1
Fo es he apy can p e en and ea dep ession: e idence om me a-analyses
1
2
Fo es he apy in ol es engaging in a combina ion o o es -based ac i i ies o imp o e
3
one’s heal h o wellbeing. We conduc ed an o e iew o sys ema ic e iews (SRs) and
4
me a-analyses o p ima y s udies o p o ide he mos comp ehensi e summa y o he
5
e ec o o es he apy on dep ession. We included 13 p ima y s udies ha ma ched ou
6
eligibili y c i e ia - all we e included in ou ecen SRs and we e conduc ed in he
7
Republic o Ko ea. We ca ied ou me a-analyses wi h da a ex ac ed di ec ly om hese
8
13 s udies and assessed hei isk o bias. Ou comes o in e es we e dep essi e
9
symp oms, empo a y eco e y om dep ession (i.e. emission), esponse o ea men
10
(i.e. ≥ 50% educ ion on dep essi e symp oms om baseline), adhe ence o ea men ,
11
and ad e se e ec s. Conside ing pooled es ima es om andomized con olled ials
12
wi h adul s, we ound ha compa ed o no in e en ion/usual ca e, o es he apy
13
p oduced a g ea e educ ion o dep essi e symp oms (Hedges’g = 1.18, 95% CI [0.86,
14
1.50], p < .00001). Also compa ed o no in e en ion/usual ca e, pa icipan s in he
15
o es he apy g oup we e 17 imes as likely o achie e emission (Risk Ra io = 17.02,
16
95% CI [3.40, 85.21], p = .0006) and h ee imes as likely o ha e a ≥ 50% educ ion on
17
dep essi e symp oms (Risk Ra io = 3.18, 95% CI [1.94, 5.21], p < .00001). Fo es
18
he apy, on a e age, educed dep essi e symp oms mo e han engaging in simila
19
ac i i ies in a hospi al o non- o es ed u ban a ea, o pa icipa ing in an in e en ion
20
ocused on die plus o es -based exe cise. We did no ind e idence ha adhe ence o
21
o es he apy is di e en om he adhe ence o al e na i e in e en ions and he
22
ad e se e ec s o o es he apy appea o be a e. These esul s indica e ha , ela i e o
23
many mo e con en ional al e na i es, o es he apy is a mo e e ec i e sho - e m
24
in e en ion o he p e en ion and ea men o dep ession in adul s.
25
2
Keywo ds: con ac wi h na u e; dys hymia; o es ba hing; men al heal h; mood
1
diso de ; shin in-yoku
2
3
1. In oduc ion
4
Dep ession is conside ed one o he mos impo an global heal h challenges (Cip iani e
5
al., 2018). Mo e han 300 million people wo ldwide su e om his diso de , which can
6
ha m many aspec s o li e (e.g. a ec i e ela ionships and wo k) and, in he wo s -case
7
scena io, lead o suicide (Wo ld Heal h O ganiza ion, 2017). Common symp oms o
8
dep ession a e sad mood, anxie y, insomnia, loss o i ali y, and lack o in e es in li e
9
(F ied, 2017). Dep ession is bes concep ualized in a con inuum anging om he
10
p esence o a ew/mild/ a e symp oms o he p esence o se e al/se e e/ equen
11
symp oms ha lead o a debili a ing li e condi ion (F ied, 2017; an de Leempu e al.,
12
2014). The diagnosis o a pe son as dep essed is based on speci ic c i e ia and should
13
ideally be done h ough a s uc u ed o semi-s uc u ed in e iew (No dgaa d e al.,
14
2013). Conside ing ha dep ession is unde s ood as a combina ion o symp oms (F ied,
15
2017; an de Leempu e al., 2014), in e en ions designed o p e en o ea
16
dep ession should no ocus solely on one symp om. By p e en ing dep ession, we e e
17
o a p ocess in which a non-dep essed pe son achie es a educ ion in dep essi e
18
symp oms. Such a educ ion helps his pe son o emain non-dep essed. By ea ing
19
dep ession, we e e o a p ocess in which a dep essed pe son achie es a educ ion in
20
dep essi e symp oms.
21
Common ea men s o dep ession a e psycho he apy and an idep essan s
22
(Cuijpe s, 2018). These ha e ad an ages, such as accessibili y o ea men , and ha e
23
been p o en o be e icien in amelio a ing dep essi e symp oms (Cuijpe s e al., 2020),
24
bu hey also p esen some disad an ages. Fo example, he use o an idep essan s can
25
3
ha e seconda y e ec s like gas oin es inal symp oms (e.g. nausea and dia hea), weigh
1
gain, and me abolic abno mali ies (Ca alho e al., 2016), and bo h an idep essan s and
2
psycho he apy lack e ec i eness in he sho - e m (Cuijpe s, 2018). Conside ing hese
3
disad an ages o psycho he apy and an idep essan s, esea che s ha e called o
4
complemen a y o e en al e na i e ea men s o dep ession (Lop es i, 2019;
5
Munkholm e al., 2019). One o hese al e na i es may be di ec con ac wi h na u e
6
(Lee e al., 2017; Rajoo e al., 2020; an Tulleken e al., 2018). Speci ically, p e ious
7
s udies ha e ound encou aging esul s ega ding he po en ial o o es he apy o
8
p e en and ea dep ession (e.g. Chun e al., 2017; Kim e al., 2009).
9
The human heal h bene i s o exposu e o ees and o es s abound and include
10
es o a i e capaci ies such as s ess educ ion as well as imp o emen in clinical men al
11
heal h ou comes (Wol e al., 2020). To maximize hese capaci ies, o es he apy is
12
o en ecommended as a o m o p e en i e medicine (Pa k e al., 2010). Fo es he apy,
13
also known as “shin in-yoku”(Oh e al., 2017; Pa k e al., 2010; Rajoo e al., 2020),
14
in ol es engaging in a combina ion o ac i i ies in a o es en i onmen o imp o e
15
one’s heal h o wellbeing (Han e al., 2016; Lee e al., 2017; Yu and Hsieh, 2020).
