Você está na página 1de 14

Complementary Therapies in Medicine (2014) 22, 930—943

Available online at www.sciencedirect.com


journal homepage: www.elsevierhealth.com/journals/ctim

Effectiveness of horticultural therapy:

A systematic review of randomized
controlled trials
Hiroharu Kamioka a,∗, Kiichiro Tsutani b, Minoru Yamada c,
Hyuntae Park d, Hiroyasu Okuizumi e, Takuya Honda f,
Shinpei Okada g, Sang-Jun Park g, Jun Kitayuguchi h,
Takafumi Abe h, Shuichi Handa e, Yoshiteru Mutoh i

Faculty of Regional Environment Science, Tokyo University of Agriculture, Tokyo 156-0072, Japan
Department of Drug Policy and Management, Graduate School of Pharmaceutical Sciences,
The University of Tokyo, Tokyo 113-0033, Japan
Human Health Sciences, Grauate School of Comprehensive Human Sciences, University of Tsukba,
Kyoto 606-8507, Japan
Department of Functioning Activation, National Center for Geriatrics and Gerontology, Aichi 474-8511,
Mimaki Onsen (Spa) Clinic, Nagano 389-0402, Japan
Japan Society for Promotion of Science, Kyoto 606-8507, Japan
Physical Education and Medicine Research Foundation, Faculty of Medical Bioechnology, Dong-A
University, Nagano 389-0402, Japan
Physical Education and Medicine Research Center Unnan, Shimane 690-2404, Japan
The Research Institute of Nippon Sport Science University, Tokyo 158-0088, Japan
Available online 1 September 2014

Aim: To summarize the evidence from randomized controlled trials (RCTs) on the effects of
Horticultural therapy;
horticultural therapy (HT).
Methods: Studies were eligible if they were RCTs. Studies included one treatment group in
which HT was applied. We searched the following databases from 1990 up to August 20, 2013:
Garden therapy;
MEDLINE via PubMed, CINAHL, Web of Science, Ichushi-Web, GHL, WPRIM, and PsyclNFO. We
also searched all Cochrane Database and Campbell Systematic Reviews up to September 20,
controlled trial;
Rehabilitation effect
Results: Four studies met all inclusion criteria. The language of all eligible publications was
English and Korean. Target diseases and/or symptoms were dementia, severe mental illness
such as schizophrenia, bipolar disorder, and major depression, frail elderly in nursing home,
and hemiplegic patients after stroke. These studies showed significant effectiveness in one
or more outcomes for mental health and behavior. However, our review especially detected

Corresponding author at: Faculty of Regional Environment Science, Tokyo University of Agriculture, 1-1-1 Sakuragaoka, Setagaya-ku,
Tokyo 156-8502, Japan. Tel.: +81 3 54772587; fax: +81 3 54772587.
E-mail address: h1kamiok@nodai.ac.jp (H. Kamioka).

0965-2299/© 2014 Elsevier Ltd. All rights reserved.
A systematic review of randomized controlled trials 931

omissions of the following descriptions: method used to generate randomization, concealment,

blinding, and intention-to-treat analysis. In addition, the results of this study suggested that the
RCTs conducted have been of relatively low quality.
Conclusion: Although there was insufficient evidence in the studies of HT due to poor method-
ological and reporting quality and heterogeneity, HT may be an effective treatment for mental
and behavioral disorders such as dementia, schizophrenia, depression, and terminal-care for
© 2014 Elsevier Ltd. All rights reserved.

Introduction Methods
Over the years, recreation activity and relaxation in a forest Criteria for considering studies included in this
environment, called ‘‘forest therapy’’ or ‘‘Shinrin-yoku’’
(e.g., forest-air bathing, and forest-landscape watching and
walking) have become a kind of climatherapy or nature
Types of studies
therapy, and are popular methods for many urban peo-
Studies were eligible if they were RCTs.
ple with mental stress conditions.1 The fields of preventive
and alternative medicine have also shown an interest
in the therapeutic effects of forest therapy.2 The green Types of participants
landscape may help one recover from stress by causing There was no restriction on patients.
lowered blood pressure, higher alpha brain wave ampli-
tude, and reduced muscle tension.3 In addition, a recent Types of intervention and language
study reported that forest bathing trips increased nat- Studies included at least one treatment group in which
ural killer (NK) cell activity, which was mediated by HT was applied. The definition of the HT was complex,
increases in the number of NK cells and the levels of and ‘therapeutic horticulture’ was also considered to have
intracellular anti-cancer proteins and phytoncides released the same meaning. The American Horticultural Therapy
from trees. The decreased production of stress hormones Association defined HT as the engagement of a person
may also partially contribute to the increased NK cell in gardening-related activities, facilitated by a trained
activity.4 therapist, to achieve a specific treatment goal.9 On the
Among nature therapies, horticultural therapy (HT), other hand, Gonzalez defined therapeutic horticulture as an
which is easily implemented, has been very popular for open program, ‘‘a process that uses plant-related activities
treatment and rehabilitation for patients and for positive through which participants strive to improve there well-
health enhancement of elderly people. The fields of preven- being through active and passive involvement’’, which can
tive and alternative medicine have also shown an interest in be easy implemented and performed by a variety of health-
the therapeutic effects of HT. care providers.10 We focused on all cure and rehabilitation
A study reported that an intervention in a hospital effects in accordance with the ICD-10 and attached impor-
horticultural garden was expected to influence healing, tance to feasibility and external validity. Therefore, in this
alleviate stress, increase feelings of well-being and pro- study, we adopted the later as HT. There was no restriction
mote participation in social life and re-employment for on the basis of language.
patients with brain damage.5 A non-randomized controlled
cohort trial reported that the addition of HT to a pain
Types of outcome measures
management program improved participants’ physical and
We focused on all cure and rehabilitation effects in
mental health and their coping ability with respect to
accordance with the ICD-10. There was no restriction on
chronic musculoskeletal pain.6 A systematic review of con-
secondary outcomes.
trolled and observational studies for nature-assisted therapy
reported that significant improvements were found for var-
ied outcomes in diverse diagnoses, spanning from obesity to Search methods for studies identification
It is well known in research design that evidence grading Bibliographic database
is highest for a systematic review (SR) with meta-analysis of We searched the following databases from 1990 up to
randomized controlled trials (RCTs). Although several stud- August 20, 2013: MEDLINE via PubMed, CINAHL, Web of
ies have reported the treatment and rehabilitation effects of Science, Ichushi Web (in Japanese), the Global Health
HT, there is no SR of the evidence based on RCTs. We checked Library (GHL), the Western Pacific Region Index Medicus
the Cochrane Review protocol HT for schizophrenia,8 but it (WPRIM), and PsyclNFO. The International Committee of
is not published at present. We focused on all treatment and Medical Journal Editors (ICMJE) recommended uniform
rehabilitation effects in accordance with the International requirements for manuscripts submitted to biomedical
Classification of Diseases-10 (ICD-10). The objective of this journals in 1993. We selected articles published (that
review was to summarize the evidence from RCTs on the included a protocol) since 1990, because it appeared
effects of HT. that the ICMJE recommendation had been adopted by the
932 H. Kamioka et al.

