1
Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
Open access
Fac o s associa ed wi h he
agg essi eness o ca e a he end o li e
o pa ien s wi h cance dying in
hospi al: a na ionwide e ospec i e
coho s udy in mainland Po ugal
Diogo Ma ins- B anco ,1,2 Sil ia Lopes ,3,4 Ri a Cana io ,1,5 Joao F ei e,2
Madalena Feio ,6 Jose Fe az- Goncal es ,7 Gab iela Sousa ,5
Nuno Lune ,8,9 Ba ba a Gomes 1,10
O iginal esea ch
To ci e: Ma ins- B ancoD,
LopesS, Cana ioR, e al.
Fac o s associa ed wi h he
agg essi eness o ca e a
he end o li e o pa ien s
wi h cance dying in hospi al:
a na ionwide e ospec i e
coho s udy in mainland
Po ugal. ESMO Open
2020;5:e000953. doi:10.1136/
esmoopen-2020-000953
Recei ed 29 July 2020
Re ised 27 Augus 2020
Accep ed 29 Augus 2020
Wa ch Video
esmoopen. bmj. com
Fo numbe ed a ilia ions see
end o a icle.
Co espondence o
D Diogo Ma ins- B anco;
diogo. mb anco@ gmail. com
© Au ho (s) (o hei
employe (s)) 2020. Re- use
pe mi ed unde CC BY- NC. No
comme cial e- use. Published
by BMJ on behal o he
Eu opean Socie y o Medical
Oncology.
ABSTRACT
In oduc ion The e is g owing conce n abou he
agg essi eness o cance ca e a he end o li e (ACCEoL),
de ined as o e ly agg essi e ea men s ha comp omise
he quali y o li e a i s end. Recognising he mos a ec ed
pa ien s is a co ne s one o imp o e oncology ca e. Ou
aim is o iden i y ac o s associa ed wi h ACCEoL o
pa ien s wi h cance dying in hospi als.
Me hods All adul pa ien s wi h cance who died in public
hospi als in mainland Po ugal (Janua y 2010 o Decembe
2015), iden i ied om he hospi al mo bidi y da abase. This
da abase p o ided indi idual clinical and demog aphic
da a. We ob ained hospi al and egion- le el a iables om
a su ey and Na ional S a is ics. The p ima y ou come is
a composi e ACCEoL measu e o 16 indica o s. We used
mul ile el andom e ec s logis ic eg ession modelling
(p<0·05).
Resul s We included 92 155 pa ien s: median age 73
yea s; 62% male; 53% wi h me as a ic disease. ACCEoL
p e alence was 71% (95% CI 70% o 71%). The mos
p e alen indica o s we e >14 days in he hospi al (43%,
42–43) and su ge y (28%, 28–28) in he las 30 days.
Olde age (p<0·001), b eas cance (OR 0·83; 95% CI 0·76
o 0·91), and me as a ic disease (0·54; 95% CI 0·50 o
0·58) we e nega i ely associa ed wi h ACCEoL. In con as ,
highe Deyo- Cha lson Como bidi y Index (p<0·001),
gas oin es inal and haema ological malignancies
(p<0·001), and dea h a cance cen e (1·31; 95% CI 1·01
o 1·72) o hospi al wi h medical oncology depa men
(1·29; 95% CI 1·02 o 1·63) we e posi i ely associa ed
wi h ACCEoL. The e was no associa ion be ween hospi al
pallia i e ca e se ices a he hospi al o dea h and
ACCEoL.
Conclusion Clinical ac o s ela ed o a be e
unde s anding o disease cou se a e associa ed wi h
ACCEoL educ ion. Pa ien s wi h mo e como bidi ies and
gas oin es inal malignancies migh ep esen g oups
wi h complex needs, and haema ological pa ien s may
be a inc eased isk because o unp edic able p ognosis.
Imp o emen o hospi al pallia i e ca e se ices could
help educe ACCEoL, pa icula ly in cance cen es and
hospi als wi h medical oncology depa men , as hose
se ices a e usually unde - esou ced, hus eaching ew.
INTRODUCTION
Towa ds he end o li e (EoL), pa ien s wi h
cance wish o eel com o able, be ea ed
wi h digni y and achie e a sense o comple-
ion.1 They also wish o a oid o e ly agg essi e
ea men s, which can comp omise symp om
con ol and ad ance ca e planning.2 App o-
p ia e managemen o EoL ca e has been
aised as a quali y- o - ca e issue and indica-
o s ha e been de eloped o iden i y heal h
sys ems ha apply o e ly in ensi e ea men s
Key ques ions
Wha is al eady known abou his subjec ?
►The e is g owing conce n abou he agg essi eness
o cance ca e a he end o li e (ACCEoL), de ined as
o e ly agg essi e ea men s ha comp omise he
quali y o li e.
►The end o li e pe iod mos s udied in ela ion o
ACCEoL is he las mon h o li e, o which Ea le e
al6 iden i ied key ACCEoL indica o s, expanded by
Lu a e al.12
►Recognising he mos a ec ed pa ien s is a co ne -
s one o imp o e his public heal h unme need.
Wha does his s udy add?
►Unchanged end o high ACCEoL in a Eu opean
coun y (Po ugal)—7 ou o 10 pa ien s wi h cance .
►Mos p e alen indica o s: >14 days in hospi al and
su ge y in las 30 days o li e.
►Olde age, b eas cance and me as a ic disease
we e associa ed wi h lowe ACCEoL.
►Como bidi ies, gas oin es inal and haema ological
cance s wi h highe ACCEoL.
How migh his impac on clinical p ac ice?