16
Fo es he apy may include o ms o o es -based exe cise bu should in ol e mo e han
17
jus physical ac i i y - ypically inco po a ing o he ac i i ies ha os e posi i e men al
18
heal h such as medi a ion, games using o es elemen s, and/o g oup ac i i ies (Bang e
19
al., 2018; Chun e al., 2017; Dje nis e al., 2019; Han e al., 2016; Lee e al., 2017;
20
Rajoo e al., 2019). In u ban en i onmen s ha inc ease s ess and hinde psychological
21
es o a ion, di ec imme sion in o es s can help people o calm down and e lec
22
(Collado e al., 2017; Kaplan, 1995; Kim e al., 2009; Ko e a e al., 2020; Maye e al.,
23
2009; Rajoo e al., 2019). The e is e idence ha e en a sho pe iod spen in a o es can
24
help people educe s ess, eco e hei a en ional capabili ies, and shi owa ds mo e
25
4
posi i e emo ions (Dje nis e al., 2019; Kaplan, 1995; Ko e a e al., 2020; Lee e al.,
1
2017; Rajoo e al., 2020; Wen e al., 2019; Yu and Hsieh, 2020). These bene i s o
2
exposu e o o es s p o ide insigh s ega ding he alue o o es en i onmen s and he
3
ele ance o conse ing and u ilizing hese se ings o imp o e human heal h (B a man
4
e al., 2019).
5
Ye , while an abundance o esea ch sugges s o es -based ac i i ies p oduce
6
posi i e heal h ou comes (Hansen e al., 2017; Pa k e al., 2010), less esea ch has
7
explo ed di ec links be ween o es he apy and dep ession (Wen e al., 2019). Fo
8
example, in Wen e al.’s (2019) s udy o he e ec s o o es ac i i ies on heal h, only
9
h ee ou o he 28 s udies included in hei analyses epo ed he e ec o o es he apy
10
on dep ession. Fu he mo e, because p e ious e idence syn hesis combined esul s om
11
dep ession measu es wi h o he cons uc s, i is di icul o disce n i o es he apy is an
12
e ec i e way o p e en ing and ea ing dep ession, speci ically (Dje nis e al., 2019;
13
Ko e a e al., 2020). I is also unknown i o es he apy is sa e (i.e. a e ad e se e ec s
14
a e?) and accep able (i.e. do people adhe e well o o es he apy?). Such knowledge
15
gaps limi he de elopmen o guidelines o p ac i ione s who migh be willing o
16
employ o es he apy o p e en and ea dep ession.
17
We he e o e conduc ed an o e iew o sys ema ic e iews (SRs) and me a-
18
analyses o answe he ollowing esea ch ques ion: Is he e su icien e idence
19
suppo ing o es he apy as an e ec i e in e en ion o p e en and ea dep ession?
20
Ou li e a u e e iew o e s ou no el con ibu ions. Fi s , o ou knowledge, his is he
21
i s e iew o ocus exclusi ely on he e ec s o o es he apy on dep ession. Second,
22
we included mo e p ima y s udies ha epo ed dep ession ou comes han p e ious SRs,
23
and used hese s udies o p o ide es ima es o he e ec o o es he apy on dep ession.
24
Thi d, in addi ion o p e ious e iews ha p ima ily ocused on he alle ia ion o
25
5
dep essi e symp oms based on s a is ical signi icance o s anda dized e ec es ima es,
1
we conside ed mo e in e p e able ou comes (Riedel e al., 2010) such as empo a y
2
eco e y om dep ession (i.e. emission), esponse o ea men (≥ 50% educ ion in
3
dep essi e symp oms om baseline), and ea men accep abili y o adhe ence (i.e.
4
d op-ou s o any eason). Fou h, di e en om p e ious SRs, we assessed he isk o
5
bias o p ima y s udies using ools ha a o he iden i ica ion o all po en ial sou ces o
6
bias (S e ne e al., 2019, 2016).
7
8
2. Me hods
9
Ou o e iew o SRs and me a-analyses o he e ec s o o es he apy on dep ession
10
was based on guidance om he la es edi ion o he Coch ane Handbook o Sys ema ic
11
Re iews o In e en ions (Higgins e al., 2019). We began he s udy by selec ing SRs
12
mos ele an o ou esea ch ques ion cen e ed on o es he apy and dep ession. We
13
loca ed SRs by es ing he u ili y o se e al da abases (e.g., MEDLINE and PsycINFO)
14
and sea ch e ms such as o es he apy, o es ba hing, and shin in-yoku (see
15
Supplemen a y File p.1). Th ough his p ocess, we ound h ee SRs ha me ou c i e ia,
16
and a ou h was la e iden i ied ia social media o esea che s. Howe e , i we we e
17
o in e p e only he esul s o hese ou SRs ( a he han o he p ima y s udies), we
18
would eme ge wi h a limi ed answe o ou esea ch ques ion. Thus, we decided o
19
analyze all eligible p ima y s udies included by hese ou SRs. We did his by
20
de eloping ou eligibili y c i e ia (Table 1) and syn hesis plan (see Supplemen a y File
21
p. 1 o 4) based on ecen guidance om me a-analysis expe s (e.g. Bende e al., 2018;
22
Higgins e al., 2019).
23
Rega ding he eligibili y c i e ia, we made impo an dis inc ions be ween o es
24
he apy, o es exposu e, o es exe cise, and o es walking. Fo es exposu e e e s o
25

6
be in a o es . Fo es exe cise in ol es doing physical exe cise in a o es , which can
1
include walking. We ope a ionalized o es he apy o include engagemen in a
2
combina ion o o es -based ac i i ies o imp o e one’s heal h o wellbeing. Thus,
3
o es -based exe cise combined wi h o he o es -based ac i i ies (e.g. medi a ion,
4
psycho he apy, g oup ac i i ies) me ou de ini ion o o es he apy. None heless, jus
5
being in a o es o exe cising in a o es was no enough o sa is y his de ini ion and
6
quali y as a o es he apy in e en ion o his s udy.
7
< Table 1 abou he e >
8
The i s au ho o he p esen s udy collec ed ele an da a om he ou SRs,
9
assessed he eligibili y o p ima y s udies (Table 1), ga he ed ele an da a on eligible
10
p ima y s udies, and assessed he isk o bias o SRs and p ima y s udies. All p ima y
11
s udies included in ou me a-analyses we e iden i ied on he ou SRs, so he i s au ho
12
sc eened p ima y s udies o eligibili y based on he in o ma ion p o ided by he ou
13
SRs. Decisions ega ding he eligibili y o p ima y s udies a ull- ex and hei isk o
14
bias assessmen we e checked by a leas one co-au ho . Mos in o ma ion desc ibing
15
p ima y s udies (e.g. sample size) was collec ed om he ou SRs and hen checked
16
wi hin he p ima y s udies (Saldanha e al., 2019). The in o ma ion used in ou me a-
17
analyses was ex ac ed di ec ly om p ima y s udies. The i s au ho checked he
18
in o ma ion om p ima y s udies a leas once a e inishing he da a ex ac ion phase.