relevant researchers and had strengthened the quality of Quality assessment of included studies
reports. In order to ensure that variation was not caused by system-
We also searched the Cochrane Database of Systematic atic errors in the study design or execution, eight review
Reviews (Cochrane Reviews), the Database of Abstracts of authors (HP, SO, HO, SH, SP, JK, TA, and TH) independently
Reviews of Effects (DARE), the Cochrane Central Register assessed the quality of articles. A full quality appraisal of
of Controlled Trials (CENTRAL), the Cochrane Methodology these papers was made using the Cochrane’s criteria list
Register (Methods Studies), the Health Technology Assess- for the methodological quality assessment.11 Disagreements
ment Database (Technology Assessments), the NHS Economic and uncertainties were resolved by discussion with other
Evaluation Database (NHS EED), The Cochrane Collabora- authors (e.g., HO, SO, and HK).
tion databases (Cochrane Groups), the Campbell Systematic Each item was scored as ‘‘yes’’ (y), ‘‘no’’ (n), ‘‘do
Reviews (the Campbell Collaboration), and all Cochrane up not know or unclear’’ (?), or ‘‘not applicable’’ (n/a).
to August 20, 2013. Depending on the study design, some items were not
All searches were performed by a specific searcher (hos- applicable. The ‘‘n/a’’ was excluded from calculation for
pital librarian) who was qualified in medical information quality assessment. We displayed the percentage of present
handling, and who was sophisticated in searches of clinical description on all 11 check items for the quality assess-
trials. ment of articles. Then, based on the percentage of risk of
poor methodology and/or bias, each item was assigned to
Search strategies the following categories: good description (80—100%), poor
The special search strategies contained the elements and description (50—79%), or very poor description (0—49%).
terms for MEDLINE, CINAHL, Web of Science, Ichushi Inter-rater reliability was calculated on a dichotomous scale
Web, GHL, WPRIM, PsyclNFO, and All Cochrane databases using percentage agreement and Cohen’s kappa coefficient
(Table 1). Only keywords relating to the above interven- (k).
tion were used for the searches. First, titles and abstracts
of identified published articles were reviewed in order to Summary of studies and data extraction
determine the relevance of the articles. Next, references in Eight review authors (HP, SO, HO, SH, SP, JK, TA, and TH)
relevant studies and identified RCTs were screened. described the summary from each article based on the rec-
ommended structured abstracts.12,13
Registry checking
We searched the International Clinical Trials Registry Plat- Benefit, harm, and withdrawals
form (ICTRP), the International Prospective Register of The GRADE Working Group14 reported that the balance
Systematic Review (PROSPERO), Clinical Trials.gov, and the between benefit and harm, quality of evidence, appli-
University Hospital Medical Information Network-Clinical cability, and the certainty of the baseline risk were all
Trials Registry (UMIN-CTR) up to August 20, 2013. considered in judgments about the strength of recom-
ICTRP in the WHO Registry Network meets specific crite- mendations. Adverse events, withdrawals, and cost for
ria for content, quality and validity, accessibility, unique intervention were especially important information for
identification, technical capacity and administration. Pri- researchers and users of clinical practice guidelines, and
mary registries meet the requirements of the ICMJE. Clinical we presented this information with the description of each
Trials.gov is a registry of federally and privately supported article.
clinical trials conducted in the United States (US) and around
the world. UMIN-CTR registers clinical trials conducted in Analysis
Japan and around the world. Pre-planned stratified analyses were: (a) trials comparing
HT with no treatment or waiting list controls, (b) trials
comparing different types of general method (e.g., physi-
Handsearching, reference checking, and other
cal therapy, occupational therapy), and (c) trials comparing
We handsearched abstracts published on HT and relevant
HT with other different intervention(s) (e.g., music ther-
journals in Japan. We checked the references of included
apy and animal-assisted therapy). We planned to express the
studies for further relevant literature.
results of each RCT, when possible, as relative risk (RR) with
corresponding 95 percent confidence intervals (95%CI) for
Review methods dichotomous data, and as standardized or weighted mean
differences (SMD) with 95%CI for continuous data. Hetero-
Selection of trials geneous results of studies that provided by inclusion criteria
In order to make the final selection of studies for the were not combined. All analyses were computed with the ‘‘R
review, all criteria were applied independently by four version 2.15.1’’, a free software environment for statistical
review authors (e.g., TH, JK, SP, TA) to the full text of arti- computing and graphics (URL:http://www.r-project.org/).
cles that had passed the first eligibility screening (Fig. 1). It compiles and runs on a wide variety of UNIX platforms,
Disagreements and uncertainties were resolved by discus- and Window.
sion with other authors (e.g., HK, KT, YM).
Studies were selected when (i) the design was a RCT Research protocol registration
and (ii) one of the interventions was a form of HT. Trials We submitted and registered our research protocol to
that were excluded are presented with reasons for exclusion the PROSPERO database (no. CRD42013005340),15 an
(Appendix). international database of prospectively registered SRs
A systematic review of randomized controlled trials 933

