.........................................................................................................
The Eu opean Jou nal o Public Heal h, Vol. 29, No. 2, 267–272
ßThe Au ho (s) 2018. Published by Ox o d Uni e si y P ess on behal o he Eu opean Public Heal h Associa ion.
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/
4.0/), which pe mi s un es ic ed euse, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
doi:10.1093/eu pub/cky215 Ad ance Access published on 10 Oc obe 2018
.........................................................................................................
Socioeconomic di e ences in heal hy and disease- ee
li e expec ancy be ween ages 50 and 75: a mul i-coho
s udy
Jenny Head
1
, Holend o Singh Chungkham
2
, Ma in Hyde
3
, Paola Zanino o
1
,
K is ina Alexande son
4
, Sa i S enholm
5,6
, Paula Salo
7,8
, Mika Ki ima
¨ki
1,9
, Ma cel Goldbe g
10
,
Ma ie Zins
10
, Jussi Vah e a
5
, Hugo Wes e lund
11
1 Depa men o Epidemiology and Public Heal h, Uni e si y College London, London, UK
2 No h-Eas Cen e, Indian S a is ical Ins i u e, Tezpu , India
3 Cen e o Inno a i e Ageing, Uni e si y o Swansea, Swansea, UK
4 Di ision o Insu ance Medicine, Depa men o Clinical Neu oscience, Ka olinska Ins i u e , S ockholm, Sweden
5 Depa men o Public Heal h, Uni e si y o Tu ku and Uni e si y Hospi al Tu ku, Tu ku, Finland
6 School o Heal h Sciences, Uni e si y o Tampe e, Tampe e, Finland
7 Depa men o Psychology, Uni e si y o Tu ku, Tu ku, Finland
8 Finnish Ins i u e o Occupa ional Heal h, Tu ku, Finland
9 Clinicum, Facul y o Medicine, Uni e si y o Helsinki, Helsinki, Finland
10 Inse m, Popula ion-based Coho s Uni -UMS 011, Villejui , F ance
11 S ess Resea ch Ins i u e, S ockholm Uni e si y, S ockholm, Sweden
Co espondence: Jenny Head, Depa men o Epidemiology and Public Heal h, Uni e si y College London, 1 - 19
To ing on Place, London WC1E 6BT, UK, Tel: +44 207 679 1683, Fax: +44 207 813 0242, e-mail: [email p o ec ed]
Backg ound: The e a e s iking socioeconomic di e ences in li e expec ancy, bu less is known abou inequali ies
in heal hy li e expec ancy and disease- ee li e expec ancy. We es ima ed socioeconomic di e ences in heal h
expec ancies in ou s udies in England, Finland, F ance and Sweden. Me hods: We es ima ed socioeconomic
di e ences in heal h expec ancies using da a d awn om epea ed wa es o he ou coho s o wo
indica o s: (i) sel - a ed heal h and (ii) ch onic diseases (ca dio ascula , cance , espi a o y and diabe es).
Socioeconomic posi ion was measu ed by occupa ional posi ion. Mul is a e li e able models we e used o
es ima e heal hy and ch onic disease- ee li e expec ancy om ages 50 o 75. Resul s: In all coho s, we ound
inequali ies in heal hy li e expec ancy acco ding o socioeconomic posi ion. In England, bo h women and men in
he highe posi ions could expec 82–83% o hei li e be ween ages 50 and 75 o be in good heal h compa ed o
68% o hose in lowe posi ions. The igu es we e 75% compa ed o 47–50% o Finland; 85–87% compa ed o
77–79% o F ance and 80–83% compa ed o 72–75% o Sweden. Those in highe occupa ional posi ions could
expec mo e yea s in good heal h (2.1–6.8 yea s) and wi hou ch onic diseases (0.5–2.3 yea s) om ages 50 o 75.
Conclusion: The e a e inequali ies in heal hy li e expec ancy be ween ages 50 and 75 acco ding o occupa ional
posi ion. These esul s sugges ha educing socioeconomic inequali ies would make an impo an con ibu ion o
ex ending heal hy li e expec ancy and disease- ee li e expec ancy.
.........................................................................................................
In oduc ion
In Eu ope, li e expec ancy con inues o ise
1
and disabili y a es in
la e li e ha e been alling in Eu ope and he USA since he
1980s.
2,3
Howe e , he e a e conce ns ha inc eases in heal hy,
disease- ee yea s o li e (heal h expec ancies), a e no longe
keeping pace wi h hose in li e expec ancy, leading o an
expansion o ime wi h mo bidi y.
4,5
Fu he mo e, he e a e
s iking socioeconomic di e ences in li e expec ancy, bu he
ex en o which hese also exis in heal h expec ancies emains
unclea .
1
The ew s udies ha ha e examined socioeconomic inequali ies in
heal h expec ancies ound ha hose in lowe socioeconomic g oups
a e doubly disad an aged by sho e li e expec ancy and mo e yea s
spen in ill heal h.