19
The isk o bias o he ou SRs was assessed using he ROBIS ool (Whi ing e
20
al., 2016). The isk o bias o he p ima y s udies was assessed using he RoB 2 o
21
andomized con olled ials (RCTs) and c oss-o e ials (S e ne e al., 2019), and
22
ROBINS-I o non- andomized con olled ials (NRCT) (S e ne e al., 2016). These a e
23
he mos comp ehensi e ools a ailable o assess po en ial bias in SRs, RCTs, c oss-
24
o e s, and NRCTs. S udies ha assigned pa icipan s o g oups based on a andom o
25
7
quasi- andom p ocess we e classi ied as RCTs (S e ne e al., 2019). S udies ha did no
1
desc ibe he andomiza ion p ocess o assigned pa icipan s o in e en ions based on a
2
non- andom c i e ion (e.g. pa icipan s’ p e e ence) we e classi ied as NRCTs (S e ne e
3
al., 2016).
4
5
2.1. Syn hesizing Da a om P ima y S udies
6
Following ecommenda ions om he Coch ane Handbook, we ocused on
7
esul s om RCTs and analyzed hem sepa a ely om c oss-o e s ials, and NRCTs
8
(Higgins e al., 2019). Ou main ou come was he s anda dized mean di e ence (SMD)
9
be ween he pos -in e en ion dep essi e symp oms o wo in e en ion g oups.
10
Dep essi e symp oms sco es a e usually calcula ed by summing he sco e o i ems on a
11
dep ession a ing scale o an indi idual. These i ems o en co e a speci ic symp om
12
equency and, some imes, symp om in ensi y. The mean we used in ou analysis was
13
he a e age sco e o he sample g oup in an in e en ion (e.g. he pos -in e en ion
14
a e age sco e o he o es he apy g oup). When p ima y s udies me all c i e ia needed
15
o be included in a me a-analysis (see Supplemen a y File p.1 o 2), we pooled hei
16
SMDs because pooled SMDs a e mo e p ecise han es ima es o e ec om single
17
s udies (Higgins e al., 2019). Following Sawilowky (2009), we in e p e ed SMDs as:
18
e y small = 0.01, small = 0.2, medium = 0.5, la ge = 0.8, e y la ge = 1.2, and huge =
19
2.0.
20
To imp o e he in e p e abili y o he e ec o o es he apy on dep ession
21
compa ed o o he in e en ions, we also conside ed educ ions in dep essi e symp oms
22
based on dicho omous ou comes such as emission om dep ession and esponse o
23
ea men (Riedel e al., 2010). Remission e e s o a empo a y eco e y om
24
dep ession and is o en assessed as “ he numbe o pa ien s wi h a sco e o dep essi e
25
8
symp oms below a speci ic cu -off on a alida ed a ing scale” (Cuijpe s e al., 2020, p.
1
93). Response o ea men is usually egis e ed as he numbe o people who exhibi ≥
2
50% educ ion o dep essi e symp oms om baseline ollowing ea men . This
3
h eshold is app op ia e o he mos commonly used scales o egis e dep ession: he
4
Hamil on Dep ession Ra ing Scale (HDRS o HAMD), Mon gome y-Asbe g
5
Dep ession Ra ing Scale (MADRS), and he Beck Dep ession In en o y (BDI), bu
6
migh no be app op ia e o o he scales (Riedel e al., 2010). None o he p ima y
7
s udies we analyzed epo ed he numbe o people who esponded o ea men (i.e.
8
esponde s). Thus, he numbe o esponde s in s udies using one o hese h ee scales
9
was es ima ed using he o mula desc ibed by Fu ukawa e al. (2005). We do no epo
10
he numbe o esponde s o p ima y s udies ha did no use one o hese h ee scales.
11
We used d op-ou o any eason as a p oxy o ea men accep abili y o adhe ence
12
(Cip iani e al., 2018; Cuijpe s e al., 2020). Fo dicho omous ou comes, we calcula ed
13
isk a ios as hey a e easie o in e p e han odds a ios (Higgins e al., 2019). We
14
epo ed he pe cen age o educ ion in dep essi e symp oms om baseline in o es
15
he apy and compa ison g oups as a desc ip i e s a is ic (Vicke s, 2001). We desc ibe in
16
Equa ion 1 how his pe cen age was calcula ed. We also collec ed and epo ed
17
in o ma ion abou any ad e se e ec s o o es he apy ea men desc ibed by p ima y
18
s udies’ au ho s.
19
20
((Pos -in e en ion mean sco e - Baseline mean sco e)/Baseline mean sco e)*100
21
22
Equa ion 1
23
24
S a is ical analyses we e pe o med using Re Man 5.3 (“Re iew Manage
25
(Re Man) [Compu e p og am],” 2014), and igu es illus a ing he isk o bias o
26
9
p ima y s udies we e c ea ed using ob is (McGuinness, 2019). Da a a e publicly
1
a ailable a : (inse ing link when published).
2
3
3. Resul s
4
3.1 Resul s o sys ema ic e iews
5
Selec ed cha ac e is ics (e.g. esea ch ques ion, eligibili y c i e ia, sea ch s a egy, isk
6
o bias assessmen , and main esul s) o he ou SRs ha we e iewed a e desc ibed in
7
Supplemen a y File p. 6. By analyzing he eligibili y c i e ia o hese SRs, we no ed ha
8
hey we e no able o include: (1) unpublished s udies; (2) s udies in languages o he
9
han English and Ko ean; (3) s udies published a e Oc obe 2019; and, (4) wi hin
10
Ko ean s udies: s udies wi h child en o adolescen s (< 18 yea s old), wi hou a
11
compa ison g oup, o published a e 2016.