Table 1 The special search strategies.

in health and social care. Key features from the review the protocol. PROSPERO is managed by CRD and funded
protocol are recorded and maintained as a permanent by the UK National Institute for Health Research (NIHR).
record in PROSPERO. This provided a comprehensive listing Registration was recommended because it encourage full
of SRs registered at inception, and enabled comparison publication of the review’s findings and transparency in
of reported review findings with what was planned in changes to methods that could bias findings.16
934 H. Kamioka et al.

Manuscripts based on databases Clinical registries and systematic review registry

MEDLINE (n=137)
CINAHL (n=36) International Clinical Trials Registry Platform (n=12 )
Web of Science (n=97) Clinical Trials.gov (n=21)
Ichushi-Web (n=42) University Hospital Medical Information Network-Clinical Trials
Global Health Library (n=27) Registry (n=4)
Western Pacific Region Index Medicus (n=6) International Prospective Register of Systematic Review (n=5)
PsycINFO (n=74)
Cochrane Database of Systematic Reviews (n=5)
Database of Abstracts of Reviews of Effects (n=0)
Cochrane Central Register of Controlled Trials (n=77)
Cochrane Methodology Register (n=1)
Health Technology Assessments Database (n=0)
NHS Economic Evaluation Database (n=0)
Cochrane Collaboration databases (n=2)
Campbell Systematic Reviews (n= 6)

Retrieved for detail evaluation by articles (n=11)

Excluded (n=7) Appendix (references to studies excluded)

Not randomized controlled trial (n=5)
Not curative effect (n=1)
Not horticultural therapy (n=1 )

Articles meeting inclusion criteria (n=4)

Figure 1 Flowchart of trial process.

Results (G31)’’.20 Because there were a variety of target diseases,

there was one article for which we could not identify a single
Study selection disease.19
Jarrott et al. reported that HTs were viable and desirable
The literature searches based on databases included poten- choices for dementia-care programs because they success-
tially relevant articles (Fig. 1). Abstracts from those articles fully engaged groups of participants who were often difficult
were assessed, and 11 papers were retrieved for further to engage in activities that elicit high levels of adaptive
evaluation (checks for relevant literature). Seven publica- behavior.17
tions were excluded because they did not meet the eligibility Kam et al. reported that the horticultural activity pro-
criteria (see Appendix). Four studies17—20 met all inclusion gram was effective in reducing stress of persons with
criteria (Table 1). psychiatric illness, but did not have a significant impact on
work behavior and quality of life.18
Tse reported that an indoor gardening program for older
Study characteristics people living in nursing homes improved life satisfaction and
social network, and decreased perception of loneliness.19
The language of all eligible publications was English17—19 Kim et al. reported that HT had effects on the improve-
and Korean.20 Target diseases and/or symptoms (Table 2) ment of mental health such as self-esteem, powerlessness,
were dementia,17 severe mental illness such as schizophre- depression, and perceptual function for hemiplegic patients
nia, bipolar disorder, and major depression,18 frail elderly after stroke.20
in nursing home,19 and hemiplegic patients after stroke.20
Based on ICD-10, we identified a disease targeted in each Quality assessment
article (Table 3). Among four studies, two studies were about
‘‘Mental and behavioral disorders (F00-01,17 F20,30-3318 )’’. We evaluated 11 items from the Cochrane’s criteria list in
There was one study in ‘‘Disease of the nervous system more detail (Table 4). Inter-rater reliability metrics for the
A systematic review of randomized controlled trials
Table 2 Brief summary of articles based on structured abstracts and additional elements.

Reference no. 17 18 19 20

Author Jarrott SE, et al. Kam MCY, et al. Tse MMY. Kim SY, et al.
Citation Am J Alzheimers Dis Other Demen Hong Kong J Occup Therapy 2010; J Clin Nurs 2010; 19: 949—958 J Korean Soc Hortic Sci 2003;
2010; 25: 657—65. 20: 80—86. 44: 780—785
Title Comparing responses to Evaluation of a horticultural Therapeutic effects of an Effect of horticultural therapy
horticultural-based and traditional activity programme for persons indoor gardening programme on functional rehabilitation in
activities in dementia care with psychiatric illness for older people living in hemiplegic patients after
programmes. nursing homes stroke
Aim/objective To identify whether participants in To investigate the effect of To explore the activities of To evaluate the influence of
horticultural-based activities applying horticulture activity on daily living and psychological horticultural therapy (HT)
(HT-based activities) exhibit stress, work performance and well-being of older people activities on the improvement
higher levels of adaptive behavior quality of life in persons with living in nursing homes and also in rehabilitation of individuals
and experience more positive psychiatric illness. to examine the effectiveness with stroke patients
affect than participants in of a gardening programme in
traditional activities (TA) during enhancing socialisation and life
the presented activities. satisfaction, reducing,
loneliness and promoting
activities of daily living for
older people living in nursing
Setting/place 8 care programmes (facilities) in All the horticulture activities took All gardening activities were National rehabilitation center
rural south west Virginia place in the five outdoor theme carried out in the multiple in Korea
gardens of the Farm, namely the function room of the nursing
Sensory garden. Activity garden, homes in Hong Kong.
Farm garden, Display garden and
Practical garden in Hong Kong.
Participants 129 persons with a diagnosis of Twenty-four participants with Fifty-three participants in Forty-two individuals among
dementia severe mental illness such as nursing homes were recruited poststroke hemiplegic patients
schizophrenia spectrum disorder to participate in this study.
(schizophrenia, schizoaffective
disorder, schizophreniform
disorder, psychosis not otherwise
specified), bipolar disorder, or
major depression were recruited
to participate in this study.

