green space and health relationship are gender
differences in the usage of green space. Although we did not collect
any information regarding the usage of green space, our supplemen-
tary analysis revealed that men had a higher level of physical activity
and were more likely than women to have spent their leisure time
outdoors (walking and gardening) (data not shown), and therefore
benefit more from green space than women do. This finding is
consistent with the results of the path analysis, showing that physical
activity is an important mediator of the association. Therefore, it is
important to take gender into account when considering any associ-
ations between urban green space and health, since men and women
may experience and respond to urban green space in different ways.
This study had several limitations. First, the volunteer subjects
tend to be more health-conscious and had a higher educational
standard compared with the general Hong Kong elderly. Second, green
space was derived from NDVI, but does not necessarily reflect the
utilization of green spaces due to accessibility issues. Nevertheless,
about 90% of Hong Kong residents have access to public parks or open
spaces which are reachable within a walking distance of 400 m from
their homes.
59
In addition, the study only focused on the association
between baseline green space and the frailty transitions, and we
lacked information on green space changes. However, it is unlikely
that green space in Hong Kong changed significantly during the study
period. Because of the limited land availability and conservation of
country parks, there have not been any large-scale deforestation and
tree planting exercise during those 2 years. Third, physical activity was
self-reported and may not necessarily have occurred within their
neighborhoods. Fourth, our analyses were also limited to fixed cova-
riates (eg, smoking status, alcohol intake, and diet quality) collected at
baseline. Time-varying covariates could be used to incorporate infor-
mation on changes that occurred during the 2-year follow-up. Fifth,
there may have been some important confounders and potential
mediators (eg, neighborhood safety, social network, social participa-
tion, sense of community, and air quality) that we were unable to
account for in the models. Sixth, the study has a short-term follow-up.
Additional longer follow-up studies will be needed to determine the
long-term effects of green space on the development of frailty.
Nevertheless, this study involved significant strengths, including the
use of a longitudinal design, the use of high-quality geographic in-
formation system data to quantify green space, and the ability to ac-
count for the influence of a wide range of individual characteristics in
the models.
Conclusions
Our findings suggest that older people living in neighborhoods that
have a higher percentage of green space had a higher likelihood of
improvement in frailty status. The observed association between
green space and frailty risk was stronger among men than women.
Although green space may promote higher levels of physical activity,
and in turn better health, better well-being, and lower risk of frailty,
other mechanisms such as reduced air pollution are likely to be
pertinent. Our findings have important implications for planning
policy to design and encourage the use of neighborhood green space
in promoting health and preventing frailty in ageing populations.
Nevertheless, because of the limited land availability in Hong Kong,
urban green spaces may not be easily expanded. It is therefore
important to maintain and improve existing urban green spaces. Our
findings also lay a path for further research to understand which
characteristics of green space (eg, safety, availability and variety of
facilities, hygiene and aesthetic features) have the strongest influence
on frailty, and to address the mechanisms by which the beneficial
effects of green space on health occur, which in turn will inform urban
green space interventions.
References
1. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: Evidence for a
phenotype. J Gerontol Ser A Biol Sci Med Sci 2001;56:M146eM156.
2. Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits.
J Gerontol Ser A Biol Sci Med Sci 2007;62:722e727.
3. Santos-Eggimann B, Cuenoud P, Spagnoli J, Junod J. Prevalence of frailty in
middle-aged and older community-dwelling Europeans living in 10 countries.
J Gerontol Ser A Biol Sci Med Sci 2009;64:675e681.
4. Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in
community-dwelling older persons: A systematic review. J Am Geriatr Soc
2012;60:1487e1492.
Green Space
Physical Activity
(PASE)
Depression
(GDS)
No. of disease
Cognitive function
(MMSE)
Two-year frailty transitions
(improvement in frailty status)
0.034*
0.012
-0.034*
0.041*
0.129*
0.134*
-0.057*
-0.137*
-0.045*
0.041*
-0.043*
0.056*
0.010
Fig. 3. GDS, Geriatric Depression Scale; PASE, Physical Activity Scale of the Elderly; MMSE, Mini-Mental State Examination. The path analysis model shows the direct and indirect
effects of neighborhood green space on the 2-year frailty transitions (improvement in frailty status from baseline to the 2-year follow-up). Coefficients within paths are standardized
regression coefficients. The model includes age, sex, marital status, socioeconomic status, current smoking status, alcohol intake, diet quality, and baseline frailty status. *P<.05.
R. Yu et al. / JAMDA 19 (2018) 528e534 533
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
5. Woo J, Yu R, Wong M, et al. Frailty screening in the community using the FRAIL
scale. J Am Med Dir Assoc 2015;16:412e419.
6. Gill TM, Gahbauer EA, Allore HG, Han L. Transitions between frailty states
among community-living older persons. Arch Intern Med 2006;166:418e423.
7. Lee JSW, Auyeung TW, Leung J, et al. Transitions in frailty states among
community-living older adults and their associated factors. J Am Med Dir Assoc
2014;15:281e286.
