To en e, P., e al (2017). The E ec s o Relaxa ion Exe cises and Pa k Walks Du ing
Wo kplace Lunch B eaks on Physiological Reco e y.
Scandina ian Jou nal o Wo k
and O ganiza ional Psychology
, 2(1): 2, 1–15, DOI: h p://doi.o g/10.16993/sjwop.19
* Uni e si y o Tampe e, FI
† Uni e si y o S ockholm, SE
Co esponding au ho : Ulla Kinnunen ([email p o ec ed])
ORIGINAL ARTICLE
The E ec s o Relaxa ion Exe cises and Pa k Walks
Du ing Wo kplace Lunch B eaks on Physiological
Reco e y
Ped o To en e*, Ulla Kinnunen*, Ma jaana Sianoja*, Jessica de Bloom*, Kale i Ko pela*,
Ma i T. Tuomis o* and Pe a Lind o s†
Conside ing he inc easing demands o a ious occupa ional in e en ions, his s udy aimed a examining
he impac o elaxa ion exe cises and pa k walks du ing lunch b eaks on physiological eco e y (i.e.,
on changes in co isol exc e ion and blood p essu e). In a ou -week andomized con olled ial, 153
knowledge wo ke s in se en companies we e alloca ed o one o h ee g oups: elaxa ion, pa k walk, o
con ol. Bo h in e en ion g oups we e equi ed o unde ake ei he a lunch ime elaxa ion exe cise o a
pa k walk on each wo king day o wo consecu i e weeks. Da a we e collec ed a baseline, du ing he
wo-week in e en ion pe iod, and in he week a e he in e en ion. Mixed-design analyses o a iance
(ANOVA) we e conduc ed. No bene icial in e en ion e ec s we e obse ed in co isol awakening esponse
(CARi) o co isol decline du ing he day (CDD). Blood p essu e dec eased signi ican ly in he a e noon
a wo k in each g oup. This dec ease was mo e p onounced in he pa k walk g oup (
d
= .51–.58) han in
he elaxa ion (
d
= .18–.28) and con ol (
d
= .31–.41) g oups. Ou s udy showed ha changing knowledge
wo ke s’ lunch ou ines o a sho pe iod o ime does no a ec co isol exc e ion, bu may lowe blood
p essu e a he end o he wo king day. This lowe ed blood p essu e also seemed o occu among he
con ols, sugges ing ha measu ing and keeping ack o blood p essu e may se e as an in e en ion. How-
e e , longe in e en ions a e needed o achie e s onge and long las ing physiological eco e y e ec s.
Keywo ds: blood p essu e; co isol; lunch b eak; pa k walk; eco e y; elaxa ion exe cise
Job s ess is linked o a wide ange o nega i e heal h
ou comes and impai ed well-being (e.g., Chandola,
B unne & Ma mo 2006; Gans e & Rosen 2013; Ki imäki
e al. 2012). Consequen ly wo kplace in e en ions coun-
e ac ing s ess a e c ucial o main aining occupa ional
heal h. In he p esen s udy we ocused on lunch b eaks
as an in e en ion se ing. Lunch b eaks ypically ep e-
sen he longes wi hin-wo kday b eaks and a e hus likely
play an impo an ole in daily job s ess eco e y (F i z
e al. 2013; T ougakos e al. 2014). By engaging employees
in speci ic eco e y ac i i ies – ei he a elaxa ion exe cise
o a pa k walk – du ing lunch ime, we aimed o dec ease
hei s ess le els and es o e hei ene gy.
Al hough he wo kplace is no he mos ob ious eco e y
se ing, wi hin-wo kday eco e y is impo an o main ain
high le els o ene gy and engagemen h oughou he
wo king day and o p e en accumula ion o job s ess
(Geu s, Becke s & Tucke 2014). Reco e y has been iden i-
ied as an impo an mechanism o explain how employees
can s ay ene ge ic, engaged, and heal hy, e en when
acing high demands and s ess a wo k (Sonnen ag & F i z
2015). Incomple e eco e y om job s ess may ini ia e a
p ocess, which e en ually culmina es in ch onic heal h
p oblems, such as p olonged a igue, sleep diso de s, and
ca dio ascula diseases (Geu s & Sonnen ag 2006; Slui e
e al. 2001; U sin 2000). In he eco e y p ocess, physi-
ological ac o s play a signi ican ole (McEwen & Seeman
1999). F om a physiological iewpoin , eco e y can be
de ined in e ms o dec easing le els o physiological
s ess ma ke s. In he p esen s udy we e alua e whe he a
wo-week lunch ime in e en ion a ec ed he physiologi-
cal s ess ma ke s co isol and blood p essu e.
Co isol and blood p essu e as physiological
ma ke s o job s ess
The hypo halamic-pi ui a y-ad enal (HPA) axis and he
sympa he ic-ad enal-medulla y (SAM) sys em a e key
physiological sys ems in ol ed in s ess- ela ed ac i a ion
bu also in eco e ing om s ess. Co isol is one o he
end p oduc s o he HPA-axis and conside ed a key ac o
linking he expe ience o s ess o ad e se heal h e ec s
(Ch ousos 2009). Blood p essu e ela es o he SAM
sys em, and high blood p essu e is associa ed wi h highe
To en e e al: Wo kplace Lunch B eak In e en ionA . 2, page 2 o 15
ca dio ascula mo bidi y and mo ali y (see Lewing on
e al. 2002, o a e iew). E idence also unde sco es he
impo ance o psychosocial ac o s a wo k in he e iology
o ca dio ascula diseases (see Backe e al. 2012; Ki imäki
e al. 2012, o e iews). Acco ding o he allos a ic load
model (McEwen 1998), s ess ho mones (i.e., co isol) a e
p ima y media o s, while he ca dio ascula sys em (i.e.,
blood p essu e) is a seconda y media o in he s ess p o-
cess. Ch onic ac i a ion o hese sys ems is a isk ac o o
de eloping ch onic disease (Gans e & Rosen 2013).