Table 2 (Continued)

Reference no. 17 18 19 20

Intervention HT-based activities were The horticulture activity Older people from the Treatment group attended a
implemented twice weekly at 4 programme was a standardized experimental groups were 1 h horticultural therapy
treatment sites for 6 weeks, while group that was conducted in 10 invited to join the eight week programme 5 times per week
regular TA were observed at consecutive days within 2 weeks. indoor gardening programme, for 6 weeks. The programme
comparison sites. Two facilitators The group objectives are: (a) to while older people in the other carried out during horticultural
developed and facilitated teach basic horticultural two nursing homes were therapy included various
HT-based activities at the knowledge and skills, and to treated as the control groups: activities such as seeding,
treatment sites. Activities were develop interests in working with they received regular care and growing, and utilization of
selected for anticipated plants, (b) to share relaxing visits each week, but not the postharvest products to make
therapeutic benefits within social, experience and coping strategies eight week indoor gardening them to experience a whole
physical, psychological, and through working with plants, (c) programme. Posttest measures life cycle of plants. For
cognitive domains. In addition, to promote sharing and social were collected on the evaluating the effect of
activities were simple, support among participations. completion of the eight week programme several scales
cost-effective, and versatile. Participants in the experimental indoor gardening programme including self-esteem scale,
Plant materials were researched group attended a 1-h horticultural for all participants and powerlessness scale, beck
to ensure safety. Group sizes at activity session for 10 days. Once participant of the experience depression inventory,
HT-based activities ranged from 4 they finished the 1-h horticulture of the experimental groups neurobehavioral cognitive
to 20 participants; the average session, they joined the regular were invited for an interview status examination, motor-free
number of participants varied at sheltered workshop training. regarding their experience of visual perception test, and
each site. When group size Participants in the central group the programme. Participants in functional independence
exceeded 8 participants, the remained in the conventional the experimental group were measure were applied.
facilitators divided participants workshop training throughout the invited to attend an interview Measurements were conducted
into 2 groups to provide study period. Based on the work conducted by the researcher at at 2 times periods; before and
step-by-step instructions, physical of previous studies, each session the end of the eight week after 6 weeks horticultural
and verbal assistance, and a of the standardized horticultural indoor gardening programme. therapy programme.
steady supply of materials to each programme have a specific theme The interviews were conducted
participant. Participation was and objectives, and all main in the function rooms of the
always voluntary and elders were activities were related to work nursing homes. Interview
informed that the focus of the with living plants. In each session, lasted from 10 to 15 min. To
activities would be gardening. there is an introduction and created before proceeding to
Horticultural therapy-based warm-up, followed by horticulture ask the participants about the
activities ranged from sowing activities and group sharing. gardening experience.
seeds and training topiaries to

H. Kamioka et al.
craft activities that incorporated
horticultural materials or themes.
A systematic review of randomized controlled trials
Main and Affect (the Apparent Affect Rating The Chinese version of Depression The Life Satisfaction Index-A Self-steem scale,
secondary Scale) and engagement (Menorah Anxiety Stress Scale 21 (DASS21), Form, The Revised UCLA powerlessness, beck
outcomes Park Engagement Scale) Work Behavior Assessment (WBA), Loneliness Scale, and The depression inventory (BDI),
and Chinese version Personal Lubben Social Network Scale. neurobehavioral cognitive
Wellbeing Index (PW!-C). status examination (NCSE),
motor-free visual perception
test (MVPT), and functional
independence measure (FIM).
Randomization No description They were assigned a number and No description No description
randomly allocated by an
independent research assistant
(voluntary helper) who was
blinded to the hypothesis and the
intervention programme.
Blinding/masking No description A blinded independent assessor No description No description
(voluntary helper) conducted all
the outcome measurements
before and after the intervention
Numbers HT-based activities group (n = 75) Experimental group (n = 12) and Experimental group (n = 26) Treatment group (n = 21) and
randomized and control group (n = 54). These control group (n = 12) and control group (n = 27). control group (n = 21)
were allocated each facility unit. They were allocated each
facility unit.
Recruitment 8 care programmes (facilities) in Twenty-four participants with An organization operating Forty-two individuals with
rural southwest Virginia severe mental illness such as resident care homes for older poststroke hemiplegic patients
schizophrenia spectrum disorder people in Hong Kong was (right or left hemiplegia,
(schizophrenia, schizoaffective approached and invited to cerebral infarct, hemorrhage)
disorder, schizophreniform participate in the study. participated in patient stroke
disorder, psychosis not otherwise rehabilitation programme of
specified), bipolar disorder, or national rehabilitation center
major depression were recruited were recruited to participate
to participate in this study. in this study.
Numbers HT-based activities group (n = 75) Experiment group (n = 10) and Experimental group (n = 26) Treatment group (n = 21) and
analyzed and control group (n = 54). control group (n = 12) and control group (n = 27). control group (n = 21)
They were allocated each
facility unit.