►Clinicians should conside cance ype, disease
s age, como bidi ies, age and he in luence o hos-
pi al oncology cul u e, o help pa ien s wi h cance
o a oid ACCEoL.
Published online
19 No embe 2020
Open access
2Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
o e minal ad anced cance pa ien s wi h e y limi ed
clinical bene i , de ined as agg essi eness o cance ca e
a he EoL (ACCEoL).3 As one componen o EoL, he
ACCEoL is in e connec ed wi h dele e ious e ec s o
pa ien s and amilies such as wo se quali y o li e and
be ea emen ou comes.4 5 This led o g owing conce n
ac oss socie ies abou he ACCEoL.6–11
Ea le e al6 epo ed one o he i s ACCEoL s udies,
based on adminis a i e da a measu ing key indica o s
wi hin he las mon h o li e, including o e use o chemo-
he apy (new egimen wi hin 30 days, o any adminis a-
ion wi hin 14 days be o e dea h), unde use o hospice
ca e, and high a es o eme gency oom isi s, hospi alisa-
ion o in ensi e ca e uni (ICU) admissions. They ound
ha each indica o was p esen in less han one- hi d o
pa ien s bu also ha ACCEoL p e alence was inc easing
in he 90s in he USA.6 S udies ollowed in o he coun-
ies, in e es ingly wi h di e en indings.7–11 Lu a e al12
p oposed an expanded lis o ACCEoL measu es based
on a sys ema ic e iew.
Despi e hese ad ances in ACCEoL esea ch, he issue
emains neglec ed ye i al o clinical oncology, in ace
o he g owing numbe o people dying wi h cance .13
Despi e wo- hi ds o he popula ion exp ess a p e e ence
o die a home in a scena io o ad anced cance 14 and
inc eases in se e al na ions on he pe cen age o cance
pa ien s dying a home a he han in hospi al,15 s ill many
pa ien s emain inc easingly exposed o o he ACCEoL
indica o s, such as ICU admission.16 The ise o hospi al
dea hs in o he coun ies sugges s a le el o dependency
on hospi al esou ces and o ACCEoL which a e con a y
o people’s p e e ences.14 17–19
Recognising he p o ile o pa ien s a isk o ecei ing
ACCEoL is c i ical o be e unde s anding his public
heal h unme need and imp o ing oncology ca e. Ou
s udy aims o iden i y ac o s associa ed wi h ACCEoL o
cance pa ien s dying in hospi als.
METHODS
S udy design and se ing
This is a na ionwide e ospec i e coho s udy o adul s
who died wi h cance in Po uguese hospi als. The s udy
ollowed he REpo ing o s udies Conduc ed using Obse -
a ional Rou inely- collec ed heal h Da a s a emen .20
Pa ien s
We included all pa ien s ha : (1) died in a public
hospi al in mainland Po ugal be ween Janua y 2010 and
Decembe 2015; (2) we e aged ≥18 yea s a he ime o
dea h and (3) had a diagnosis o cance eco ded in he
episode leading o dea h, using In e na ional Classi ica-
ion o Diseases, nin h Re ision, Clinical Modi ica ion
(ICD-9- CM) codes om he chap e ‘neoplasms’ (codes
140–239), excluding benign neoplasms, ca cinoma in
si u, neoplasms o unce ain beha iou o unspeci ied
na u e (210-239).
P ima y ou come
S udy ou come is a composi e bina y measu e o ACCEoL,
posi i e in he p esence o a leas one o 16 indi idual
indica o s (S1 - online supplemen al ile 1). The lis o
indica o s om Ea le e al6 was expanded based on he
sys ema ic e iew by Lu a e al.12 A na ional expe panel
assessed con en alidi y. Fo all pa ien s, all indica o s
we e measu ed o he las 30 days o li e, excep he use
o chemo he apy and immuno he apy/biological agen s,
which was sho ened o he las 14 days o li e, ollowing
Ea le e al’s c i e ia.6
Da a sou ces
We used he hospi al mo bidi y da abase (HMD) o iden-
i y pa ien s and ob ain indi idual- le el da a. The HMD
con ains ou inely- collec ed da a om all public hospi als
in mainland Po ugal o unding pu poses, since 1989.21
The da ase was anonymised by he Po uguese Heal h
Sys em Cen al Adminis a ion and included: (1) demo-
g aphic da a: sex, age a dea h and bo ough o esidence;
(2) clinical da a coded by ICD-9- CM: me as a ic disease
s a us, p ocedu es and main o seconda y diagnoses used
o he calcula ion o he Deyo- Cha lson Como bidi y
Index (DCCI)— used o p edic mo ali y ha de i es
om he sum o he sco e a ibu ed o each como bidi y
ou o 17 ch onic medical condi ions22 23 and (3) admin-
is a i e da a: da e o admission, discha ge o dea h, ype
o ea men (medical s su gical) and hospi al o dea h.
In Sep embe 2016, we su eyed all hospi al adminis a-
ion boa ds and di ec o s o pallia i e ca e se ices (PCS)
c ea ed be o e Janua y 2016. Following ecommenda ions
om a Coch ane Re iew on me hods o inc ease esponse
o pos al and elec onic ques ionnai es,24 we ob ained a
78% esponse a e. We used a semis uc u ed ques ion-
nai e o ob ain he ollowing in o ma ion: (1) hospi al:
hospi al ype (gene al s cance cen e), hospi al dimen-
sion (numbe o beds), exis ence o medical oncology
depa men (MOD); (2) PCS: exis ence and c ea ion
da e, exis ence o pallia i e ca e uni and numbe o beds.