Co isol
Co isol le els ypically ollow a diu nal hy hm peaking
du ing he i s hou a e waking – known as he co -
isol awakening esponse (CAR) – ollowed by g adual
dec eases eaching hei lowes le els a ound midnigh
(Clow e al. 2010). In his s udy we ocused on bo h he
co isol awakening esponse index (CARi) and he decline
in co isol le els om he mo ning o he e ening (CDD).
The CARi index is measu ed by calcula ing he di e -
ence be ween co isol le els on awakening and 30 o 45
minu es he ea e (P uessne e al. 1997). Me a-analy ic
indings (Chida & S ep oe 2009) sugges ha high CARi is
posi i ely bu weakly associa ed wi h a numbe o psycho-
social ac o s including job s ess and o he o ms o li e
s ess. Job s ess and gene al li e s ess we e mos com-
monly ela ed o highe CARi, whe eas a igue, bu nou ,
and exhaus ion we e ela ed o lowe CARi. This la e
e ec may ela e o a si ua ion o p olonged exposu e o
s ess, whe e he bodily sys ems may unde go a dys egula-
ion (coun e - egula o y esponse) and co isol ebounds
o lowe le els han is ypical (Fekedulegn e al. 2012). This
is o en e e ed o as hypoco isolism (Heim, Ehle &
Hellhamme 2000).
The decline o co isol le els du ing he day (CDD,
also e e ed o as a decline owa ds e ening o simply a
decline) has been less esea ched in psychophysiological
s udies han he CAR. CDD can be measu ed as he di e -
ence be ween he peak le el (measu ed 30 o 45 minu es
a e awakening) and he e ening co isol le el (meas-
u ed in he e ening be o e going o bed; e.g., Hansen,
Hogh & Pe sson 2011). A me a-analysis (Chandola,
He aclides & Ku ami 2010) o he esul s om 16 s ud-
ies examining ela ionships be ween job s ess and
pos -mo ning (gene ally a e noon and e ening) diu nal
co isol le els epo ed inconclusi e indings and called
o u he esea ch on he opic. Howe e , many s ud-
ies sugges ha high co isol le els in he e ening may
e lec an insu icien abili y o eco e (e.g., Ha is e al.
2007; McEwen 1998; U sin & E iksen 2004). In addi ion,
a sligh decline om mo ning o e ening, ha is, a la -
ened co isol p o ile, has been associa ed wi h psycho-
social ac o s such as hea y wo kload and job demands
(Caplan, Cobb & F ench 1979; Ka lson e al. 2011), high
e o - ewa d imbalance (Liao, B unne & Kuma i 2013),
bu also wi h s ess- ela ed exhaus ion (Lindebe g e al.
2008; Nicolson & an Dies 2000; Sjög en, Leande son &
K is enson 2006). In sum, despi e inconsis encies
conce ning CDD- indings, i seems ha la co isol
diu nal p o iles a e o en ela ed o s ess ul s a es and
en i onmen s (Mille , Chen & Zhou 2007).
Blood p essu e
High blood p essu e – a majo ca dio ascula disease isk
ac o – is ano he physiological indica o ha has been
linked o job s ess (see Gans e & Rosen 2013, o a e iew).
Job s ain, long wo king hou s, and high job demands
ha e equen ly been associa ed wi h highe a e age
daily blood p essu e (e.g., Ligh , Tu ne & Hinde li e
1992; Schnall e al. 1998). Despi e his, o he s udies
epo no ela ion o e en an in e se ela ion be ween
s ess and blood p essu e (see Nyklícek, Vinge hoe s &
Van Heck 1996, o a e iew). Fo example, Hassoun
e al. (2015) showed ha less pe cei ed li e s ess was
associa ed wi h lowe sys olic and dias olic blood p es-
su e le els, whe eas exposu e o objec i ely measu ed job
s esso s (i.e., physical and psycho-social s esso s) based
on s anda d occupa ional classi ica ions was un ela ed
o blood p essu e. Such disc epancies in s udy indings
a e likely o be explained by con ounding mode a o a i-
ables, such as di e ences in he measu es used o assess
job s ess.
Relaxa ion exe cise and pa k walk as eco e y
ac i i ies
I is well known ha elaxa ion in e en ions a e e ec-
i e in educing job s ess (see Richa dson & Ro hs ein
2008; an Dixhoo n & Whi e 2005, o e iews). The
mos equen ly used elaxa ion echniques o achie e
a elaxed s a e, cha ac e ized by low physiological ac i-
a ion and posi i e a ec , a e deep-b ea hing and con-
scious elease o muscle ension e e ed o as p og es-
si e muscle elaxa ion (McCallie, Blum & Hood 2006). In
addi ion, some elaxa ion exe cises include mind ulness
componen s (awa eness and accep ance), which ha e
posi i e e ec s on psychological ou comes (see Ho -
mann e al. 2010, o a e iew). Al hough e y spa se,
some empi ical indings show ha elaxa ion echniques
ha e he po en ial o dec ease co isol sec e ion and/o
blood p essu e in in e en ions conduc ed in wo kplaces
(Nyklícek e al. 2013; Richa dson & Ro hs ein 2008; Yung
e al. 2004).
Howe e , as a as we know, elaxa ion exe cises pe -
o med du ing lunch b eaks ha e been examined in only
one s udy. In a small-scale andomized con olled ial
(RCT) among 14 call-cen e agen s, K ajewski and col-
leagues (2011) showed ha p og essi e muscle elaxa-
ion educed employees’ co isol awakening esponse
(CARi) as well as lunch ime and bed ime co isol le -
els. In e es ingly, he lowe co isol le els a lunch ime
and bed ime we e obse ed in he sho un (a e 0.25
mon hs) while lowe ed educed CARi was only ound in
he long un (a e 5–6 mon hs). The in e en ion co -
e ed a pe iod o six mon hs, du ing which he pa icipan s
engaged in a 20-minu e elaxa ion exe cise du ing lunch
b eaks while he ma ched con ol g oup engaged in small
alk. Sel - epo ed s ain (K ajewski, Wieland & Saue land
2010) and sleepiness (Schniede e al. 2013) also dec eased
To en e e al: Wo kplace Lunch B eak In e en ion A . 2, page 3 o 15
immedia ely a e he lunch b eak and a he end o he
a e noon in he elaxa ion g oup.