Table 2 (Continued)

Reference no. 17 18 19 20

Outcome Results revealed no differences There was a significant difference There were significant In treatment group, the
between groups on affective in change scores of the DASS21 improvements in life self-esteem scale and the
domains. Levels of adaptive (p = 0.01) between experimental satisfaction and social network powerlessness were
behavior differed between the and control group. There were no and a significant decrease in significantly improved, and
groups, with the treatment group significant differences in change perception of loneliness for beck depression inventory
demonstrating higher levels of scores of the PWI-C between the older people in the score, motor-free visual
active, passive, and other two groups. experimental group after the perception test raw score, and
engagement and the comparison eight week indoor gardening functional independence
group demonstrating higher levels programme, while the measure were also significantly
of self-engagement. activities of daily living were decreased. Among the items of
unchanged for both groups neurobehavioral cognitive
after programme. status examination, similarity
score was significantly
Harm Nothing No description No description No description
Conclusion Our results highlight the value of Horticultural therapy is effective The present study Horticultural therapy was very
HT-based programmes and the in decreasing the levels of anxiety, demonstrates the therapeutic effective as one of the
importance of simultaneously depression and stress among effects of an indoor gardening alternative medicines and
capturing participants’ affective participants in this pilot study, but programme. Findings provide played an important role in
and behavioral responses. the impact of the programme on evidence for the inclusion of an rehabilitation of the stroke
work behavior and quality of life indoor gardening programme in patients, especially in
will need further exploration. nursing home care practice neuropsychological aspects.
with older people.
Trial No description No description No description No description
Fund Alzheimer’s and Related Diseases No description The Hong Kong Polytechnic No description
Research Award Fund from the University supported the study.
Virginia Center on Aging (VCOA),
Fund # 04-3.
Withdrawals Nothing There were two drop-outs in the Nothing Nothing
experiment group. These reasons
were not described.

H. Kamioka et al.
Cost of No description No description No description No description
A systematic review of randomized controlled trials 939

Table 3 International Classification of target diseases in each article.

Chapter ICD code Classification Reference no. (Detail ICD code)

1 A00-B99 Certain infectious and parasitic diseases

2 C00-D48 Neoplasms
3 D50-D89 Diseases of the blood and blood-forming
organs and certain disorders involving the
immune mechanism
4 E00-E90 Endocrine, nutritional and metabolic
5 F00-F99 Mental and behavioral disorders 17(F00-01), 18(F20,30-33)
6 G00-G99 Diseases of the nervous system 20(G81)
7 H00-H59 Diseases of the eye and adnexa
8 H60-H95 Diseases of the ear and mastoid process
9 I00-I99 Diseases of the circulatory system
10 J00-J99 Diseases of the respiratory system
11 K00-K93 Diseases of the digestive system
12 L00-L99 Diseases of the skin and subcutaneous tissue
13 M00-M99 Diseases of the musculoskeletal system and
connective tissue
14 N00-N99 Diseases of the genitourinary system
15 O00-O99 Pregnancy, childbirth and the puerperium
16 P00-P96 Certain conditions originating in the
perinatal period
17 Q00-Q99 Congenital malformations, deformations
and chromosomal abnormalities
18 R00-R99 Symptoms, signs and abnormal clinical and
laboratory finding not elsewhere classified
19 S00-T98 Injury, positioning and certain other
consequences of external causes
20 V00-Y98 External causes of morbidity and mortality
21 Z00-Z99 Factors influencing health status and
contact with health services
22 U00-U99 Code for special purpose
— Unidentified Because many illnesses were mixed, we 19
could not identify it.

quality assessment indicated substantial agreement for all Withdrawals and adverse events
44 items (percentage agreement 86% and k = 0.794).
This assessment evaluated the quality of how the main Three studies18—20 did not describe adverse events or harm
findings of the study were summarized in the written report. (Table 2). Three studies17,19,20 reported no withdrawals
There was a remarkable lack of execution and/or description (dropouts), and one study showed two dropouts, but reasons
in the randomization, concealment, blinding, and intention- were not described.
to-treatment (ITT) analysis, in general. The items for which
the description was lacking (very poor; <50%) in many studies
Costs of intervention
were as follows: ‘‘Was the method of randomization ade-
quate?’’ (25%); ‘‘Was the treatment allocation concealed?’’
All studies17—20 provided no information about the costs of
(25%); ‘‘Was the patient blinded to the intervention?’’ (0%);
intervention (Table 2).
‘‘Was the care provider blinded to the intervention?’’ (0%);
‘‘Was the outcome assessor blinded to the intervention?’’
(25%); and ‘‘Did the analysis include an ITT analysis?’’ Trial registration
There were no descriptions about this information in all
We could not perform meta-analysis because of hetero-
geneity of all outcome measurements and intervention This is the first SR of the effectiveness of HT based on RCTs.
methods. Our study is unique because it summarized the evidence for
940 H. Kamioka et al.

Table 4 Evaluation of the quality of methodology for each article.

No Criteria list 17 18 19 20 Present description**

No/4 Rate (%)

1 Was the method of randomization ? y ? n 1 25%

2 Was the treatment allocation concealed? ? y ? n 1 25%
3 Were the groups similar at baseline y y y y 4 100%
regarding the most important prognostic
4 Was the patient blinded to the intervention? n n ? n 0 0%
5 Was the care provider blinded to the n n n n 0 0%
6 Was the outcome assessor blinded to the n y n n 1 25%
7 Were cointerventions avoided or similar? y y y y 4 100%
8 Was the compliance acceptable in all group? ? y y n 2 50%
9 Was the drop-out rate described and n y y n 2 50%
10 Was the timing of the outcome assessment y y y y 4 100%
in all groups similar?
11 Did the analysis include an n n y n 1 25%
intention-to-treat analysis?
Present description no/11 3 8 6 3 —
Rate (%) 27% 73% 55% 27%
Yes: y, no: n, do not know or unclear: ?, not applicable: n/a.