We made ollow- up con ac s and consul ed he na ional
di ec o y o PCS o comple e missing da a. Hospi al- le el
da a we e linked o indi idual- le el da a by hospi al o
dea h a iable om HMD.
We ob ained lis s o bo oughs classi ied by heal h
egion, u banisa ion le el (p edominan ly u ban, mid-
u ban, p edominan ly u al) based on Census 2011 da a
and by dep i a ion le el based on Census 2001 da a
(Eu opean Dep i a ion Index).25 Region- le el da a we e
linked o indi idual and hospi al- le el da a by bo ough o
esidence a iable om HMD.
S a is ical analysis
We i s desc ibed he s udy popula ion, compa ing
he cha ac e is ics o pa ien s who ecei ed ACCEoL
wi h hose ha did no , using Pea son’s χ2 o Mann-
Whi ney es s. We hen de e mined he p e alence o he
composi e measu e o ACCEoL and each indi idual indi-
ca o o he whole popula ion and by me as a ic disease
Open access
3
Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953 Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
s a us, p ima y cance si e and ype o hospi al o dea h.
We examined ends in composi e and indi idual indica-
o s om 2010 o 2015, using χ2 es o end. We consid-
e ed a di e ence o >5% as clinically meaning ul.
Taking he composi e ACCEoL measu e as he depen-
den a iable, unadjus ed odds a io (ORs) wi h 95%
con idence in e al (CI) we e calcula ed o each inde-
penden a iable. We used mul ile el andom e ec s
logis ic eg ession modelling (accoun ing o indi idual,
hospi al and egion le els), including yea o dea h and
all independen a iables ha showed associa ion wi h
ACCEoL on unadjus ed analysis. Finally, we conduc ed
a subg oup analysis o pa ien s wi h me as a ic disease.
All analyses we e based on comple e cases using STATA.
IC12.1 (p<0·05).
RESULTS
Pa ien cha ac e is ics
We included 92 155 pa ien s (S2 - online supplemen al ile
1), 62% male and median age o 73 yea s (in e qua ile
ange - IQR, 62–81). Fi y- h ee pe cen had me as a ic
disease and he mos common p ima y solid umou si es
we e lung (16%) and colo ec al (10%). Twel e pe cen
had haema ological malignancies. The median DCCI
sco e was eigh poin s (IQR, 4–9). Only 15% died in a
cance cen e bu nea ly all pa ien s died in a hospi al
wi h MOD (93%). Mos (66%) died in a hospi al wi h
hospi al PCS (hPCS). The cha ac e is ics o he g oups
wi h and wi hou ACCEoL we e s a is ically di e en a :
i) indi idual- le el: all a iables excep yea o dea h; ii)
hospi al- le el: hPCS, MOD, hospi al ype, hospi al dimen-
sion, heal h egion; and iii) egion- le el: pallia i e ca e
uni beds/100 dea hs- yea ( able 1).
PCS esou ces
Su ey esul s showed ha he pe cen age o hospi al
cen es wi h hPCS inc eased om 42% (13/31) in 2010
o 74% (23/31) in 2015 (p<0·05), e lec ed in an inc ease
o he pe cen age o pa ien s who died in hospi als wi h
hPCS (50% in 2010 o 82% in 2015, p<0·001) (S3 - online
supplemen al ile 1).
ACCEoL p e alence
The p e alence o he composi e ACCEoL measu e was
71% (95% CI 70% o 71%). The di e ence by me a-
s a ic disease s a us was only 3% (pa ien s wi h me a-
s a ic disease: 70% s o he s: 73%). ACCEoL also a ied
by p ima y cance si e (63% in b eas cance o 79% in
haema ological malignancies) and ype o hospi al o
dea h (74% in cance cen e s 69% in gene al hospi al)
(S4 - online supplemen al ile 1).
The mos p e alen indica o s we e >14 days o leng h
o s ay in hospi al (43%, 95% CI 42 o 43) and su ge y
(28%, 95% CI 28 o 28) wi hin he las 30 days, wi h no
clinically meaning ul di e ences be ween pa ien s wi h
e sus wi hou me as a ic disease excep o ICU admis-
sion (4% s 10%, espec i ely), mechanical en ila ion
(2% s 8%) and inse ion o endo acheal ube (15% s
22%) ( igu e 1).
The p ima y ou come emained s able o e ime ( om
71% in 2010 o 72% in 2011) and despi e some indica-
o s showing s a is ically signi ican changes om 2010 o
2015, none we e conside ed clinically meaning ul (S5 -
online supplemen al ile 1).
Fac o s associa ed wi h ACCEoL
In mul i a ia e analysis, olde age (p<0·001), b eas
cance (OR 0·83; 95% CI 0·76 o 0·91), and me as a ic
disease (OR 0·54; 95% CI 0·50 o 0·58) we e nega i ely
associa ed wi h ACCEoL. In con as , highe DCCI
(p<0·001), gas oin es inal and haema ological malig-
nancies (p<0·001), and dea h a a cance cen e (OR
1·31; 95% CI 1·01 o 1·72) o a a hospi al wi h MOD (OR
1·29; 95% CI 1·02 o 1·63) we e posi i ely associa ed wi h
ACCEoL. Adjus ing o con ounde s, he e was no asso-
cia ion be ween exis ence o hPCS and ACCEoL. The e
was a con ex ual e ec o hospi al- le el wi h a median OR
o 1·20 (95% CI 1·15 o 1·27), bu no e ec o egion
(S6 - online supplemen al ile 1). The subg oup analysis
o pa ien s wi h me as a ic disease showed also a nega i e
associa ion wi h ACCEoL o male sex (OR 0·93, 95% CI
0·88 o 0·98) and cance as main diagnosis o las hospi al
admission (OR 0·88, 95% CI 0·83 o 0·94), and a posi i e
associa ion o yea o dea h (OR 1·03, 95% CI 1·01 o
1·05) ( able 2).