8. Theou O, Stathokostas L, Roland KP, et al. The effectiveness of exercise in-
terventions for the management of frailty: A systematic review. J Aging Res
2011;2011:569194.
9. Mytton OT, Townsend N, Rutter H, Foster C. Green space and physical activity:
An observational study using Health Survey for England data. Health Place
2012;18:1034e1041.
10. Gong Y, Gallacher J, Palmer S, Fone D. Neighbourhood green space, physical
function and participation in physical activities among elderly men: The
Caerphilly prospective study. Int J Behav Nutr Phy Act 2014;11:40.
11. Sugiyama T, Leslie E, Giles-Corti B, Owen N. Associations of neighbourhood
greenness with physical and mental health: Do walking, social coherence and
local social interaction explain the relationships? J Epidemiol Community
Health 2008;62:e9.
12. Beyer KMM, Kaltenbach A, Szabo A, et al. Exposure to neighborhood green
space and mental health: Evidence from the survey of the health of Wisconsin.
Int J Environ Res Public Health 2014;11:3453e3472.
13. Tamosiunas A, Grazuleviciene R, Luksiene D, et al. Accessibility and use of
urban green spaces, and cardiovascular health: Findings from a Kaunas cohort
study. Environ Health 2014;13:20.
14. Feng J, Glass TA, Curriero FC, et al. The built environment and obesity: A sys-
tematic review of the epidemiologic evidence. Health Place 2010;16:175e190.
15. Lachowycz K, Jones AP. Greenspace and obesity: A systematic review of the
evidence. Obes Rev 2011;12:e183ee189.
16. Bowler DE, Buyung-Ali LM, Knight TM, Pullin AS. A systematic review of evi-
dence for the added benefits to health of exposure to natural environments.
BMC Public Health 2010;10:456.
17. Gascon M, Triguero-Mas M, Martinez D, et al. Residential green spaces and
mortality: A systematic review. Environ Int 2016;86:60e67.
18. Wang D, Lau KK, Yu R, et al. Neighbouring green space and mortality in
community-dwelling elderly Hong Kong Chinese: A cohort study. BMJ Open
2017;7:e015794.
19. Maas J, Verheij RA, Groenewegen PP, et al. Green space, urbanity, and health:
How strong is the relation? J Epidemiol Community Health 2006;60:587e592.
20. Wong SYS, Kwok T, Woo J, et al. Bone mineral density and the risk of peripheral
arterial disease in men and women: Results from Mr. and Ms Os, Hong Kong.
Osteoporosis Int 2005;16:1933e1938.
21. NASA. Landsat 7 Science Data Users Handbook. Greenbelt, MD: Goddard Space
Flight Center, NASA; 2009.
22. Grahn P, Stigsdotter UA. Landscape planning and stress. Urban Forestry Urban
Green 2003;2:1e18.
23. Nielsen TS, Hansen KB. Do green areas affect health? Results from a Danish
survey on the use of green areas and health indicators. Health Place 2007;13:
839e850.
24. Annerstedt M, Ostergren PO, Bjork J, et al. Green qualities in the neighbour-
hood and mental healthdresults from a longitudinal cohort study in Southern
Sweden. BMC Public Health 2012;12:337.
25. Bjork J, Albin M, Grahn P, et al. Recreational values of the natural environment
in relation to neighbourhood satisfaction, physical activity, obesity and well-
being. J Epidemiol Community Health 2008;62:e2.
26. Rhew IC, Vander Stoep A, Kearney A, et al. Validation of the normalized dif-
ference vegetation index as a measure of neighborhood greenness. Ann Epi-
demiol 2011;21:946e952.
27. Washburn RA, Smith KW, Jette AM, Janney CA. The physical-activity scale for
the elderly (PASE)ddevelopment and evaluation. J Clin Epidemiol 1993;46:
153e162.
28. Woo J, Lynn H, Leung J, Wong SY. Self-perceived social status and health in
older Hong Kong Chinese women compared with men. Women Health 2008;
48:209e234.
29. Woo J, Leung J, Chan R, Chau PH. Influence of income and self-rated socio-
economic position on lifestyle, and physical and psychological function in older
Chinese adults aged 65 years and over. Public Health 2013;127:878e881.
30. Kim S, Haines PS, Siega-Riz AM, Popkin BM. The diet quality index-
international (DQI-I) provides an effective tool for cross-national comparison
of diet quality as illustrated by China and the United States. J Nutr 2003;133:
3476e3484.
31. Lee HCB, Chiu HFK, Kwok WY, et al. Chinese elderly and the GDS short form: A
preliminary study. Clin Gerontol 1994;14:37e42.
32. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state.”A practical method
for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;
12:189e198.
33. Katzman R, Zhang MY, Ouang Ya Q, et al. A Chinese version of the mini-mental
state examination: Impact of illiteracy in a Shanghai dementia survey. J Clin
Epidemiol 1988;41:971e978.