each target disease according to ICD-10 classification. We congruent with previous studies that used other research
assume that this study will be helpful to researchers who designs.21,22
want to understand the effects of HT comprehensively, and Horticultural environment may play a significant role in
it could provide indispensable information for the organi- the improvement of mental health in patients of various
zation that is going to make guidelines according to each types, but there is some suggestion that the influence of
disease. facilitators is very important for outcome.17,23,24 An anal-
Among four RCTs that were identified, target diseases ysis of facilitators’ interaction styles and their knowledge
and/or symptoms included dementia, severe mental illness of dementia-care practices would also supplement the evi-
such as schizophrenia, bipolar disorder, and major depres- dence and guide subsequent practice.17
sion, frail elderly in nursing home, and hemiplegic patients
after stroke.
Overall evidence and quality assessment
The Cochrane’s list is the most important tool related to the
Tendency of target disease and outcome internal validity of trials. In this SR, there were serious prob-
lems with the conduct and reporting of the target studies.
The most reported target diseases were ‘‘Mental and behav- Our review especially detected omissions of the following
ioral disorders’’,17,18 and the effect of HT on these diseases descriptions: method used to generate randomization, con-
was improved mental health (e.g., anxiety and depression) cealment, blinding, and ITT analysis. Descriptions of these
and adaptive behavior. The main reason given in these arti- items were lacking (very poor; <50%) in many studies. In
cles for improved mental health might be that participants addition, the results of this study suggested that the RCTs
could relax from work stress and felt joyful in the horticul- conducted have been of relatively low quality.
tural environment during the program.18 This trend was also In the Cochrane Review,25 the eligibility criteria for a
seen in a research study for hemiplegic patients after stroke, meta-analysis are strict, and for each article, heterogene-
with the target disease category ‘‘Disease of the nervous ity and low quality of reporting must first be excluded. We
system (G31)’’.20 The improvement on adaptive behavior could not perform a meta-analysis. Due to poor method-
for patients with dementia might also be that HT supported ological and reporting quality and heterogeneity, there was
participants’ attainment of the adaptation level, which was insufficient evidence in the studies of HT, and we are there-
characterized by optimal person-environment fit.17 Another fore unable to offer any clear conclusions about the effects
study reported an unidentified targeted disease,19 and of HT based on RCTs. Moreover, the CONSORT 201026 and
indoor HT for older people living in a nursing home improved the CONSORT for nonpharmacological trials checklists27 are
mental-psychosocial outcomes. These common results were relatively new, but it was shown that the study protocol
A systematic review of randomized controlled trials 941

Table 5 Overall evidence and future research agenda to build evidence of horticulture therapy.

Overall evidence in the present Research agenda

Due to poor methodological and reporting quality 1. Satisfactory description and methodology
and heterogeneity, although there was including the CONSORT 2010 and the CONSORT for
insufficient evidence in the studies of HT, it nonpharmacological trials
may be an effective treatment for mental and 2. Description of adverse effects (e.g. allergy)
behavioral disorders such as dementia, and withdrawals
schizophrenia, depression, terminal-care of 3. Description of intervention dose (if pragmatic
cancer. intervention)
4. Description of cost
5. Power analysis
6. Development of the original check-list for HT
7. Effect on symptomatic relief of PTSD* caused
by disaster or war
* Post-traumatic stress disorder (PTSD).

description and implementation for HT studies should be and vegetable consumption, physical activity, and physical
subjected to these checklists. function in cancer survivors.31
Although further accumulation of strict RCT data is nec- In addition, HT as an intervention is unique and com-
essary, HT may be effective treatment for the following pletely different than pharmacological or traditional
diseases and symptoms: mental and behavioral disorders rehabilitation methods. Therefore, it may be necessary
such as dementia, depression, and schizophrenia, and vari- to add some original items like herbal intervention,32
ous other clinical conditions. acupancture,33 traditional Chinese medicine34 and
balneotherapy35 to the CONSORT 2010 checklist as
alternative and/or complementary medicines.
Future research agenda to build evidence Wang et al.36 suggested that gardening may be an activ-
Table 5 shows the future research agenda for studies ity that promotes overall health and quality of life, physical
on the treatment effect of HT. In a recent review for strength, fitness and flexibility, cognitive ability, and social-
gardening as a mental health intervention, Clatworthy ization. Furthermore we expect new research on victims of
et al.28 also suggested that there is a need further high- the East Japan Great Earthquake Disaster and the Collapse of
quality research such as adequate outcome measures in this the Fukushima Nuclear Power Plant that occurred on March
field. Researchers should use the appropriate checklists for 11, 2011.37 A research protocol has been developed to inves-
research design, appropriate analysis methods such as power tigate whether intervention using HT provides symptomatic
analysis, and intervention methods, which would lead to relief of mild post-traumatic stress disorder (PTSD) caused
improvement in the quality of a study, and would contribute by the disaster.
to the accumulation of evidence. Researchers should also
present not only efficacy data, but also a description of any
adverse events or harmful phenomena and the reasons for
withdrawals and non-participation. Strength and limitations
As a gradual increase of intervention is necessary in cure This review had several strengths: (1) the methods and
and rehabilitation programs, it is easy to assign settings like implementation registered high on the PROSPERO database;
‘‘Stage’’ for the intervention, such as first stage and second (2) it was a comprehensive search strategy across multi-
stage. Therefore, we also expect to understand the results ple databases with no data restrictions; and (3) it involved
and detailed descriptions of ‘‘pragmatic trials’’29 as well as detailed data extraction to allow for collecting all arti-
‘‘explanatory trials’’ for the treatment effect of HT. Jarrott cles’ content into a recommended structured abstract. The
et al. emphasized that systematically assessing the dosage conduct and reporting of this review also aligned with the
of therapeutic programming needed to achieve desired out- PRISMA statement38 for transparent reporting of SRs and
comes would help program planners maximize benefit.17 meta-analyses.
Bowen et al.30 suggested that public health is moving This review also had several limitations that should be
toward the goal of implementing evidence based inter- acknowledged. Firstly, some selection criteria were common
vention. But the feasibility of possible interventions, and across studies, as described above; however, bias remained
whether comprehensive and multilevel evaluations are due to differences in eligibility for participation in each
needed to justify them, must be determined. It is at least study. Secondly, publication bias was a limitation. Although
necessary to show the cost of such interventions. Intro- there was no linguistic restriction in the eligibility crite-
duction of an interventional method must be based on its ria, we searched studies with only English and Japanese key
cost—benefit, cost—effectiveness, and cost—utility. A recent words. In addition, this review reported on a relatively small
preliminary study reported on the feasibility and acceptabil- and heterogeneous sample of studies. Moreover, we could
ity of a mentored gardening intervention, and suggested that not follow standard procedures for estimating the effects of
it may offer a novel and promising strategy to improve fruit moderating variables.
942 H. Kamioka et al.