DISCUSSION
Ou s udy showed an unchanged end o high p e a-
lence o ACCEoL in a Eu opean coun y. Se en ou o 10
adul cance pa ien s dying in Po uguese public hospi als
be ween 2010 and 2015 ecei ed ACCEoL, mo e han in
o he Wes e n coun ies (71% s 22%–65%).6 7 10 26
The high a e o hospi alisa ion may e lec no only he
in ensi y o clinical ca e, bu also he ex en o social and
clinical suppo in communi y and home se ings. F ee
access o he Na ional Heal h Se ice o cance pa ien s
acili a es hospi alisa ion in a con ex o sca ce commu-
ni y suppo . This should be conside ed when es ab-
lishing compa isons wi h di e en heal h sys ems, whe e
he p e alence o he mos common indi idual indica o
o ou s udy, >14 days in hospi al wi hin he las mon h
o li e, was lowe (43% s 11%–30%).6 9 On he o he
hand, we ound ICU admission p e alence a he bo om
o he p e iously epo ed ange (6% s 3%–25%)5–9 and
less p e alen in pa ien s wi h me as a ic disease (4% s
10%), sugges ing ha he educed use o limi ed and
me ely clinically d i en hospi al esou ces may e lec an
adequa e in ensi y o clinical ca e, despi e he high a e o
ACCEoL mainly due o hospi alisa ion.
Howe e , mo e han one- ou h o he pa ien s we e
submi ed o a su gical in e en ion a he EoL, he mos
common p ocedu e and he second mos p e alen
indi idual indica o . Al hough pallia i e su gical in e -
en ions in ad anced s ages and complica ions om a
Open access
4Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
Table 1 Pa ien s cha ac e is ics
Cha ac e is ics
All pa ien s
(n=92 155)
Pa ien s who died
wi h ACCEoL
(n=65 564)
Pa ien s who died
wi hou ACCEoL
(n=26 591)
Tes s (χ2/Mann-
Whi ney (M- W))
Pa ien
and heal h
condi ion
Sex (% male) 61.9 62.2 61.2 χ2
p=0.004
Agea dea h Median in yea s
(IQR)
73 (62–81) 72 (61–80) 75 (65–83) M- W p<0.001
18–39 (%) 2.0 2.3 1.1 M- W p<0.001
40–49 (%) 5.2 5.7 4.0
50–59 (%) 13.1 13.9 11.1
60–69 (%) 21.4 22.5 18.7
70–79 (%) 29.5 29.8 28.6
80–89 (%) 24.7 22.5 29.9
≥90 (%) 4.3 3.3 6.6
P ima y cance
ype
Lung (%) 15.9 14.4 16.5 χ2p<0.001
Colo ec al (%) 10.2 9.5 8.6
Gas ic (%) 8.7 8.5 6.7
P os a e (%) 8.5 6.7 9.8
B eas (%) 5.8 4.7 7.1
Haema ological (%) 11.9 12.4 8.0
O he (%) 41.3 43.7 43.4
Deyo- Cha lson Como bidi y Index,
median sco e (IQR)
8 (4–9) 8 (4–9) 8 (4–9) M- W NS
Less han mild: 2
(%)
11.6 10.4 14.3 M- W p<0.001
Mild: 3–4 (%) 16.1 15.9 16.7
Mode a e: 5–6 (%) 8.2 8.8 6.9
Se e e:>6 (%) 64.1 64.9 62.1
Cance as main diagnosis o las
hospi al admission (%)
65.9 66.8 63.6 χ2p<0.001
Me as a ic disease (%) 53.0 52.1 55.4 χ2p<0.001
Yea o dea h (%) 2010 16.0 15.9 16.1 M- W NS
2011 16.4 16.5 16.2
2012 16.4 16.4 16.3
2013 16.8 16.9 16.7
2014 17.0 17.0 17.1
2015 17.4 17.3 17.5
Con inued
Open access
5
Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953 Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
Cha ac e is ics
All pa ien s
(n=92 155)
Pa ien s who died
wi h ACCEoL
(n=65 564)
Pa ien s who died
wi hou ACCEoL
(n=26 591)
Tes s (χ2/Mann-
Whi ney (M- W))
Hospi al
whe e pa ien
died
Hospi al pallia i e ca e se ice (% yes) 65.8 67.2 62.5 χ2p<0.001
Median yea s o exis ence (IQR) 4.6 (2.3–8.6) 4.6 (2.4–8.4) 4.6 (2.2–9.0) M- W NS
Qua ile 1 (0.003–2.3 yea s) (%) 16.5 16.5 16.5 χ2p<0.001
Qua ile 2 (2.3–4.6 yea s) (%) 16.4 17.1 14.8
Qua ile 3 (4.6–8.6 yea s) (%) 16.5 17.0 15.2
Qua ile 4 (8.6–23.0 yea s) (%) 16.4 16.7 15.9
Medical oncology depa men (% yes) 92.9 93.5 91.6 χ2p<0.001
Hospi al ype (% cance cen e) 14.7 15.2 13.4 χ2p<0.001
Hospi al dimension, median no. beds
(IQR)
380 (319–568) 380 (319–570) 380 (319–565) M- W p=0.011