12
The ou SRs we e deemed as a high isk o bias because o limi a ions ha
13
could hinde he adequa e iden i ica ion, selec ion, da a ex ac ion, app aisal, o
14
syn hesis o ele an p ima y s udies (Table 2). Fo example, he syn hesis me hods
15
used in he ou SRs conduc ed be o e ou e iew we e limi ed. Mo eo e , no SR
16
ocused exclusi ely on he e ec o o es he apy on dep ession. I is possible,
17
howe e , o ex ac some in o ma ion abou he e ec o o es he apy on dep ession
18
because h ee o he ou SRs epo ed esul s o each p ima y s udy. Dje nis e al.’s
19
(2019) SR was he only one ha did no epo esul s o each p ima y s udy. Ins ead,
20
he au ho s p o ided, o example, a pooled es ima e o he e ec o o es ac i i ies on
21
a combina ion o psychological cons uc s. Ko e a e al.’s (2020) SR p o ided es ima es
22
o he e ec o o es he apy on dep ession, bu only o h ee o he 13 s udies
23
analyzed in ou s udy. The me a-analysis pe o med by Ko e a e al. (2020) combined
24
esul s om measu es o dep ession wi h esul s collec ed using a measu e o mood
25
16
Only one o he me a-analyses we conduc ed (Supplemen a y Fig. 1 in
1
Supplemen a y File p. 7) p oduced subs an ial s a is ical he e ogenei y, bu we did no
2
y o explain his he e ogenei y (e.g. using me a- eg ession) due o he small numbe o
3
s udies included in his me a-analysis (Higgins e al., 2019).
4
5
4. Discussion
6
. In he p esen s udy, we analyzed ou ecen SRs ha explo ed connec ions
7
be ween o es he apy and dep ession, ocusing on o es he apy’s capaci y o educe
8
dep essi e symp oms in compa ison o no in e en ion/usual ca e and ou al e na i e
9
in e en ions. Ou indings show he e ec o o es he apy on dep ession is g ea e
10
han he e ec o any al e na i e in e en ion. People in o es he apy g oups had a
11
highe educ ion in dep essi e symp oms han people in he o he g oups.
12
13
When compa ed o simila in e en ions in non- o es se ings, he bene i s o
14
o es he apy we e clea . Being in ol ed in he apeu ic ac i i ies in a o es appea s o
15
be mo e e ec i e han pa icipa ing in such ac i i ies in a hospi al o in an u ban (non-
16
o es ed) a ea. Assuming he only di e ence be ween he o es he apy g oup and he
17
o he g oups was he in e en ion se ing, hese indings sugges ha exposu e o a
18
o es en i onmen may p o ide addi ional bene i s beyond he he apeu ic ac i i ies
19
hemsel es. This is in line wi h esea ch showing he bene i s o exposu e o na u e, and
20
o es s speci ically (Wen e al., 2019; Wol e al., 2020). Fo example, Bowle e al.
21
(2010) conduc ed me a-analyses o s udies compa ing he e ec o he same ac i i y
22
conduc ed in a na u al e sus a syn he ic en i onmen . They ound ha indi iduals who
23
conduc ed he ac i i ies in na u al en i onmen s exp essed less ange , a igue, and
24
sadness han indi iduals in syn he ic en i onmen s. Ou esul s a e also in line wi h
25

17
esea ch conduc ed unde s ess educ ion heo y (Ul ich e al., 1991) and a en ion
1
es o a ion heo y (Kaplan, 1995), which ha e consis en ly shown ha exposu e o
2
na u al en i onmen s a o s s ess educ ion, mood imp o emen , and he eco e y o
3
a en ional capabili ies mo e han non-na u al en i onmen s.
4
Ou indings also e ealed ha o es he apy was mo e e ec i e han dep ession
5
ea men egimens ocused on die plus o es -based exe cise. Whe eas die , exe cise,
6
and o es exposu e may educe dep essi e symp oms, he g ea e e ec o o es
7
he apy on dep ession may be explained by he ac i i ies in which he o es he apy
8
g oups pa icipa ed (Dje nis e al., 2019; Kim e al., 2009). The dis inc ion be ween
9
o es he apy (a combina ion o ac i i ies posi i e o men al heal h) and o es exe cise
10
is impo an . Whe eas g een exe cise can yield a a ie y o posi i e heal h ou comes
11
(Bowle e al., 2010; Gladwell e al., 2013), o es he apy, which includes o he
12
ac i i ies posi i e o men al heal h, may be mo e e ec i e o p e en and ea
13
dep ession. Simila conclusions can be d awn o he compa ison be ween o es
14
he apy and walking in a o es , as he o es he apy g oup gene ally had a highe
15
educ ion in dep essi e symp oms and g ea e esponse o ea men han he o es
16
walking g oup (al hough he con idence in e al o he es ima e o he di e ence
17
be ween hese in e en ions o e laps ze o).
18
Analyzing he numbe o d op-ou s in he in e en ion g oups ac oss he s udies
19
syn hesized, we ound no e idence ha o es he apy was a less accep able ea men
20
han o he al e na i es (Supplemen a y Fig. 4 in he Supplemen a y File p. 8). We also
21
ound ha he ad e se e ec s o o es he apy may be a e.
22
23
4.1. Limi a ions
24
18
Se e al limi a ions should be conside ed when in e p e ing he esul s o ou
1
s udy. Fi s , we selec ed only ou ele an SRs. None heless, as he las SR we included
2
(Ko e a e al., 2020) did no p o ide any new p ima y s udy, i seems ha ou app oach
3
was su icien ly comp ehensi e. Also suppo ing his iew, no new eligible p ima y
4
s udy was ound in a ecen ly published SR abou he e ec o o es ac i i ies on
5
physiological and psychological ou comes (Rajoo e al., 2020).
6
Second, he sc eening p ocess o p ima y s udies and subsequen da a ex ac ion
7
was no checked by ano he e iewe . We judged ha i was no necessa y o ha e
8
ano he e iewe in ol ed in he sc eening because his p ocess was ela i ely simple.
9
Rega ding da a ex ac ion, he i s au ho checked he in o ma ion desc ibing p ima y
10
s udies and in o ma ion used in me a-analyses a leas once a e inishing he da a
11
ex ac ion phase.