Pa k walks combine wo elemen s, which ha e been
shown o be bene icial o well-being and heal h: expo-
su e o na u e and physical ac i i y. Among heal hy popu-
la ions e en sho exposu e o na u al en i onmen s can
educe men al a igue and s ess and imp o e well-being
(Ba on & P e y 2010; Be man, Jonides & Kaplan 2008;
B a man e al. 2015). Sys ema ic e iews (Bowle e al.
2010; Thompson Coon e al. 2011) ha e shown ha exe -
cising in a na u al en i onmen (so-called “g een exe cise”)
imp o es psychological well-being mo e han indoo o
u ban a ea exe cise. Also, in s uden samples, posi i e
e ec s o na u e walks on ca dio ascula measu es such
as educed blood p essu e o hea a e a iabili y ha e
been epo ed (Ha ig e al. 2003; Pa k e al. 2010; Song
e al. 2015). Ye he e idence o changes in neu oendo-
c ine measu es such as co isol sec e ion is inconsis en .
Findings show ha ou side he occupa ional con ex ,
walks in na u al se ings lowe ed co isol le els mo e han
u ban walks (Lee e al. 2011; Pa k e al. 2010), ha e had
no e ec s in compa ison o u ban walks (Ty äinen e
al. 2014), o ha e had posi i e e ec s simila o hose o
u ban walks (Gidlow e al. 2016).
Rega ding he en i onmen , one in e en ion s udy
in es iga ed he e ec o physical ac i i y ei he in na u-
al o buil en i onmen s du ing he lunch b eak (B own
e al. 2014). This eigh -week RCT among 73 o ice wo k-
e s showed ha sel - epo ed men al heal h imp o ed in
he na u al en i onmen lunch ime walking g oup, bu
no in he buil en i onmen walking g oup, o among
he con ols. The e was also a dec ease in sys olic blood
p essu e in he na u al en i onmen walking g oup com-
pa ed o he buil en i onmen walking g oup. Howe e ,
as a dec ease in he sys olic blood p essu e also occu ed
in he con ol g oup, his may sugges ha he posi i e
changes in he na u al en i onmen walking g oup we e
a ibu able o some hing o he han he ial i sel . No
changes in o he heal h pa ame e s (e.g., hea a e a ia-
bili y, body mass index) we e ound. The au ho s poin ou
ha he 20-minu e walking exe cise was comple ed wice
a week and adhe ence was poo . This means ha he non-
signi ican esul s may be due o insu icien exposu e.
The aim o he p esen s udy
The aim o his s udy was o examine he impac o lunch-
ime elaxa ion exe cises and pa k walks on co isol sec e-
ion and blood p essu e. This RCT encompassed a wo-
week in e en ion pe iod du ing which he pa icipan s
we e eques ed o unde ake ei he a 15-minu e elaxa ion
exe cise, a pa k walk o spend hei lunch b eaks as usual
(con ol). This pe iod is a he sho , bu we expec ed posi-
i e in e en ion e ec s since me a-analy ic indings ha e
shown ha sho e in e en ions (1–4 weeks) a e mo e
e ec i e han longe ones (Richa dson & Ro hs ein 2008).
We based ou expec a ions on he s ess eac ion model,
acco ding o which he elease o s ess is conside ed as a
dec ease in s ain (Zap , Do mann & F ese 1996). In ou
s udy, his means dec eased physiological s ess eac ions.
Thus we expec ed o ind a educ ion in co isol esponses
(CARi, CDD) du ing he in e en ion weeks in bo h elaxa-
ion exe cise and pa k walk g oups. Conce ning blood
p essu e, which was measu ed h ee imes pe day (mo n-
ing, a e noon, e ening), we expec ed blood p essu e in
he a e noon o show he mos ma ked dec ease in bo h
in e en ion g oups as his measu emen was he closes
in ime ollowing he in e en ion du ing lunch ime. The
ollow-up measu emen s o co isol and blood p essu e
we e scheduled o one week a e he in e en ion o
examine po en ially pe sis ing sho - e m changes. Based
on he accumula ion model o s ess (Zap e al. 1996),
i is easonable o expec long- e m in e en ion e ec s
as well. Speci ically, he accumula ion model s a es ha
e en when s ess is eleased, i may ake longe o an
indi idual’s physiological sys em o ge id o he s ain.
Howe e , wi hou any de ailed heo y o change ega ding
he ime (when), du a ion (how long) and he shape ( o m
o e ime) o he ela ionships be ween he e ec s o ou
in e en ions and hei ou comes, i is impossible o make
heo e ically based choices o ime-lags (see Kelloway &
F ancis 2013).
Me hod
Design and p ocedu e
This s udy u ilized a andomized con olled ial design
and las ed ou wo king weeks in o al, wi h wo weeks
cons i u ing he in e en ion pe iod. Du ing he wo
in e en ion weeks pa icipan s we e ins uc ed o engage
o 15 minu es du ing hei wo king day lunch b eaks
(al oge he 10 days) in one o he ac i i ies ha hey we e
andomly assigned o: 1) elaxa ion exe cise, 2) pa k walk-
ing, 3) usual b eak ac i i ies (con ol g oup). Pa icipan s
in he elaxa ion and pa k walking g oups we e ins uc ed
o ea lunch be o e engaging in hei in e en ions. The
andomiza ion (based on compu e ized andom numbe s)
was done a he o ganiza ional le el, ha is, he e we e
pa icipan s om each o he h ee g oups in e e y o gan-
iza ion aking pa in he s udy.