Qua ile 1 (6–319 beds) (%) 29.7 30.3 28.3 M- W p<0.001
Qua ile 2 (320–380 beds) (%) 21.0 20.2 22.8
Qua ile 3 (381–568 beds) (%) 24.5 23.6 26.5
Qua ile 4 (568–1299 beds) (%) 24.8 25.8 22.4
Heal h egion (HR) Lisbon and Tagus
Valley (%)
40.8 41.4 39.4 χ2p<0.001
No h (%) 31.4 30.5 33.8
Cen e (%) 19.1 19.7 17.7
Alen ejo and
Alga e (%)
8.6 8.4 9.2
Region whe e
pa ien li ed
Pallia i e ca e uni beds, median no o
pallia i e ca e uni beds/100 dea hs- yea ,
pe HR (IQR)
1.00
(0.81–1.22)
1.00
(0.81–1.22)
1.00
(0.80–1.22)
M- W NS
Qua ile 1 (0–0.8) (%) 29.8 29.5 30.4 χ2p<0.001
Qua ile 2 (0.8–1.0) (%) 22.4 22.2 23.2
Qua ile 3 (1.0–1.2) (%) 23.5 24.1 22.1
Qua ile 4 (1.2–19.4) (%) 24.3 24.2 24.4
Home pallia i e ca e se ices (% yes) 11.6 11.5 11.8 NS
Median yea s o exis ence (IQR) 5.3 (2.3–15.8) 5.5 (2.3–15.8) 4.9 (2.2–15.6) NS
Qua ile 1 (0.003–2.3 yea s) (%) 2.9 2.8 3.0 NS
Qua ile 2 (2.3–5.4 yea s) (%) 2.9 2.8 3.0
Qua ile 3 (5.4–15.9 yea s) (%) 2.9 2.9 2.9
Qua ile 4 (15.9–59.8 yea s) (%) 2.9 3.0 2.7
Eu opean Dep i a ion Index, median
sco e (IQR)
−1.24
(−2.34-(−0.01))
−1.24
(−2.34–0.00)
−1.24
(−2.34-(−0.02))
NS
Qua ile 1 (leas dep i ed: −7.31 o
−2.34) (%)
25.5 25.6 25.2 NS
Qua ile 2 (-2.34 o −1.24) (%) 24.5 24.5 24.6
Qua ile 3 (-1.24 o −0.01) (%) 25.0 24.8 25.4
Qua ile 4 (mos dep i ed: −0.01 o
13.47) (%)
25.0 25.1 24.8
U banisa ion le el P edominan ly
u ban (%)
69.0 69.0 68.8 NS
Mid- u ban (%) 15.2 15.0 15.6
P edominan ly u al
(%)
15.8 15.9 15.6
ACCEoL, agg essi eness o cance ca e a he end o li e; NS, non- signi ican .
Table 1 Con inued

Open access
6Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
p ima y umou esec ion p ocedu e wi h cu a i e in en
migh be causes, u he s udies should ocus on his indi-
idual ou come. We ound a ela i ely low p e alence o
chemo he apy adminis a ion in he las 14 days o li e,
compa ed wi h p e ious epo s (7% s 2%–24%).6–9
Howe e , his indica o migh be unde es ima ed since i
did no measu e he adminis a ion o o al agen s.
B eas cance was he p ima y cance ype associa ed
wi h he lowes a e o ACCEoL, p obably due o be e
knowledge o clinical ajec o ies o disease sub ypes. On
he o he hand, we ound ha pa ien s wi h gas oin es-
inal and haema ological malignancies a e a inc eased
isk o ACCEoL. In pa ien s wi h gas oin es inal malig-
nancies, i may be due o highe a e o pos ope a i e
complica ions in ea ly s ages and diges i e haemo hage
o malignan obs uc ions in la e s ages.26 Pa ien s wi h
haema ological malignancies ep esen a subg oup wi h
mo e unp edic able p ognosis and highe pe cen age o
cu a i e in en ea men s, and he e o e, as p e iously
epo ed,7 10 associa ed wi h highe ACCEoL. Also as
p e iously epo ed,6 7 pa ien s wi h highe DCCI we e a
inc eased isk o ACCEoL, ce ainly ela ed wi h highe
complexi y o ca e. As expec ed, pa ien s wi h me as a ic
disease we e less likely o expe ience ACCEoL, mainly
when cance was he main diagnosis o las hospi al
admission, sugges ing ha hospi alisa ion due o la e
s age disease p og ession is ecognised. Despi e ends
emaining unchanged in he o e all sample, ACCEoL is
inc easing o e ime in me as a ic disease. This migh be
explained by he scien i ic ad ances on sys emic an ineo-
plas ic ea men s, mos ly expe ienced in ad anced s ages.
Aligned wi h li e a u e,7 9 age was he mos in luen ial
Figu e 1 P e alence o composi e and indi idual indica o s o ACCEoL (%), o e all and by disease s age. ACCEoL,
agg essi eness o cance ca e a he end o li e; Immuno h, Immuno he apy; Biol, biological.