12
Thi d, al hough he indings om ou me a-analyses a e encou aging and
13
unde sco e he po en ial o o es he apy o p e en and ea dep ession, we should
14
in e p e hese indings wi h some cau ion due o he limi a ions o exis ing p ima y
15
s udies. O e all, he p ima y s udies included a conside able di e si y o pa icipan s
16
(e.g. heal h and unheal hy people), in e en ions, compa ison g oups, ou come
17
measu es, and se ings. None heless, young adul s we e unde ep esen ed, as he mean
18
age/age ange o adul s was abo e 39 yea s and only one s udy included child en (and i
19
yielded inconclusi e esul s). Thus, i emains unclea i he e ec o o es he apy on
20
dep ession is highe o lowe on child en, adolescen s, and young adul s.
21
Fou h, cau ion should also be execu ed when in e p e ing ou esul s because all
22
he s udies included in ou e iew we e conduc ed in he Republic o Ko ea, as hese
23
we e he only s udies we loca ed om anywhe e in he wo ld ha e alua ed o es
24
he apy in e en ions and epo ed esul s o dep ession ou come measu es.
25
19
Conside ing ha o es ac i i ies a e popula elsewhe e in Asia (Yu and Hsieh, 2020),
1
u u e esea ch could sys ema ically sea ch o p ima y s udies conduc ed in coun ies
2
such as Japan and Taiwan. This also highligh s he need o wide geog aphic co e age
3
in esea ch abou he o es he apy e ec on dep ession.
4
Fi h, ou inabili y o inco po a e unpublished esea ch also cas s some doub
5
abou he ue e ec i eness o o es he apy, as he e is e idence ha s udies a e mo e
6
likely o be published i hey e eal s a is ically signi ican ea men e ec s (Higgins e
7
al., 2019). Al hough he e ec o o es he apy migh ha e been o e es ima ed (o
8
unde es ima ed), he publica ion o o es he apy indings is unlikely o be hea ily
9
in luenced by inancial in e es s, which a o he publica ion o posi i e esul s, as is he
10
case o an idep essan s udies (Munkholm e al., 2019). We examined he epo ed
11
unding and con lic o in e es o p ima y s udies and ound no e idence o any con lic
12
o in e es .
13
Finally, me hodological limi a ions o he p ima y s udies in ou analyses
14
inc ease unce ain y abou he ue e ec o o es he apy on dep ession. Only ou
15
RCTs and wo c oss-o e ials we e included in ou me a-analyses, and all had a high
16
isk o bias. The main limi a ions o hese s udies we e he lack o blinding, which is
17
in easible in o es he apy in e en ions because people expe ience he ea men
18
en i onmen , and he lack o a p o ocol desc ibing he analysis plan. The o me
19
inc eases he isk o de ia ions om he in ended in e en ion and can in oduce bias in
20
he assessmen o ea men - ela ed ou comes. The la e hinde s he possibili y o
21
assessing selec i e epo ing. Limi a ions we e also iden i ied in he andomiza ion
22
p ocess and in he way some s udies deal wi h missing ou come da a. Whe eas mos o
23
hese limi a ions a e ypically associa ed wi h an exagge a ion o expe imen al
24
in e en ion e ec s (in ou case, o es he apy), in some ci cums ances he e ec can be
25
20
unde es ima ed (S e ne e al., 2019, 2016). Mo eo e , a NRCT was he only sou ce o
1
e idence o he impac o o es he apy on child en’s dep essi e symp oms, and o he
2
NRCTs p o ided addi ional in o ma ion o wo compa isons (i.e. o es he apy e sus
3
no in e en ion/usual ca e, and o es he apy e sus simila ac i i ies in a hospi al).
4
Simila o RCTs and c osso e ials, he NRCTs we e no blinded and did no publish
5
an analysis plan. Besides hese limi a ions, NRCTs end o ha e a highe isk o bias
6
han RCTs due o con ounding, as he assignmen o pa icipan s o he in e en ion can
7
be ela ed o baseline a iables ha in luence he ou come (i.e. p ognos ic ac o s).
8
9
5. Fu u e esea ch oppo uni ies and conclusions
10
In summa y, u u e li e a u e e iews may ex end he e idence we ha e
11
syn hesized by sys ema ically sea ching o unpublished s udies, s udies in o he
12
languages han English and Ko ean, and s udies wi h child en and adolescen s. Fu u e
13
p ima y s udies should documen i (and wha ) ad e se e ec s occu ed in he o es
14
he apy and compa ison g oup(s) and epo emission om dep ession and esponse o
15
ea men . Fu u e esea ch could also examine whe he ce ain aspec s o o es he apy
16
(e.g., medi a ion s. g een exe cise) a e mo e bene icial han o he s when i comes o
17
ea ing dep ession. I includes assessing i o how di e en equency and du a ion o
18
o es he apy in e en ions may in luence he educ ion in dep essi e symp oms. The
19
adop ion o ele an Consolida ed S anda ds o Repo ing T ials (CONSORT) should
20
acili a e he assessmen o p ima y s udies' isk o bias and he in e p e a ion o hei
21
esul s (Mohe e al., 2010). Ou indings also highligh he need o mo e
22
me hodologically igo ous RCTs examining he e ec s o o es he apy on dep ession.
23
Despi e some limi a ions, ou e iew o SRs and p ima y s udies examining he
24
e ec i eness o o es he apy as a p e en i e measu e and ea men o dep ession
25
21
yielded he mos conclusi e e idence o da e. Compa ed o an idep essan s, simila
1
ac i i ies in a hospi al o non- o es ed u ban se ings, o e en die and o es -based
2
exe cise, o es he apy appea s o be mo e likely o p oduce ou comes like emission
3
and esponse o ea men , wi h adequa e accep abili y o adhe ence. Thus, while mo e
4
s udies a e needed, we belie e p ac i ione s should conside he use o o es he apy as
5
bo h a p e en i e measu e and a ea men o dep ession in adul s. This is in line wi h
6
he g owing suppo o he inco po a ion o he apeu ical ac i i ies in con ac wi h
7
na u e as a mains eam in e en ion o he p e en ion and ea men o men al heal h
8
p oblems (Buckley e al., 2018).
9
10
Re e ences
11
Bang, K.-S., Kim, S., Song, M., Kang, K., Jeong, Y., 2018. The e ec s o a heal h
12
p omo ion p og am using u ban o es s and nu sing s uden men o s on he
13
pe cei ed and psychological heal h o elemen a y school child en in ulne able
14
popula ions. In . J. En i on. Res. Public Heal h 15.