34. Astell-Burt T, Feng XQ, Kolt GS. Is neighborhood green space associated with a
lower risk of type 2 diabetes? Evidence from 267,072 Australians. Diabetes
Care 2014;37:197e201.
35. Cohen DA, McKenzie TL, Sehgal A, et al. Contribution of public parks to physical
activity. Am J Public Health 2007;97:509e514.
36. Mowen A, Orsega-Smith E, Payne L, et al. The role of park proximity and social
support in shaping park visitation, physical activity, and perceived health
among older adults. J Phys Act Health 2007;4:167e179.
37. Duncan M, Mummery K. Psychosocial and environmental factors associated
with physical activity among city dwellers in regional Queensland. Prev Med
2005;40:363e372.
38. Giles-Corti B, Donovan RJ. The relative influence of individual, social and
physical environment determinants of physical activity. Soc Sci Med 2002;54:
1793e1812.
39. Abildso CG, Zizzi S, Abildso LC, et al. Built environment and psychosocial
factors associated with trail proximity and use. Am J Health Behav 2007;31:
374e383.
40. Michael YL, Perdue LA, Orwoll ES, et al. Physical activity resources and changes
in walking in a cohort of older men. Am J Public Health 2010;100:654e660.
41. Atkinson HH, Rosano C, Simonsick EM, et al. Cognitive function, gait speed
decline, and comorbidities: The health, aging and body composition study.
J Gerontol A Biol Sci Med Sci 2007;62:844e850.
42. Raji MA, Al Snih S, Ostir GV, et al. Cognitive status and future risk of frailty
in older Mexican Americans. J Gerontol A Biol Sci Med Sci 2010;65:
1228e1234.
43. Berman MG, Kross E, Krpan KM, et al. Interacting with nature improves
cognition and affect for individuals with depression. J Affect Disord 2012;140:
300e305.
44. Gidlow CJ, Jones MV, Hurst G, et al. Where to put your best foot forward:
Psycho-physiological responses to walking in natural and urban environments.
J Environ Psychol 2016;45:22e29.
45. Triguero-Mas M, Dadvand P, Cirach M, et al. Natural outdoor environments and
mental and physical health: Relationships and mechanisms. Environ Int 2015;
77:35e41.
46. Pope D, Tisdall R, Middleton J, et al. Quality of and access to green space in
relation to psychological distress: Results from a population-based cross-
sectional study as part of the EURO-URHIS 2 project. Eur J Public Health; 2015
[Epub ahead of print].
47. Ulrich RS. Aesthetic and affective response to natural environment. In:
Altman I, Wohlwill JF, editors. Behavior and the Natural Environment. Boston,
MA: Springer; 1983. p. 85e125.
48. Ulrich RS, Simons RF, Losito BD, et al. Stress recovery during exposure to
natural and urban environments. J Environ Psychol 1991;11:201e230.
49. Kaplan R, Kaplan S. The Experience of Nature: A Psychological Perspective.
Cambridge: Cambridge University Press; 1989.
50. Kaplan S. The Restorative benefits of naturedtoward an integrative frame-
work. J Environ Psychol 1995;15:169e182.
51. Kaplan S. Meditation, restoration, and the management of mental fatigue.
Environ Behav 2001;33:480e506.
52. Kaplan R, Kaplan S. Well-being, reasonableness, and the natural environment.
Appl Psychol Health Well Being 2011;3:304e321.
53. Tilley S, Neale C, Patuano A, Cinderby S. Older people’s experiences of mobility
and mood in an urban environment: A mixed methods approach using elec-
troencephalography (EEG) and interviews. Int J Environ Res Public Health
2017;14:151.
54. Liu CF, Li XM. Carbon storage and sequestration by urban forests in Shenyang,
China. Urban Forestry Urban Green 2012;11:121e128.
55. Nowak DJ, Greenfield EJ, Hoehn RE, Lapoint E. Carbon storage and sequestra-
tion by trees in urban and community areas of the United States. Environ Pollut
2013;178:229e236.
56. Maddocks M, Kon SS, Canavan JL, et al. Physical frailty and pulmonary reha-
bilitation in COPD: A prospective cohort study. Thorax 2016;71:988e995.
57. Richardson EA, Mitchell R. Gender differences in relationships between urban
green space and health in the United Kingdom. Soc Sci Med 2010;71:568e575.
58. van den Bosch MA, Ostergren PO, Grahn P, et al. Moving to serene nature may
prevent poor mental healthdresults from a Swedish longitudinal cohort study.
Int J Environ Res Public Health 2015;12:7974e7989.
59. Planning Department. Hong Kong in the Global Context. HKSAR; 2016.
R. Yu et al. / JAMDA 19 (2018) 528e534534
Downloaded for Anonymous User (n/a) at Chinese University of Hong Kong from ClinicalKey.com by Elsevier on January 04, 2019.
For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.