Open access
7
Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953 Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
Table 2 Fac o s associa ed wi h ACCEoL: mul ile el logis ic eg ession model
Fac o s
Unadjus ed
Adjus ed
(n=57 683)
Adjus ed (subg oup o
pa ien s wi h me as a ic
disease) (n=31 199)
NOR (95% CI) P alue OR (95% CI) P alue OR (95% CI) P alue
Pa ien
and heal h
condi ion
Sex ( emale— e e ence) 92 155
Male 57 071 1.04 (1.01 o 1.08) <0.01 0.97 (0.93 o 1.01) 0.12 0.93 (0.88 o 0.98) 0.01
Age a dea h (in yea s) 92 155
18–39 1 800 Re e ence <0.001 Re e ence <0.001 Re e ence <0.001
40–49 4 790 0.71 (0.61 o 0.81) 0.81 (0.68 o 0.97) 0.78 (0.62 o 0.97)
50–59 12 038 0.61 (0.54 o 0.70) 0.70 (0.59 o 0.82) 0.65 (0.53 o 0.80)
60–69 19 745 0.59 (0.52 o 0.67) 0.65 (0.55 o 0.76) 0.57 (0.47 o 0.70)
70–79 27 147 0.51 (0.45 o 0.58) 0.54 (0.46 o 0.63) 0.48 (0.39 o 0.59)
80–89 22 718 0.37 (0.33 o 0.42) 0.40 (0.34 o 0.46) 0.36 (0.29 o 0.44)
≥90 3 917 0.25 (0.22 o 0.28) 0.27 (0.23 o 0.32) 0.23 (0.18 o 0.30)
P ima y cance ype 92 155
Lung 14 695 Re e ence <0.001 Re e ence <0.001 Re e ence <0.001
Colo ec al 9 419 1.27 (1.19 o 1.34) 1.55 (1.44 o 1.67) 1.17 (1.07 o 1.28)
Gas ic 8 024 1.45 (1.36 o 1.54) 1.73 (1.60 o 1.87) 1.47 (1.33 o 1.62)
P os a e 7 825 0.81 (0.77 o 0.86) 1.04 (0.97 o 1.13) 1.08 (0.97 o 1.20)
B eas 5 346 0.77 (0.72 o 0.82) 0.83 (0.76 o 0.91) 0.72 (0.65 o 0.80)
Haema ological 11 013 1.77 (1.67 o 1.88) 1.73 (1.60 o 1.87) ·
O he 38 046 1.13 (1.08 o 1.17) 1.29 (1.23 o 1.36) 1.12 (1.04 o 1.20)
Deyo- Cha lson Como bidi y
Index
92 155
Less mild: 2 10 646 Re e ence <0.001 Re e ence <0.001 · ·
Mild: 3–4 14 864 1.31 (1.24 o 1.38) 1.31 (1.22 o 1.40) ·
Mode a e: 5–6 7 590 1.73 (1.62 o 1.84) 1.68 (1.54 o 1.84) ·
Se e e:>6 59 055 1.43 (1.37 o 1.49) 2.20 (2.02 o 2.38) ·
Cance as main diagnosis
o las hospi al admission
(no— e e ence)
92 155
Yes 60 732 1.15 (1.12 o 1.19) <0.001 1.04 (0.99 o 1.08) 0.10 0.88 (0.83 o 0.94) <0.001
Me as a ic disease
(no— e e ence)
92 155
Yes 48 870 0.88 (0.85 o 0.90) <0.001 0.54 (0.51 o 0.58) <0.001 · ·
Yea o dea h (con inuous) 92 155 1.00 (0.99 o 1.01) 0.71 1.01 (1.00 o 1.03) 0.13 1.03 (1.01 o 1.05) 0.01
Con inued
Open access
8Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
ac o on ACCEoL, wi h he oldes pa ien s ha ing 73%
lowe odds o ecei ing ACCEoL compa ed wi h he
younges . Male sex in me as a ic pa ien s was associa ed
wi h dec eased isk o ACCEoL, a dispa i y no clea ly
unde s ood and he opposi e o he epo ed in o he
con inen s,7 27 maybe esul ing om cul u al di e ences.
Conside ing en i onmen al ac o s, dea h a a cance
cen e and dea h a a hospi al wi h MOD we e associa ed
Fac o s
Unadjus ed
Adjus ed
(n=57 683)
Adjus ed (subg oup o
pa ien s wi h me as a ic
disease) (n=31 199)
NOR (95% CI) P alue OR (95% CI) P alue OR (95% CI) P alue
Hospi al
whe e
pa ien
died
Hospi al pallia i e ca e
se ice
67 262
No exis ence 23 001 Re e ence <0.001 Re e ence 0.18 Re e ence 0.42
0.003–2.3 yea s o exis ence 11 075 1.15 (1.09 o 1.20) 1.02 (0.94 o 1.11) 1.02 (0.92 o 1.13)
2.3–4.6 yea s o exis ence 11 038 1.32 (1.26 o 1.39) 1.10 (1.00 o 1.22) 1.09 (0.96 o 1.23)
4.6–8.6 yea s o exis ence 11 076 1.28 (1.22 o 1.34) 1.05 (0.92 o 1.18) 1.02 (0.87 o 1.18)
8.6–23.0 yea s o exis ence 11 063 1.20 (1.14 o 1.26) 0.97 (0.79 o 1.20) 1.01 (0.79 o 1.30)
Medical Oncology
Depa men
(no— e e ence)
62 465
Yes 58 055 1.31 (1.23 o 1.40) <0.001 1.29 (1.02 o 1.63) 0.03 1.26 (0.98 o 1.62) 0.07
Hospi al ype
(Gene al hospi al— e e ence)
92 155
Cance cen e 13 536 1.16 (1.12 o 1.21) <0.001 1.31 (1.01 o 1.72) 0.04 1.38 (1.04 o 1.83) 0.03
Hospi al dimension 57 683
6–319 beds 17 156 Re e ence <0.001 Re e ence 0.20 Re e ence 0.10
320–380 beds 12 104 0.83 (0.79 o 0.87) 1.11 (1.00 o 1.24) 1.12 (0.99 o 1.28)
381–568 beds 14 123 0.83 (0.79 o 0.87) 1.08 (0.96 o 1.24) 1.14 (0.98 o 1.32)
568–1299 beds 14 300 1.08 (1.03 o 1.13) 1.03 (0.88 o 1.21) 1.10 (0.91 o 1.33)
Region
whe e
pa ien
li ed
Pallia i e ca e uni beds* 92 155
0–0.8 beds/100 dea hs, pe HR 27 427 Re e ence <0.01 Re e ence 0.24 Re e ence 0.13
0.8–1.0 beds/100 dea hs, pe
HR
20 687 0.96 (0.91 o 1.00) 0.97 (0.92 o 1.02) 0.95 (0.88 o 1.02)