Conclusion of included studies and data extraction. KT, MY, and YM are
the guarantors. HK, SO, HO, and SH designed the study. TH
Although there was insufficient evidence in the studies of conducted the statistical analyses. SJP, TH, and HK assessed
HT due to poor methodological and reporting quality and the quality of articles. All authors critically revised the
heterogeneity, HT may be an effective treatment for mental manuscript for important intellectual content.
and behavioral disorders such as dementia, schizophrenia,
depression, and terminal-care for cancer. Funding
To most effectively assess the potential benefits of HT, it
will be important for future research to utilize and describe Supported by Grant-in-Aid for Scientific Research (C) number
(1) RCT methodology when appropriate, (2) adverse effects 23500817 from the Ministry of Education, Culture, Sports,
and withdrawals, (3) intervention dose, (4) intervention Science and Technology (MEXT), Japan, 2012.
cost, and (5) development of the original check-list for HT.
Conflict of interest statement
We would like to express our appreciation to Ms. Rie
None declared. Higashino and Ms. Rinako Kai (paperwork) and Ms. Satoko
Sayama (all searches of studies) for their assistance in this
Author contribution study.

HO, MY, SO, SP, TH, SH, and HK conceived the study and Appendix.
take responsibility for the quality assessment and summary

References to studies excluded in this review

Exclusion no. Author. Journal (year) Title Reason of exclusion

1 Verra ML, et al. Altern Horticultural therapy for patients Not randomized
Ther Health Med (2012) with chronic musculoskeletal controlled trial
pain: results of a pilot study
2 Annerstedt M, et al. Nature-assisted therapy: Not randomized
Scand J Public Health systematic review of controlled controlled trial
(2011) and observational studies.
3 Luk KY, et al. Int J The effect of horticultural Not randomized
Geriatr Psychiatry (2011) activities on agitation in nursing controlled trial
home residents with dementia.
4 Garland K, et al. Am J A comparison of two treatments Not horticultural therapy
Geriatr Psychiatry (2007) of agitated behavior in nursing
home residents with dementia:
simulated family presence and
preferred music
5 Verra ML, et al. Horticultural therapy for patients Not randomized
Alternative Therapies with chronic musculoskeletal controlled trial
Health Med (2012) pain: results of a pilot study
6 Rice JS, et al. J Offender Impact of horticultural therapy on Not curative effect
Rehabil (1998) psychosocial functioning among
urban jail inmates
7 Lee YA, et al. Korean J Effect of horticultural therapy Not randomized
Horticultural Sci Technol using pressed flower based upon controlled trial
(2007) logotherapy on the improvement
of the purpose in life and ego
identity of middle-aged women
A systematic review of randomized controlled trials 943