The elaxa ion exe cise was based on he elease-only
e sion o p og essi e muscle elaxa ion (Ös 1987) and
a deep b ea hing and accep ance exe cise (Tuomis o
2007). The elease-only e sion o p og essi e elaxa ion
p ima ily a ge s muscle elaxa ion and he deep b ea h-
ing exploi s he po en ial o he agal in luence on he
au onomic ne ous sys em. Accep ance e e s o e ain-
ing om e alua ing o labeling one’s expe iences, ha is,
being a en i e in an accep ing and nonjudgmen al way
(Bae 2003). The elaxa ion exe cise was augh by a psy-
chologis o a ained psychology s uden o hal an hou
in a g oup se ing a he o ganiza ion si e be o e s a ing
he in e en ion pe iod. The pa icipan s we e ad ised
ha each elaxa ion session du ing he in e en ion
pe iod should las 15 minu es and we e also gi en w i en
elaxa ion ins uc ions.
The pa icipan s assigned o he pa k walking g oup
walked a p ede e mined ou e in he nea es pa k a
a low-in ensi y pace. We had made su e ha all he
pa icipa ing companies had a pa k nea by (wi hin
To en e e al: Wo kplace Lunch B eak In e en ionA . 2, page 4 o 15
i e minu es’ walking dis ance). The pa icipan s we e
asked o pay a en ion o hei su oundings and o
a oid alking du ing his 15-minu e walk. Be o e he
in e en ion pe iod s a ed, he ou e was walked once
oge he in a g oup guided by a aine , and we e gi en
maps showing he ou e. Du ing he in e en ion weeks
he pa icipan s could ei he walk alone o in g oups.
Those walking in g oups we e discou aged om alking
o each o he , as he idea was o ocus on he na u al
su oundings.
Du ing he ou wo king weeks, measu emen s o
co isol and blood p essu e we e aken on eigh days: on
Tuesday and Thu sday be o e he in e en ion, du ing
he wo in e en ion weeks, and he week immedia ely
a e he in e en ion. Fo p ac ical easons he ou -week
RCT was ca ied ou in wo phases in 2014: sp ing (weeks
19–22) and all (weeks 36–39). The p ocedu e was exac ly
he same bo h imes, meaning ha he all in e en ion
s udy cons i u ed a eplica ion o he sp ing in e en ion
(see Figu e 1, o he s udy design).
A e he s udy, all pa icipan s ecei ed w i en indi-
idual eedback on hei physiological indica o s and
we e in i ed o a end a lec u e abou he bene i s o
na u al en i onmen s and elaxa ion. We also a led
h ee a el ouche s wo h 400€ in o al among all
hose who had comple ed he online ques ionnai e a
he beginning o he s udy. The esea ch was app o ed
by he E hics Commi ee o he Tampe e Region
(S a emen 10/2014) and he ial was duly egis e ed
(NCT02124837). The ull s udy p o ocol has been pub-
lished by De Bloom, Kinnunen and Ko pela in 2014.
Pa icipan s
The pa icipan s we e ec ui ed wi h he help o a
Finnish company supplying occupa ional heal h ca e
se ices in he ci y o Tampe e and i s su oundings. The
ollowing exclusion c i e ia we e applied: a) shi wo k
o ex emely i egula wo king hou s and b) se ious
illness o alle gies ende ing walking ou doo s impos-
sible. O he app oxima ely 2,226 people app oached,
279 eplied and me he c i e ia (see Figu e 2).
In he end, people om se en di e en companies
pa icipa ed.
Fo p ac ical easons a company was included in he
s udy when a leas six o i s employees olun ee ed
o pa icipa e. This p econdi ion educed he g oup
o pa icipan s om 279 o 225. O he emaining
225 pa icipan s, 53 d opped ou ei he be o e (48) o
du ing (5) he in e en ion o a ious easons (e.g.,
sickness, a el o holiday plans du ing he in e en-
ion weeks). Fi e o hese 53 indi iduals d opped ou
be o e he pa icipan s had been andomized in o
he s udy g oups. O he 48 la e d opou s, 18 (24%)
le he pa k walk g oup, 20 (26%) le he elaxa ion
g oup and 10 (14%) le he con ol g oup. The numbe
o d opou s be ween he s udy g oups did no di e
signi ican ly (χ2 (2, 220) = 3.50, p > .05). Fu he mo e,
19 pa icipan s had o be excluded because 1) hey did
no engage in he pa k walking/ elaxa ion exe cise
in ensi ely enough (less han six ou o en imes,
13 pa icipan s) o 2) hei da a we e la gely missing
(six pa icipan s). The inal sample included 153
pa icipan s (see Figu e 2).
Figu e 1: A e age awakening imes and aking he measu emen s [C (AW) = Co isol a e awakening; C (AW30) =
Co isol 30 minu es a e awakening; C (E) = E ening co isol; BP = Blood p essu e] a e shown in he uppe pa
o he igu e. Measu emen days [d1–8 = measu emen day; T1 = baseline, T2 = i s in e en ion week, T3 = second
in e en ion week, T4 = ollow-up] a e shown in he lowe pa o he igu e.
To en e e al: Wo kplace Lunch B eak In e en ion A . 2, page 5 o 15
Figu e 2: Flowcha o pa icipan s.
Measu emen s
Be o e he s udy s a ed, we o ganized aining sessions
in e e y o ganiza ion du ing wo king hou s. In hese ses-
sions we wen h ough he esea ch p o ocol, p ac iced
he in e en ion ac i i ies and explained how o collec
sali a samples and measu e blood p essu e. All ins uc-
ions we e also included in he pape -pencil bookle s in
which he pa icipan s epo ed hei measu emen s. A
he end o he aining session, all pa icipan s p o ided
w i en in o med consen .