15
h ps://doi.o g/10.3390/ije ph15091977
16
Bende , R., F iede, T., Koch, A., Kuss, O., Schla mann, P., Schwa ze , G., Skipka, G.,
17
2018. Me hods o e idence syn hesis in he case o e y ew s udies. Res. Syn h.
18
Me hods. h ps://doi.o g/10.1002/j sm.1297
19
Bowle , D.E., Buyung-Ali, L.M., Knigh , T.M., Pullin, A.S., 2010. A sys ema ic e iew
20
o e idence o he added bene i s o heal h o exposu e o na u al en i onmen s.
21
BMC Public Heal h 10, 456. h ps://doi.o g/10.1186/1471-2458-10-456
22
B a man, G.N., Ande son, C.B., Be man, M.G., Coch an, B., de V ies, S., Flande s, J.,
23
Folke, C., F umkin, H., G oss, J.J., Ha ig, T., Kahn, P.H., Kuo, M., Lawle , J.J.,
24
Le in, P.S., Lindahl, T., Meye -Lindenbe g, A., Mi chell, R., Ouyang, Z., Roe, J.,
25

22
Sca le , L., Smi h, J.R., an den Bosch, M., Wheele , B.W., Whi e, M.P., Zheng,
1
H., Daily, G.C., 2019. Na u e and men al heal h: An ecosys em se ice
2
pe spec i e. Sci. Ad . 5, eaax0903. h ps://doi.o g/10.1126/sciad .aax0903
3
Buckley, R.C., B ough, P., Wes away, D., 2018. B inging ou doo he apies in o
4
mains eam men al heal h. F on . Public Heal. 6.
5
h ps://doi.o g/10.3389/ pubh.2018.00119
6
Ca alho, A.F., Sha ma, M.S., B unoni, A.R., Vie a, E., Fa a, G.A., 2016. The sa e y,
7
ole abili y and isks associa ed wi h he use o newe gene a ion an idep essan
8
d ugs: A c i ical e iew o he li e a u e. Psycho he . Psychosom. 85, 270–288.
9
h ps://doi.o g/10.1159/000447034
10
Choi, Y.H., Ha, Y.S., 2014. The e ec i eness o a o es -expe ience-in eg a ion
11
in e en ion o communi y dwelling cance pa ien s’ dep ession and esilience. J.
12
Ko ean Acad. Communi y Heal. Nu s. 25, 109.
13
h ps://doi.o g/10.12799/jkachn.2014.25.2.109
14
Chun, M.H., Chang, M.C., Lee, S.-J., 2017. The e ec s o o es he apy on dep ession
15
and anxie y in pa ien s wi h ch onic s oke. In . J. Neu osci. 127, 199–203.
16
h ps://doi.o g/10.3109/00207454.2016.1170015
17
Cip iani, A., Fu ukawa, T.A., Salan i, G., Chaimani, A., A kinson, L.Z., Ogawa, Y.,
18
Leuch , S., Ruhe, H.G., Tu ne , E.H., Higgins, J.P.T., Egge , M., Takeshima, N.,
19
Hayasaka, Y., Imai, H., Shinoha a, K., Tajika, A., Ioannidis, J.P.A., Geddes, J.R.,
20
2018. Compa a i e e icacy and accep abili y o 21 an idep essan d ugs o he
21
acu e ea men o adul s wi h majo dep essi e diso de : A sys ema ic e iew and
22
ne wo k me a-analysis. Lance 391, 1357–1366. h ps://doi.o g/10.1016/S0140-
23
6736(17)32802-7
24
Collado, S., S aa s, H., Co aliza, J.A., Ha ig, T., 2017. Res o a i e en i onmen s and
25
23
heal h, in: Na a o, O., Fleu y-Bahi, G., Pol, E. (Eds.), Handbook o
1
En i onmen al Psychology and Quali y o Li e Resea ch. Sp inge , New Yo k, pp.
2
127–148. h ps://doi.o g/10.1007/978-3-319-31416-7_7
3
Cuijpe s, P., 2018. The challenges o imp o ing ea men s o dep ession. JAMA 320,
4
2529–2530. h ps://doi.o g/10.1001/jama.2018.17824
5
Cuijpe s, P., Noma, H., Ka yo aki, E., Vinke s, C.H., Cip iani, A., Fu ukawa, T.A.,
6
2020. A ne wo k me a‐analysis o he e ec s o psycho he apies,
7
pha maco he apies and hei combina ion in he ea men o adul dep ession.
8
Wo ld Psychia y 19, 92–107. h ps://doi.o g/10.1002/wps.20701
9
Dje nis, Le s up, Poulsen, S igsdo e , Dahlgaa d, O’Toole, 2019. A Sys ema ic e iew
10
and me a-analysis o na u e-based mind ulness: e ec s o mo ing mind ulness
11
aining in o an ou doo na u al se ing. In . J. En i on. Res. Public Heal h 16,
12
3202. h ps://doi.o g/10.3390/ije ph16173202
13
F ied, E.I., 2017. The 52 symp oms o majo dep ession: Lack o con en o e lap
14
among se en common dep ession scales. J. A ec . Diso d. 208, 191–197.
15
h ps://doi.o g/10.1016/j.jad.2016.10.019
16
Fu ukawa, T.A., Cip iani, A., Ba bui, C., B ambilla, P., Wa anabe, N., 2005. Impu ing
17
esponse a es om means and s anda d de ia ions in me a-analyses. In . Clin.
18
Psychopha macol. 20, 49–52. h ps://doi.o g/10.1097/00004850-200501000-00010
19
Gladwell, V.F., B own, D.K., Wood, C., Sande cock, G.R., Ba on, J.L., 2013. The
20
g ea ou doo s: How a g een exe cise en i onmen can bene i all. Ex em. Physiol.
21
Med. 2, 1–7. h ps://doi.o g/10.1186/2046-7648-2-3
22
Han, J.-W., Choi, H., Jeon, Y.-H., Yoon, C.-H., Woo, J.-M., Kim, W., 2016. The
23
E ec s o Fo es The apy on Coping wi h Ch onic Widesp ead Pain: Physiological
24
and Psychological Di e ences be ween Pa icipan s in a Fo es The apy P og am
25
24
and a Con ol G oup. In . J. En i on. Res. Public Heal h 13.
1
h ps://doi.o g/10.3390/ije ph13030255
2
Hansen, M.M., Jones, R., Tocchini, K., 2017. Shin in-yoku ( o es ba hing) and na u e
3
he apy: A s a e-o - he-a e iew. In . J. En i on. Res. Public Heal h 14.