1.0–1.2 beds/100 dea hs, pe
HR
21 675 1.13 (1.08 o 1.18) 1.05 (0.97 o 1.12) 1.10 (1.00 o 1.21)
1.2–19.4 beds/100 dea hs,
pe HR
22 366 0.98 (0.94 o 1.02) 0.96 (0.88 o 1.04) 0.94 (0.84 o 1.05)
Home pallia i e ca e eam 92 155
No exis ence 81 438 Re e ence 0.06 · · · ·
0.003–2.3 yea s o exis ence 2 662 0.93 (0.85 o 1.01) · ·
2.3–5.4 yea s o exis ence 2 663 0.92 (0.85 o 1.00) · ·
5.4–15.9 yea s o exis ence 2 662 1.00 (0.91 o 1.08) · ·
15.9–59.8 yea s o exis ence 2 660 1.07 (0.98 o 1.17) · ·
Eu opean Dep i a ion Index 85 688
leas dep i ed: −7.31 o −2.34 21 830 Re e ence 0.41 · · · ·
−2.34 o −1.24 21 024 0.98 (0.94 o 1.02) · ·
−1.24 o −0.01 21 415 0.96 (0.92 o 1.00) · ·
mos dep i ed: −0.01 o 13.47 21 419 1.00 (0.95 o 1.04) · ·
U banisa ion le el 85 790
P edominan ly u ban 59 170 Re e ence 0.32 ·· · ·
Mid- u ban 13 033 0.96 (0.92 o 1.00) · ·
P edominan ly u al 13 587 1.02 (0.97 o 1.06) · ·
*Median numbe o pallia i e ca e uni beds/100 dea hs- yea , pe heal h egion.
ACCEoL, agg essi eness o cance ca e a he end o li e; HR, heal h egion.
Table 2 Con inued
Open access
9
Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953 Ma ins- B ancoD, e al. ESMO Open 2020;5:e000953. doi:10.1136/esmoopen-2020-000953
wi h highe isk o ecei ing ACCEoL. These indings
migh be due o highe complexi y o cases ea ed in
cance cen es o easie access o an ineoplas ic ea -
men s and clinical ials when MOD exis s a he hospi al
o admission. On he o he hand, and in con as wi h wha
li e a u e epo s,28 ou s udy showed no associa ion o
a ailabili y o hPCS and educ ion o ACCEoL. This esul
migh be in luenced by he ac ha hospi als wi h hPCS
a e o en hose wi h a case mix o mo e complex pa ien s,
wi h highe isk o ACCEoL. I also may be because hPCS
e ec was measu ed a a hospi al- le el, since he HMD
did no p o ide indi idual da a on hPCS in e en ion (we
could only measu e he exis ence o hPCS a he hospi al
whe e a gi en pa ien died). Mo eo e , la e e e al o
limi ed human esou ces in hPCS could hinde hei
impac on ACCEoL. The e was no e ec o egion- le el
cha ac e is ics on ACCEoL, as opposed o o he s udies.7
This is a obus na ionwide s udy ha used mul i a ia e
me hods o con ol o con ounde s a di e en le els.
Howe e , he use o ou inely- collec ed da a gene a ed
o adminis a i e pu poses does no allow o conclude
o wha ex en he ACCEoL was adequa e o inadequa e
o each indi idual, as he HMD does no con ain he
cause o dea h o he se ing o he an icance ea men
(cu a i e s pallia i e). To o e come his limi a ion, we
planned he subg oup analysis o me as a ic disease as
a sensi i i y analysis, since he ea men in en in hese
pa ien s is mos ly pallia i e. We expanded he Ea le e al’s
amewo k,6 bu eme gency depa men isi s we e no
included in his s udy, hus he p e alence o ACCEoL
could be e en highe han we es ima ed (measu emen
bias). In con as , he s udy was es ic ed o people who
died in hospi al, who a e likely o ecei e highe ACCEoL
han hose who died elsewhe e (selec ion bias).
CONCLUSIONS
This s udy un a elled impo an da a on ACCEoL and
associa ed isk ac o s, expanding he ea lie amewo k
o measu ing ACCEoL. We con i med ha clinical ac o s
ela ed wi h a be e unde s anding o disease cou se a e
associa ed wi h ACCEoL educ ion. In con as , we iden-
i ied g oups o pa ien s a inc eased isk o ACCEoL
such as pa ien s wi h mo e como bidi ies, gas oin es inal
and haema ological malignancies. The e o e, clinicians
should seek o be e in eg a ion o p ognos ic es ima-
ions wi h adequa e iming o an icipa ed discussion
o pa ien s and amilies’ p e e ences and expec a ions,
pa icula ly wi hin he high- isk g oups iden i ied by ou
s udy. E o s should be made o empowe men and ein-
o cemen o he g owing numbe o hospi al and home
PCS, wi h human esou ces, ea lie e e al and in eg a-
ion in o a comp ehensi e cance ca e plan, mainly in
cance cen es and hospi als wi h MOD. The s udy ecog-
nised need o u he esea ch on he impac o he
social and clinical communi y suppo on hospi alisa ion
a es, de e minan s o su ge y p ocedu e, managemen
o haema ological malignancies a he EoL, and highe
ACCEoL in pa ien s who deceased a cance cen es o
hospi als wi h MOD.