Be o e and a e he daily lunch b eak in e en ion, pa -
icipan s in he in e en ion g oups epo ed hei le el

To en e e al: Wo kplace Lunch B eak In e en ionA . 2, page 6 o 15
o ension on a scale om 0 o 100 in he pape -pencil
bookle . The scale was ancho ed as ollows: 0 = ex emely
elaxed, such as si ing on he couch a e exe cising and
ha ing a sauna ba h, 50 = no mal le el o elaxa ion/
ension, 100 = ex emely ense, such as be o e a majo ,
s ess ul li e e en o an impo an speaking appea ance
in public.
Co isol was sampled om sali a using Sali e e swabs.
To measu e CAR, he samples we e aken immedia ely on
waking (AW) and 30 minu es a e waking (AW30). The
hi d sample was collec ed in he e ening be o e going
o sleep (see Figu e 1). The pa icipan s w o e down
he exac imes o awakening and sali a sampling in he
pape -pencil bookle . They we e also asked o e ain
om ea ing, d inking, hea y physical exe cise and b ush-
ing hei ee h o 30 minu es be o e sali a sampling. To
assess compliance, he pa icipan s we e asked o epo
in hei bookle s whe he hey had engaged in any o
hese beha io s o had aken medica ion (“Please ick
[ he box] i you did any o he ollowing ac i i ies wi hin
he las 30 minu es”) be o e sampling sali a o measu ing
blood p essu e le els.
The sali a samples we e s o ed in he pa icipan s’ home
e ige a o s un il sample collec ion was comple ed, and
hen collec ed by he esea che s in chilled boxes, mailed
o and analyzed a he Finnish Ins i u e o Occupa ional
Heal h, as hese samples ole a e he p e ailing empe -
a u es du ing shipmen (S alde e al. 2016). Values o
sali a y co isol we e analyzed using a LIA ki (LIA, IBL,
Hambu g, Ge many) and epo ed in nmol/l.
We calcula ed CARi by de e mining he di e ence
be ween co isol alues a awakening +30 (AW30) and
awakening (AW). A smalle (less s eep) CARi indica es
be e physiological eco e y. CDD was calcula ed by
de e mining he di e ence be ween co isol alues a
awakening +30 (AW30) and bed ime. Unlike he CARi,
a CDD index inc ease is ela ed o be e physiological
eco e y. We a e aged he co isol alues o Tuesday and
Thu sday o e e y week, as he e we e no signi ican di -
e ences be ween days ( = −2.08‒1.31; p alues > .05,
using Bon e oni co ec ion). Mo eo e , a e aging he
wo days also made he indexes mo e obus week-le el
indica o s and less p one o i ele an daily luc ua ions.
Be o e combining he sp ing and all da a se s in he
analyses, we checked ha he sp ing and au umn RCTs
did no di e signi ican ly in co isol (AW, AW30, e ening;
= −2.15‒0.74; p alues > .05).
Employees measu ed hei blood p essu e (BP) using a
digi al blood p essu e moni o (Om on M2) h ee imes a
day (a e waking, in he a e noon abou 30-60 minu es
be o e he end o he wo king day and be o e going o
bed). A e e y measu emen , pa icipan s measu ed hei
es ing BP a e a i e-minu e es ing pe iod. They always
measu ed hei blood p essu e wice wi h a wo-minu e
ime pe iod be ween measu emen s o ensu e measu e-
men eliabili y. The pa icipan s w o e down in he book-
le he exac alues o hei sys olic BP (SBP) and dias olic
BP (DBP) in mmHg (as au oma ically displayed by he
moni o ). We calcula ed an a e age o hese wo highly
co ela ed measu emen s ( = .85–.91, p < .001, and =
.92–.94, p < .001), o all sys olic and dias olic BP measu e-
men s. We a e aged he alues on Tuesday and Thu sday
o e e y week. O he 24 - es s pe o med, only SBP on
awakening in he i s week (T1) was signi ican ly highe
on he Tuesday han on he Thu sday a e Bon e oni co -
ec ion ( = 3.39, p = .001). Addi ionally, as he BP alues
o he sp ing and au umn RCTs did no di e signi ican ly
( = ‒1.77 ‒ 0.76; p alues > .05 using Bon e oni co ec ion),
we used he whole sample in he analyses.
Backg ound in o ma ion (gende , educa ion, age, weekly
wo king hou s, wo kload, au onomy and social suppo
a wo k, job exhaus ion, diagnosed psychia ic o endo-
c ine diseases, and hype ension) was collec ed h ough
an online ques ionnai e a he beginning o he s udy.
Wo kload, au onomy, and suppo desc ibe he key cha -
ac e is ics o any job based on Ka asek’s model (Ka asek &
Theo ell 1990). Wo kload was measu ed wi h h ee i ems
(e.g., “How o en does you job equi e you o wo k unde
ime p essu e? C onbach’s alpha = .88”) om he QWI
(Spec o & Jex 1998). Job au onomy (5 i ems, e.g., “I ha e
lexibili y in se ing my own wo king hou s”, C onbach’s
alpha = .77) and social suppo om colleagues and
supe iso s (6 i ems, e.g., “I needed, I can ge suppo
and help wi h my wo k om my cowo ke s”, “My wo k
achie emen s a e app ecia ed by my immedia e supe io ,”
C onbach’s alpha = .78) we e measu ed wi h i ems om
he QPS No dic-ADW (Dallne e al. 2000). All i ems we e
a ed on a scale anging om 1 ( e y seldom o ne e )
o 5 ( e y o en o always). Job exhaus ion was measu ed
wi h i e i ems (e.g., “I eel emo ionally d ained om my
wo k”, C onbach’s alpha = .94) om he MBI—Gene al
Su ey (Maslach, Jackson & Lei e 1996), which has been
alida ed in Finland (Kalimo, Hakanen & Toppinen-Tanne
2006). The esponse scale anged om 0 (ne e ) o 6
(e e y day). Diseases we e assessed by single i ems (yes = 1,
no = 0).