4
h ps://doi.o g/10.3390/ije ph14080851
5
Higgins, J.P.T., Thomas, J., Chandle , J., Cumps on, M., Li, T., Page, M.J., Welch, V.A.
6
(Eds.), 2019. Coch ane Handbook o Sys ema ic Re iews o In e en ions, 2nd
7
ed. John Wiley & Sons, Chiches e .
8
Hong, S.-S., Kim, H.-C., Cho, S.-H., 2013. The E ec s o Fo es s Healing o Cogni i e
9
Func ion. J. O ien . Neu opsychia y 24, 63–74.
10
h ps://doi.o g/10.7231/jon.2013.24.1.063
11
Hong, S.-S., Lee, J.-E., Kim, H.-C., Cho, S.-H., 2012. The E ec s o Fo es s Healing
12
o Hwa-Byung. J. O ien . Neu opsychia y 23, 169–182.
13
h ps://doi.o g/10.7231/jon.2012.23.4.169
14
Kaplan, S., 1995. The es o a i e bene i s o na u e: Towa d an in eg a i e amewo k.
15
J. En i on. Psychol. 15, 169–182. h ps://doi.o g/10.1016/0272-4944(95)90001-2
16
Kim, W., Lim, S.-K., Chung, E.-J., Woo, J.-M., 2009. The e ec o cogni i e beha io
17
he apy-based psycho he apy applied in a o es en i onmen on physiological
18
changes and emission o majo dep essi e diso de . Psychia y In es ig. 6, 245.
19
h ps://doi.o g/10.4306/pi.2009.6.4.245
20
Ko e a, Y., Richa dson, M., She ield, D., 2020. E ec s o Shin in-Yoku (Fo es
21
Ba hing) and Na u e The apy on Men al Heal h: A Sys ema ic Re iew and Me a-
22
Analysis.
23
Lee, I., Choi, H., Bang, K.-S., Kim, S., Song, M., Lee, B., 2017. E ec s o o es
24
he apy on dep essi e symp oms among adul s: A sys ema ic e iew. In . J.
25
25
En i on. Res. Public Heal h 14. h ps://doi.o g/10.3390/ije ph14030321
1
Lim, Y., Kim, D., Yeoun, P., 2014. Changes in dep ession deg ee and sel -es eem o
2
senio si izens in a nu sing home acco ding o o es he apy p og am. J. Ko ean
3
Ins . Fo . Rec ea . 18, 1–11. h ps://doi.o g/10.34272/ o es .2014.18.1.001
4
Lop es i, A.L., 2019. I is ime o in es iga e in eg a i e app oaches o enhance
5
ea men ou comes o dep ession? Med. Hypo heses 126, 82–94.
6
h ps://doi.o g/10.1016/j.mehy.2019.03.008
7
Maye , F.S., F an z, C.M.P., B uehlman-Senecal, E., Dolli e , K., 2009. Why is na u e
8
bene icial? The ole o connec edness o na u e. En i on. Beha . 41, 607–643.
9
h ps://doi.o g/10.1177/0013916508319745
10
McGuinness, L., 2019. ob is: An R package and web applica ion o isualising isk-
11
o -bias assessmen s.
12
Mohe , D., Hopewell, S., Schulz, K.F., Mon o i, V., Gø zsche, P.C., De e eaux, P.J.,
13
Elbou ne, D., Egge , M., Al man, D.G., 2010. CONSORT 2010 explana ion and
14
elabo a ion: Upda ed guidelines o epo ing pa allel g oup andomised ials. In .
15
J. Su g. 10, 28–55. h ps://doi.o g/h ps://doi.o g/10.1136/bmj.c869
16
Munkholm, K., Paludan-Mülle , A.S., Boesen, K., 2019. Conside ing he
17
me hodological limi a ions in he e idence base o an idep essan s o dep ession:
18
A eanalysis o a ne wo k me a-analysis. BMJ Open 9, e024886.
19
h ps://doi.o g/10.1136/bmjopen-2018-024886
20
No dgaa d, J., Sass, L.A., Pa nas, J., 2013. The psychia ic in e iew: Validi y,
21
s uc u e, and subjec i i y. Eu . A ch. Psychia y Clin. Neu osci. 263, 353–364.
22
h ps://doi.o g/10.1007/s00406-012-0366-z
23
Oh, B., Lee, K.J., Zaslawski, C., Yeung, A., Rosen hal, D., La key, L., Back, M., 2017.
24
Heal h and well-being bene i s o spending ime in o es s: sys ema ic e iew.
25
Woo 2012
Pa ien s wi h
majo
dep ession
aking
an idep essan s
45.68
----
Fo es ac i i ies included
cogni i e-beha io al he apy,
medi a ion, and elaxa ion aining.
Fou weeks
Once a week
( o 3 hou s)
28
1: (N = 21) Simila
ac i i ies in a hospi al;
2: (N = 15) T ea men
as usual;
3: (N = 17) Walking in
a o es .
HDRS-17, MADRS,
and BDI
Randomized
con olled ial
Seoul
A bo e um
T1: Jus be o e
ea men ; T2: jus
a e ea men
Bang 2018
Elemen a y-
school s uden s
in g ades 4 o 6
a i e
communi y
cen e s
11.79
55.56
Fo es ac i i ies included i e
senses expe ience, walking, and
games.
Ten weeks
Once a week
( o 60 min)
24
(N = 28) No mal daily
ou ines
Child en’s
Dep ession
In en o y
Non-
andomized
con olled ial
U ban o es
T1: Jus be o e
ea men ; T2: jus
a e ea men
Kim 2009
Pa ien s wi h
majo
dep ession
aking
an idep essan s
46.2
85.7
Fo es ac i i ies included
cogni i e-beha io al he apy,
posi i e psychology ools, and
mind ulness medi a ion on b ea h,
wind, o es , and sounds.
Fou weeks
Once a week
( h ee
hou s/session)
23
1: (N = 19) Simila
ac i i ies in a hospi al;
2: (N = 21) T ea men
as usual.