References 21. Dijkstra K, Pieterse ME, Pruyn A. Stress-reducing effects in the

built healthcare environment: the mediating role of perceived
1. Ohtsuka Y, Yabunaka N, Takayama S. Shinrin-yoku (forest-air attractiveness. Prev Med 2008;47:279—83.
bathing and walking) effectively decreases blood glucose levels 22. Wichrowski M, Whiteson J, Haas F, Mola A, Rey MJ. Effects
in diabetic patients. Int J Biometeorol 1998;41:125—7. of horticultural therapy on mood and heart rate in patients
2. Frumkim K. Beyond toxicity human health and the natural envi- participating in an inpatient cardiopulmonary rehabilitation
ronment. Am J Prev Med 2001;20:234—40. programme. J Cardiopulm Rehabil 2005;25:270—4.
3. Ulrich RS. Natural versus urban scenes: some psycho- 23. Burgio LD, Allen-Burge R, Roth DL, Bourgeois MS, Dijkstra K,
physiological effects. Environ Behav 1981;13:523—56. Gerstle J, et al. Come talk with me: improving communication
4. Li Q. Effect of forest bathing trips on human immune function. between nursing assistants and nursing home residents during
Environ Health Prev Med 2010;15:18—26. care routines. Gerontologist 2001;41:449—60.
5. Söderback I, Söderström M, Schälander E. Horticultural therapy: 24. Jarrott SE, Gigliotti CM. Layers of influence: important con-
the ‘healing garden’ and gardening in rehabilitation measures at textual factors in directing dementia care programs. J Appl
Danderyd Hospital Rehabilitation Clinic, Sweden. Pediatr Reha- Gerontol 2011;30:113—22.
bil 2005;8:235—7. 25. Higgins JPT, Green S. Cochrane handbook for systematic
6. Verra ML, Angst F, Beck T, Lehmann S, Brioschi R, Schneiter R, reviews of interventions Version 5.1.0 [updated March 2011].
et al. Horticultural therapy for patients with chronic muscu- The Cochrane Collaboration; 2011. http://www.cochrane.
loskeletal pain: results of a pilot study. Altern Ther Health Med org/training/cochrane-handbook [accessed 05.06.13].
2012;18:44—50. 26. Moher D, Hopwell S, Schulz KF, Montori V, Gotzsche PC, Dev-
7. Annerstedt M, Wahrborg P. Nature-assisted therapy: ereaux PJ, et al. CONSORT 2010 explanation and elaboration:
systematic review of controlled and observational updated guidelines for reporting parallel group randomised tri-
studies. Scand J Public Health 2011;39:371—88, als. BMJ 2010;340:c869.
http://dx.doi.org/10.1177/1403494810396400. 27. Boutron I, Moher D, Altman DG, Schulz KF, Ravaud P, CON-
8. Roberts S, Bradley AJ. Horticultural therapy for SORT group. Methods and processes of the CONSORT group:
schizophrenia. Cochrane Database Syst Rev 2011, example of an extension for trials assessing nonpharmacologic
http://dx.doi.org/10.1002/14651858.CD009413. treatments. Ann Intern Med 2008;148:W60—6.
9. American Horticultural Therapy Association. http://ahta.org/ 28. Schwartz D, Lellouch J. Explanatory and pragmatic attitudes in
sites/default/files/DefinitionsandPositions.pdf [accessed therapeutical trials. J Clin Epidemiol 2009;62:499—505.
22.08.13]. 29. Clatworthy J, Hinds J, Camic PM. Gardening as a mental health
10. Gonzalez MT, Hartig T, Patil GC, Martinsen EW, Kirkevold M. intervention: a review. Mental Health Rev J 2013;18:214—25.
Therapeutic horticulture in clinical depression: a prospective 30. Bowen DJ, Kreuter M, Spring B, Cofta-Werpel L, Linnan L,
study. Res Theory Nurs Pract 2009;23:312—28. Weiner D, et al. How we design feasibility studies. Am J Prev
11. van Tulder M, Furlan A, Bombardier C, Bouter L, The Edi- Med 2009;36:452—7.
torial Board of the Cochrane Collaboration Back Review 31. Blair CK, Madan-Swain A, Locher JL, Desmond RA, Loss
Group. Updated method guidelines for systematic reviews Santos JD, Affuso O, et al. Harvest for health gardening
in the Cochrane Collaboration Back Review Group. Spine intervention feasibility study in cancer survivors. Acta Oncol
2003;28:1290—9. 2013;52:1110—8.
12. Hopwell S, Clarke M, Moher D, Wager E, Middleton P, Altman 32. Gagnier JJ, Boon H, Rochon P, Moher D, Barnes J, Bombardier C,
DG, et al. CONSORT for reporting randomized trials in journal et al. Reporting randomized, controlled trials of herbal inter-
and conference abstracts. Lancet 2008;371:281—3. ventions: an elaborated CONSORT statement. Ann Intern Med
13. Hopewel S, Ravaud P, Baron G, Boutron I. Effect of editors’ 2006;144:364—7.
implementation of CONSORT guidelines on the reporting of 33. MacPherson H, White A, Cummings M, Jobst K, Rose K, Niemtzow
abstracts in high impact medical journals: interrupted time R. Standards for reporting interventions in controlled trials of
series analysis. BMJ 2012;344:e4178. acupuncture: the STRICTA recommendations. Complement Ther
14. Atkins D, Best D, Briss PA, Eccles M, Falck-Ytter Y, Flottorp S, Med 2001;9:246—9.
et al. Grading quality of evidence and strength of recommen- 34. Wu T, Shang H, Bian Z, Zhang J, Li T, Li Y, et al. Recommenda-
dations. BMJ 2004;328:1490—7. tions for reporting adverse drug reactions and adverse events of
15. International Prospective Register of Systematic Reviews; traditional Chinese medicine. J Evid-Based Med 2010;3:11—7.
August 9, 2013. http://www.crd.york.ac.uk/NIHR PROSPERO/ 35. Kamioka H, Kawamura Y, Tsutani K, Maeda M, Hayasaka Y,
display record.asp?ID=CRD42013005340 [accessed 28.08.13]. Okuizumi H, et al. A checklist to assess the quality of reports
16. Booth A, Clarke M, Ghersi D, Peticrew M, Stewart L. An on spa therapy and balneotherapy trials was developed using
international registry of systematic review protocols. Lancet the Delphi consensus method: the SPAC checklist. Complement
2011;377:108—9. Ther Med 2013;21:324—32.
17. Jarrott SE, Gigliotti CM. Comparing responses to horticultural- 36. Wang D, Macmillan T. The benefits of gardening for older
based and traditional activities in dementia care pro- adults: a systematic review of the literature. Act Adapt Aging
grams. Am J Alzheimers Dis Other Demen 2010;25:657—65, 2013;37:153—81.
http://dx.doi.org/10.1177/1533317510385810. 37. Kotozaki Y. Intervention study of psychotherapy on elderly
18. Kam MCY, Siu AMH. Evaluation of a horticultural activity pro- women who live in disaster area. University hospital Medical
gramme for persons with psychiatric illness. Hong Kong J Occup Information Network-Clinical Trial Registration (UMIN-CTR
Therapy 2010;20:80—6. 000008936). https://upload.umin.ac.jp/cgi-open-bin/ctr/
19. Tse MM. Therapeutic effects of an indoor gardening pro- ctr.cgi?function=brows&action=brows&type=summary&
gramme for older people living in nursing homes. J Clin Nurs recptno=R000010226&language=J [accessed 05.09.13].
2010;19:949—58. 38. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, loan-
20. Kim SY, Son KC, Jung HJ, Yoo JH, Kim BS, Park SW. Effect nidis JPA, et al. The PRISMA statement for reporting systematic
of horticultural therapy on functional rehabilitation in hemi- reviews and meta-analyses of studies that evaluate health care
plegic patients after stroke. Korean J Hortic Sci Technol interventions: explanation and elaboration. Ann Intern Med
2003;44:780—5 [in Korean with English abstract]. 2009;151:W65—94.