Da a exclusion and s a is ical analyses
When analyzing co isol da a, we sc eened he da a o
ensu e he alidi y o measu emen s. We excluded ou
pa icipan s wi h a physician-diagnosed psychia ic dis-
ease and 25 wi h endoc ine diseases om he sample.
Fu he mo e, 36 pa icipan s we e excluded as hey had
sys ema ically la ened o nega i e CAR p o iles (i.e., on
mo e han on 50% o he days he ise om awakening o
awakening +30 measu emen was less han 2.5 nmol/l;
Wüs e al. 2000). These so-called CAR non- esponden s
we e equally dis ibu ed among he in e en ion and con-
ol g oups (χ2 (2, 87) = .38, p = .83). The inal co isol sam-
ple included 88 pa icipan s (see Figu e 2).
Single co isol samples ( alues) we e also excluded: Each
o he h ee measu emen s pe day was epea ed on wo
days pe week o e ou weeks, hus ou s udy esul ed in
24 co isol alues pe pe son and 3,672 co isol alues in
o al. Ou lie s beyond h ee s anda d de ia ions om he
mean (137 alues, 3.7%) we e i s emo ed. In addi ion,
when a pa icipan had consumed alcohol be o e collec -
ing he sample (9 alues), did no collec samples a he
To en e e al: Wo kplace Lunch B eak In e en ion A . 2, page 7 o 15
appoin ed imes (i.e., o e 15 minu es’ sel - epo ed delay
in he mo ning sample compa ed o awakening ime, 78
alues), o had an occasional la o nega i e CAR (330 al-
ues). These condi ions accoun ed o 417 excluded alues
(11.4%).
Co isol da a had skewed dis ibu ions and we e Log10
ans o med. Nex , we analyzed he pa e ns o miss-
ing da a. The missing da a i ems anged om 13.6% o
19.3% o CARi and om 13.5% o 21.6% o CDD pe
measu emen ime. De ailed inspec ion o he missing
da a i ems showed ha he highes a es o missing da a
we e o be ound in he las week o he s udy (T4). In
hese ci cums ances, da a impu a ion was needed o
inc ease bo h s a is ical powe and ex e nal alidi y (see
Newman 2009). The mos common and ecommended
missing da a impu a ion me hod is mul iple impu a-
ion based on maximum likelihood es ima ion (Ho on
& Kleinman 2007; Scha e & G aham 2002). This me hod
comple es missing da a wi h alues andomly gene a ed
om an adjus ed dis ibu ion o he ou come a iable
and co a ia es, epea ing he p ocedu e mul iple imes.
The esul s a e hen pooled o ob ain poin and a iance
es ima es o he a iable o in e es . Acco dingly, we
conduc ed mul iple impu a ions using logis ic eg ession
wi h i e impu a ion subsamples o gene a e he pooled
impu ed da ase .
Simila ly, when analyzing BP da a, we sc eened he da a
o ensu e he alidi y o he measu emen s. As each o he
h ee measu emen s pe eigh days we e epea ed a wo-
minu e in e als and p oduced bo h SBP and DBP alues,
ou s udy esul ed in 96 BP alues pe pe son and hus in
a o al o 14,688 BP alues. Clea ly e oneous single alues
(4 alues, 0.03%) and ou lie s beyond h ee s anda d de i-
a ions om he mean (63 alues, 0.4%) we e emo ed.
Single alues we e also emo ed i pa icipan s had con-
sumed alcohol be o e measu ing BP (which occu ed on
nine occasions be o e e ening measu emen s esul ing in
36 excluded alues, 0.3%) o had exe cised hea ily (which
occu ed on 14 occasions be o e e ening measu emen s
esul ing in 56 excluded alues, 0.4%). Fu he mo e, we
checked ha he measu emen s we e aken a he imes
ad ised (six alues, 0.04%, excluded o his eason). We
excluded wo cases wi h diagnosed hype ension ha -
ing sys ema ically and a ypically high BP alues (mo e
han wo s anda d de ia ions abo e he sample a e age).
A e his da a cleaning, he inal sample o BP analy-
ses included 151 pa icipan s (see Figu e 2). Impu a ion
was no needed o he BP da a as he numbe o missing
obse a ions emained low.
We used mixed-design ANOVAs o es whe he he
h ee g oups ( elaxa ion, pa k walk, and con ol) di e ed
in CARi, CDD, SBP, and DBP o e he in e en ion pe iod
(T1–T4). The in e en ion and con ol g oups ac ed as a
be ween-g oups ac o and ime was a epea ed measu e.
A signi ican G oup × Time in e ac ion e ec would sug-
ges a signi ican e ec o he in e en ion, ha is, he
physiological indexes would show di e en changes ac oss
ime in he h ee g oups. To unde s and he magni ude o
hese di e ences, we also calcula ed e ec sizes (Cohen d)
compa ing he physiological indexes a baseline (T1) and
in he second in e en ion week (T3). We expec ed ha
he e ec o he in e en ion would be la ges owa ds
he end o he wo-week in e en ion pe iod. E ec sizes
smalle han 0.2 a e conside ed i ial, d’s be ween 0.2
and 0.5 a e de ined as small, d’s be ween 0.5 and 0.8 a e
conside ed medium, and d’s g ea e han 0.8 a e in e -
p e ed as la ge e ec sizes (Cohen 1992; Sulli an & Feinn,
2012). All analyses we e pe o med using SPSS 22.
Resul s
Main sample cha ac e is ics and adhe ence o p o ocol
The e we e no di e ences in backg ound ac o s (gen-
de , educa ion, age, weekly wo king hou s, wo kload,
au onomy, social suppo , exhaus ion) and waking imes
be ween he in e en ion and con ol g oups when he
co isol da a a baseline we e analyzed. Fo BP da a, how-
e e , di e ences in educa ion eme ged: he in e en ion
g oups had highe educa ional le els han did he con ol
g oup (χ2 (2, 150) = 7.45, p = .02; see Tables 1 and 2).