BDI, HDRS,
MADRS
Randomized
con olled ial
Hong-Reung;
44-ha
a bo e um
T1: Jus be o e
ea men ; T2: T1 +
1 week;
T3: T1 + 2 weeks;
T4: T1 + 3 weeks
No e: *When esul s o mo e han one ou come measu e we e a ailable, we ga e p e e ence o esul s om one ou come measu e based on
speci ic c i e ia (see Supplemen a y File p. 2 o mo e in o ma ion). The chosen measu es a e in bold. Unde lined s udies we e w i en in Ko ean.
BDI = Beck Dep ession In en o y; HDRS = Hamil on Dep ession Ra ing Scale; MADRS = Mon gome y-Asbe g Dep ession Ra ing Scale.

33
Table 4. Pe cen age o educ ion in dep essi e symp oms om baseline o o es he apy and o he in e en ions, lis ed by p ima y s udy
Fi s au ho and
yea
G oups
Fo es he apy
Simila ac i i ies in a
hospi al
Simila ac i i ies in an
u ban a ea
No in e en ion/usual
ca e
Die plus exe cise in he
o es
Walking in a
o es
Randomized con olled ials
Chun 2017
-77.46
NA
-1.39
NA
NA
NA
Kim 2009
-50.08
-19.91
NA
-7.57
NA
NA
Shin 2012
-64.04
NA
NA
0.20
NA
NA
Woo 2012
-50.27
-34.81
NA
-10.10
NA
-32.49
Mean (SD)
-60.46 (13.08)
-27.36 (10.54)
-1.39
-5.82 (5.37)
NA
-32.49
Median
-57.16
-27.36
-1.39
-7.57
NA
-32.49
C oss-o e ials
Hong 2013
-46.43
NA
NA
NA
-5.10
NA
Hong 2012
-56.04
NA
NA
NA
-9.91
NA
Mean (SD)
-51.24 (6.80)
NA
NA
NA
-7.51 (3.40)
NA
Median
-51.24
NA
NA
NA
-7.51
NA
Non- andomized con olled ials
Han 2016
-46.08
NA
NA
-15.85
NA
NA
Lim 2014
-27.71
-16.76
NA
2.18
NA
NA
You 2014
-27.62
NA
NA
-0.58
NA
NA
Choi 2014
-9.61
NA
NA
3.43
NA
NA
M.-H. Kim 2015
-13.03
NA
NA
6.99
NA
NA
Y. G. Kim 2015
-9.79
NA
NA
3.93
NA
NA
Mean (SD)
-22.31 (14.33)
-16.76
NA
0.02 (8.15)
NA
NA
Median
-20.33
-16.76
NA
2.81
NA
NA
Non- andomized con olled ial wi h child en
Bang 2018
-21.13
NA
NA
-8.73
NA
NA
Table 5. Agg ega ed sample size (N) and s udies p o iding da a o each compa ison
Compa ison
N
P ima y s udies’ i s au ho and yea
Fo es he apy e sus no
in e en ion/usual ca e
525
Kim 2009; Woo 2012; Shin 2012; You
2014; Lim 2014; Choi 2014; M-H. Kim
2015; Y-G. Kim 2015; Bang 2018
Fo es he apy e sus simila
ac i i ies in a hospi al
134
Kim 2009; Woo 2012; Lim 2014
Fo es he apy e sus simila
ac i i ies in an u ban a ea
59
Chun 2017
Fo es he apy e sus die plus
o es -based exe cise
31
Hong 2013; Hong 2012
Fo es he apy e sus walking in
a o es
43
Woo 2012
No e: Some o es he apy g oups we e coun ed in mo e han one compa ison. Randomized
con olled ials a e in bold and c oss-o e ials a e unde lined. The o he s udies a e non-
andomized con olled ials.
Fig. 1. Flow diag am illus a ing he selec ion p ocess o p ima y s udies.
Reco ds iden i ied h ough he ou
sys ema ic e iews (N = 101)
Reco ds excluded because hey
a e duplica es o did no
measu e dep ession (N = 82)
Full- ex a icles assessed
o eligibili y
(N = 19)
Sc eening
Included
Eligibili y
Iden i ica ion
Full- ex a icles excluded
wi h easons (N = 6)
# S udies ha in ol ed jus
exe cising in o es (N = 3);
# S udies in se ings o he han
o es (N = 2);
# We could no ha e access o
one s udy.
S udies syn hesized
(N = 13)
Reco ds sc eened
(N = 101)
Fig. 2. Resul s o andomized con olled ials: (a) Compa ison o he pos -in e en ion mean sco e o o es he apy g oups e sus no in e en ion/usual ca e
using he in e se a iance ixed-e ec me a-analysis. (b) Compa ison o he isk o empo a y eco e y om dep ession (i.e. emission) be ween o es
he apy g oups and usual ca e o dep ession g oups, using he Man el-Haenszel ixed-e ec me a-analysis. (c) Compa ison o he isk o esponse o
ea men (i.e. ≥ 50% educ ion on dep essi e symp oms) be ween o es he apy g oups and no in e en ion/usual ca e g oups, using he Man el-Haenszel
ixed-e ec me a-analysis. E en s e e o cases o emission (b) o esponse (c). G een squa es e e o s anda dized mean di e ences and blues squa es o
isk a ios. Bigge squa es indica ed mo e pa icipan s in a s udy o mo e e en s and a bigge diamond indica es g ea e unce ain y in he es ima e. C oss-
o e s and non- andomized con olled ials we e analyzed sepa a ely.
Fig. 3. Resul s o andomized con olled ials: (a) Compa ison o he pos -in e en ion mean sco e o o es he apy g oups e sus simila ac i i ies in hospi al
g oups using he in e se a iance ixed-e ec me a-analysis. (b) Compa ison o he isk o empo a y eco e y om dep ession (i.e. emission) be ween o es
he apy g oups and simila ac i i ies in hospi al g oups, using he Man el-Haenszel ixed-e ec me a-analysis. (c) Compa ison o he isk o esponse o
ea men (i.e. ≥ 50% educ ion on dep essi e symp oms) be ween o es he apy g oups and simila ac i i ies in hospi al g oups, using he Man el-Haenszel
ixed-e ec me a-analysis. E en s e e o cases o emission (b) o esponse (c). G een squa es e e o s anda dized mean di e ences and blues squa es o
isk a ios. Bigge squa es indica ed mo e pa icipan s in a s udy o mo e e en s and a bigge diamond indica es g ea e unce ain y in he es ima e. C oss-o e
and non- andomized con olled ials we e analyzed sepa a ely.