Au ho a ilia ions
1Cicely Saunde s Ins i u e o Pallia i e Ca e, Policy and Rehabili a ion, King’s College
London, London, UK
2Medical Oncology Depa men , Ins i u o Po uguês de Oncologia de Lisboa
F ancisco Gen il, Lisbon, Po ugal
3NOVA Na ional School o Public Heal h, Public Heal h Resea ch Cen e,
Uni e sidade NOVA de Lisboa, Lisbon, Po ugal
4Comp ehensi e Heal h Resea ch Cen e, Uni e sidade NOVA de Lisboa, Lisbon,
Po ugal
5Medical Oncology Depa men , Ins i u o Po uguês de Oncologia de Coimb a
F ancisco Gen il, Coimb a, Po ugal
6Hospi al Pallia i e Ca e Suppo Team, Ins i u o Po uguês de Oncologia de Lisboa
F ancisco Gen il, Lisbon, Po ugal
7Pallia i e Ca e Se ice, Ins i u o Po uguês de Oncologia do Po o F ancisco Gen il,
Po o, Po ugal
8Depa men o Public Heal h and Fo ensic Sciences and Medical Educa ion,
Uni e si y o Po o Medical School, Po o, Po ugal
9EPIUni , Ins i u e o Public Heal h, Uni e si y o Po o, Po o, Po ugal
10Facul y o Medicine, Uni e si y o Coimb a, Coimb a, Po ugal
Twi e Diogo Ma ins- B anco @DMB anco and Ba ba a Gomes @B_Gomes_
Acknowledgemen s This p ojec was pe o med as pa o he DINAMO P ojec ,
which aims o enhance ad anced aining and esea ch o op imise home pallia i e
ca e in Po ugal (P incipal In es iga o – Ba ba a Gomes, Scien i ic Di ec o – I ene
J. Higginson, o he membe s – P.L. Fe ei a, H. Aguia , A.F. Lace da, V.P. Sa men o,
D. Soa es, R. Caná io, M. de B i o, C. Ribei o, D. Ma ins- B anco). HMD was p o ided
by Po uguese Heal h Sys em Cen al Adminis a ion. Eu opean Dep i a ion Index
was ob ained by cou esy o one o i s au ho s – Ana Isabel Ribei o. Scien i ic and
ins i u ional suppo om he Po uguese Socie y o Oncology. This s udy was
possible also due o he pa icipa ion o all he cen es and eams who answe ed
o he na ionwide su ey. The s udy was pa ly p esen ed as a mee ing/con e ence
abs ac a : Ame ican Socie y o Clinical Oncology, Pallia i e and Suppo i e Ca e
in Oncology Symposium 2018, San Diego, Cali o nia, USA; and Eu opean Socie y o
Medical Oncology Annual Mee ing 2017, Mad id, Spain.
Con ibu o s All au ho s ha e pa icipa ed in he de elopmen o he s udy, da a
collec ion, analysis, and in e p e a ion o he esul s, w i ing and e iewing o he
manusc ip and app o e he inal ex .
Funding This wo k was suppo ed by he Calous e Gulbenkian Founda ion as pa
o he DINAMO P ojec (g an numbe 127 988 o BG); and he Núcleo Regional do
Sul – Liga Po uguesa Con a o Canc o (43/2015 and 35/2016 o DM- B).
Compe ing in e es s The i s au ho decla es hono a ia om Me ck Sha p &
Dohme, Angelini, P ize , As aZeneca, No a is, and B is ol- Mye s Squibb, and
mee ing/ a el g an s om Janssen, Roche, Labo a ó ios Vi ó ia, No a is, Gilead
Sciences, Ipsen, Me ck Sha p & Dohme, and Pie e Fab e.
Pa ien consen o publica ion No equi ed.
E hics app o al We used anonymised ou inely collec ed da a hence no e hical
app o al was equi ed.
P o enance and pee e iew No commissioned; ex e nally pee e iewed.
Da a a ailabili y s a emen Da a a e a ailable on easonable eques . Da a may
be ob ained om a hi d pa y and a e no publicly a ailable. All da a ele an o
he s udy a e included in he a icle o uploaded as supplemen a y in o ma ion.
Raw da a om HMD was p o ided by he Po uguese Heal h Sys em Cen al
Adminis a ion. S udy p o ocol, p og amming code and da a om he na ional
su ey o he hospi al adminis a ion boa ds and di ec o s o PCS is s o ed by he
co esponding au ho . Lis s o bo oughs classi ied by heal h egion and u banisa ion
le el is based on Census 2011 da a. Eu opean Dep i a ion Index was ob ained by
cou esy o one o i s au ho s.
Supplemen al ma e ial This con en has been supplied by he au ho (s). I has
no been e ed by BMJ Publishing G oup Limi ed (BMJ) and may no ha e been
pee - e iewed. Any opinions o ecommenda ions discussed a e solely hose
o he au ho (s) and a e no endo sed by BMJ. BMJ disclaims all liabili y and
esponsibili y a ising om any eliance placed on he con en . Whe e he con en
includes any ansla ed ma e ial, BMJ does no wa an he accu acy and eliabili y