Howe e , educa ion did no co ela e wi h BP alues
( = −.08‒.12; p > .05). Age was posi i ely associa ed wi h
BP ( = .23‒.44; p < .05), bu he e we e no signi ican age
di e ences be ween he in e en ion and con ol g oups
(F (2, 132) = 1.31, p = .27).
Time o waking co ela ed nega i ely wi h CARi on
he i s ( = −.34, p < .05) and i h ( = −.30, p < .05)
measu emen days, indica ing ha he ea lie he pa ici-
pan s woke up he s eepe he CARi index. Addi ionally,
CDD co ela ed posi i ely wi h waking ime on he six h
day ( = .26, p < .05), ha is, he la e pa icipan s woke,
he highe hei CDD index. Howe e , he e we e no di -
e ences be ween he wo in e en ion and he con ol
g oups in waking imes on any o he eigh measu emen
days (F = 0.32–2.21, p > .05).
Tables 1 and 2 show – besides he main cha ac e is ics
o he sample – co isol and BP alues a baseline o each
g oup. No g oup di e ences in co isol o in BP alues a
baseline eme ged.
Rega ding adhe ence o p o ocol, we examined whe he
he e we e signi ican g oup di e ences in e ms o he num-
be o exe cises comple ed, exe cise du a ion, and he leng h
o he lunch b eaks. On a e age, pa icipan s engaged in
elaxa ion exe cises o pa k walks 8.6 imes ou o en du ing
he wo-week in e en ion pe iod. The e we e no di e ences
in he equency o elaxa ion exe cises o pa k walks (F (1, 95)
= 0.14, p = .71). This held o pa icipan s wi h bo h co isol
and BP da a. In he elaxa ion g oup lunch b eaks las ed on
a e age 27 minu es and he exe cise on a e age 14 minu es
( ange: 8–20). In he pa k walk g oup lunch b eaks las ed on
a e age 28 minu es, o which he pa icipan s engaged in
he exe cise o 15 minu es ( ange: 8–20), and o he con-
ol g oup, lunch b eaks las ed 27 minu es on a e age ( ange:
10–60). Pa k walks las ed on a e age 1–2 minu es longe han
he elaxa ion exe cises (F (1, 88) = 15.2, p < .01). This was he
case o bo h co isol and BP da a. The e we e no s a is ically
signi ican g oup di e ences in he lunch b eak du a ion (F
(2, 150) = 0.05, p = .95). In bo h in e en ion g oups he le els
o ension dec eased signi ican ly on e e y in e en ion day in
To en e e al: Wo kplace Lunch B eak In e en ionA . 2, page 8 o 15
Va iables Relaxa ion (n = 28) Pa k walk (n = 29) Con ol (n = 31) p†
Female, % 89 90 87 .94
Highe educa ion1, % 46 50 47 .53
Age in yea s (mean) 46 47 48 .67
Wo king h/w (mean) 38.2 38.3 38.7 .87
Wo kload 3.8 3.7 3.9 .69
Au onomy 3.4 3.5 3.4 .92
Social suppo 4.1 4.1 4.0 .77
Exhaus ion 1.9 2.0 2.2 .67
Awakening ime, h (mean) 6:23 6:17 6:10 .44
Co isol2 AW 16.0 18.5 16.0 .65
Co isol2 AW+30 min 30.3 31.1 28.7 .77
Co isol2 bed ime 4.8 4.3 4.7 .85
Va iables Relaxa ion (n = 46) Pa k walk (n = 51) Con ol (n = 54) p†
Female, % 91 92 87 .64
Highe educa ion1, % 54 44 28 .02
Age in yea s (mean) 46 48 48 .27
Wo king h/w (mean) 38.6 38.2 38.4 .89
Wo kload 3.8 3.8 3.9 .84
Au onomy 3.2 3.4 3.4 .60
Social suppo 4.0 4.0 4.0 .99
Exhaus ion 1.8 2.0 2.2 .21
Awakening ime, h (mean) 6:18 6:15 6:15 .61
Mo ning SBP (mean) 118.1 117.4 115.0 .43
Mo ning DBP 74.7 75.1 72.8 .35
A e noon SBP 127.4 125.5 124.2 .45
A e noon DBP 79.3 78.6 77.5 .51
E ening SBP 118.3 117.0 115.3 .43
E ening DBP 72.8 73.0 71.6 .57
No es. Abb e ia ion: AW = Awakening ime.
1 Highe educa ion = Mas e ’s deg ee o highe .
2 Co isol in nmol/l wi hou log- ans o ma ion.
p† Re e s o a chi-squa e es o ca ego ical a iables o o a F- es o con inuous a iables.
No es. Abb e ia ions: SBP = Sys olic blood p essu e; DBP = Dias olic blood p essu e; measu ed in mmHg OK.
1 Highe educa ion = Mas e ’s deg ee o highe .
p† Re e s o a chi-squa e es o ca ego ical a iables o o a F- es o con inuous a iables.
he in e en ion g oups. In he pa k g oup, ension dec eased
om an a e age o 57.1 be o e he lunch b eak o an a e age
o 42.4 a e he lunch b eak ( (44) = 7.24, p < .001). In he
elaxa ion g oup ension dec eased om 58.7 o 40.9 ( (35)
= 6.83, p < .001).
ANOVA esul s
Mixed-design ANOVAs we e conduc ed o add ess he
in e en ion e ec s. Table 3 shows he esul s o co isol.
The e we e no di e ences in he CARi index in e ms o
ime, g oup o in e ac ion e ec s. Thus he in e en ion
Table 1: Main sample cha ac e is ics o co isol o he baseline week (n = 88).
Table 2: Main sample cha ac e is ics o blood p essu e o he baseline week (n =151).
To en e e al: Wo kplace Lunch B eak In e en ion A . 2, page 9 o 15
did no ha e any e ec on lowe ing he co isol awaken-
ing esponse in he in e en ion g oups compa ed o he
con ol g oup, which was con a y o expec a ion. The
esul s o he CDD index, add essing co isol decline
h oughou he day, sugges a signi ican in e ac ion e ec
be ween g oup and ime. Speci ically, changes in CDD in
he con ol g oup di e ed om hose in he in e en ion
g oups. The mean alues o he CDD index emained s a-
ble in he in e en ion g oups compa ed o he con ol
g oup, whe e he e was mo e a ia ion wi h s eepe ises
a T2 and T4 (see Table 3).
In Table 3 he mean di e ences in he CARi index
be ween T3 (a he end o he in e en ion) and T1 (a base-
line be o e he in e en ion) a e posi i e. This indica es
a s eepe inc ease in he awakening co isol esponse a
he end o he in e en ion compa ed o baseline be o e
he in e en ion in all h ee g oups. The g ea es inc ease
om T1 o T3 was seen in he elaxa ion g oup (d = .38).
Thus he changes obse ed in CARi we e con a y o ou
expec a ions. Ins ead, he di e ences in he CDD we e in
he expec ed (posi i e) di ec ion, showing a sligh decline
be ween T3 and T1, bu wi h i ial e ec sizes.
Fo BP (Table 4), signi ican ime e ec s eme ged o
ou ou o he six alues (SBP and DBP a mo ning, a e -
noon, and e ening), bu no signi ican in e ac ion o
g oup e ec s we e obse ed. Thus, in iew o he ANOVA
esul s, he in e en ion and con ol g oups did no di -
e in BP o e ime. The signi ican ime e ec s we e
obse ed in he a e noon a he end o he wo king day,
bu also in he mo ning. BP showed a dec easing end
o e ime.
Rega ding he di e ences be ween T3 and T1, he
end in ol ed dec easing SBP and DBP in e e y measu e-
men (mo ning, a e noon, and e ening) o all g oups.
Howe e , he dec ease was mos p onounced in he walk-
ing g oup (d = .28–.58) and smalles in he elaxa ion
g oup (d = .00–.28). The g ea es dec ease in he walking
g oup was seen in SBP (d = .58) and DBP (d = .51) in he
a e noon.
Discussion
This s udy wi h i s ocus on he physiological e ec s o a
eco e y in e en ion du ing he wo king day is among
he i s RCTs in his esea ch a ea. Speci ically, we exam-
ined whe he physiological s ess ma ke s o co isol and
blood p essu e would change du ing and a e a wo-week
in e en ion consis ing ei he o a 15-minu e elaxa ion
exe cise o a pa k walk aken on e e y wo king day a
lunch ime compa ed o a con ol g oup spending hei
lunch b eaks as usual. Al hough he e idence o changes
in physiological ma ke s is limi ed ega ding in e en ions
du ing wo king days (see Richa dson & Ro hs ein 2008,
o a e iew), bo h elaxa ion exe cises and pa k walks
ha e shown posi i e e ec s, especially o blood p essu e,
bu mos ly in s udies conduc ed ou side he occupa ional
con ex (Ha ig e al. 2003; Nyklícek e al. 2013; Pa k e al.
2010; Song e al. 2015).
Ou hypo hesis ha he co isol awakening esponse
(CAR) would show a la ened p o ile and ha he e
would be a decline in co isol exc e ion o e he day
(CDD) in he in e en ion g oups a e he wo-week
in e en ion pe iod was no suppo ed. Thus ou indings
did no suppo he hypo heses o he s ess eac ion o
accumula ed s ess models (Zap e al. 1996). In ac , we
no iced an opposi e endency – a s eepe p o ile – in he
co isol awakening esponse a he end o he in e en-
ion compa ed o he baseline be o e he in e en ion.
Impo an ly, his endency was mo e p onounced in he
elaxa ion g oup (d = .38) han in he walking o con ol
g oups. This means ha ou indings a e opposed o hose
epo ed by K ajewski e al. (2011) who, in examining he
e ec s o elaxa ion exe cises du ing lunch b eaks showed
a la ened CAR p o ile in he elaxa ion g oup. Howe e ,
his educed CAR was only obse ed in he long un (a e
5–6 mon hs). Conce ning pa k walks, he e is no esea ch
on co isol compa able o ou lunch b eak in e en ion
s udy, al hough some s udies ha e shown ha walks in
na u al su oundings lowe co isol alues mo e han do
u ban walks (Lee e al. 2011; Pa k e al. 2010).
T1 T2 T3 T4 F-s a is ic
M SD M SD M SD M SD Δ T3–T1 Cohen’s dTime G oup G × T
CARi
Relax 14.5 9.6 14.2 9.1 19.2 11.1 16.3 8.7 4.7 .38 1.60 .85 .55
Walk 12.8 8.1 12.0 5.5 15.7 10.8 18.1 15.6 2.9 .20
Con ol 12.7 5.0 13.3 8.1 14.1 8.5 14.0 13.5 1.4 .16
CDD
Relax 12.1 8.2 11.0 9.9 13.1 8.7 12.7 4.9 1.0 .09 2.26 .20 2.30*
Walk 11.8 9.3 11.6 8.8 12.3 8.9 12.1 8.1 0.5 .05
Con ol 10.9 8.8 16.6 14.4 11.8 9.6 15.4 11.5 0.9 .11
No es. Means (M), s anda d de ia ions (SD), mean di e ences and Cohen d we e calcula ed using non- ans o med
co isol alues; n = 28, 29, and 31, o elaxa ion, walk, and con ol g oups espec i ely.
* p < .05.
Table 3: Mixed ANOVA esul s o co isol.