6,7
Resul s showed ha people wi h highe
educa ion no only li e longe bu spend mo e yea s in be e
heal h han hose wi h a lowe educa ion.
8–12
E en ewe s udies
ha e in es iga ed occupa ional socioeconomic posi ion, a mo e
p oximal measu e which may be e e lec adul hood ci cum-
s ances,
13
and hese also sugges ha he e a e socioeconomic
di e ences in heal h expec ancies.
14,15
Gi en ha many go e nmen s
expec people o ex end hei wo king li es and ha good heal h is
associa ed wi h ex ended wo king, i is impo an o s udy
socioeconomic inequali ies in heal h expec ancies. In his con ex ,
occupa ional posi ion may be a mo e ele an indica o han
educa ion.
The aim o his mul i-coho s udy is o con ibu e o knowledge
on inequali ies in heal h expec ancies a olde ages. We examined
occupa ional class di e ences in bo h heal hy li e expec ancy and
ch onic disease- ee li e expec ancy be ween he ages o 50 and
75 yea s o men and women om ou coho s wi h longi udinal
da a on heal h spanning se e al yea s.
Me hods
The da a came om ou coho s udies in England, Finland, F ance
and Sweden. People we e included om he i s obse a ion wi h
alid da a on sel - a ed heal h and ch onic diseases when hey we e
aged 50 yea s o olde .
16
We limi ed ou es ima ion o pa ial li e
267
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019
expec ancy o an uppe age o 75 as no all coho s had pa icipan s
aged 75 and olde .
Samples
The English da a we e om he i s six wa es o he English
Longi udinal S udy o Ageing (ELSA), an open-access, na ionally
ep esen a i e biennial longi udinal su ey o hose aged 50 and
o e li ing in p i a e households in England. The sample size was
11 391 people a he i s wa e in 2002.
17
We included 9213 pa ici-
pan s aged 50–75 a he i s wa e wi h alid da a on sel - a ed heal h
and ch onic diseases.
The Finnish da a came om i e wa es o he Finnish Public
Sec o s udy (FPS). The FPS, es ablished in 1997/1998, comp ises
all employees wi h 6 mon h job con ac in any yea om 1991/
2000 o 2005 in 10 owns and 5 hospi al dis ic s in Finland. Su ey
da a we e collec ed a 4-yea in e als on 103 866 coho membe s,
employed in he pa icipa ing o ganiza ions du ing he su ey yea s
1997/1998, 2000/2002, 2004, 2008 and/o 2012. Follow-up su eys
o esponden s who had e i ed o le he o ganiza ions we e
conduc ed in 2005, 2009 and 2013. Da a om 42 978 people aged
50–75 a he i s wa e wi h alid heal h measu es we e analyzed.
The F ench da a we e aken om he GAZEL Coho S udy, se up
in 1989 among E
´lec ici e
´de F ance-Gaz de F ance wo ke s, he
F ench na ional u ili y company, wi h annual wa es o da a
collec ion up o 2014. A incep ion in 1989, he GAZEL Coho
S udy included 20 625 olun ee s (15 011 men and 5614 women)
aged 35–50 yea s.
18
Da a om 18 263 people om he i s wa e hey
we e aged 50 plus and had heal h da a we e analyzed.
The da a o Sweden we e om i e wa es o he Swedish
Longi udinal Occupa ional Su ey o Heal h (SLOSH).
19
The
SLOSH coho is d awn om he Swedish Wo k En i onmen
Su ey (SWES), a c oss-sec ional, biennial su ey o a andom
s a i ied sample o employed people aged 16–64 yea s who
esponded o he Labou Fo ce Su ey in he same yea . The i s
wa e o SLOSH in 2006 was a ollow-up o esponden s o he 2003
SWES. A wa e 2 in 2008, new esponden s o he 2005 SWES we e
added, gi ing an o e all sample o 18 915. This sample was e-
su eyed in 2010, 2012 and 2014. Da a om 8186 people aged 50–
75 we e analyzed.
Ou come measu es
We de ined wo ou comes: (i) heal hy li e expec ancy using sel - a ed
heal h and (ii) ch onic disease- ee li e expec ancy.
Sel - a ed heal h: All pa icipan s we e asked abou hei heal h
s a us a each wa e. Responses we e ca ego ized in o good and
poo heal h as ollows. GAZEL used an eigh poin scale, whe e
he pa icipan indica ed whe he he /his heal h was e y good (8)
o e y bad (1). In line wi h p e ious esea ch,
20
he ou op
esponse op ions we e coded as good heal h. In ELSA, FPS and
SLOSH he ques ion used a i e poin esponse o ma , anging
om excellen / e y good o e y poo . Fo FPS and SLOSH, he
op wo esponse op ions we e g ouped and ca ego ized as ‘good
heal h’, and in ELSA, he op h ee esponse op ions we e
ca ego ized as ‘good heal h’. The ques ion wo dings o each
coho a e gi en in he online Supplemen a y appendix.
Ch onic diseases: P esence o he ollowing ch onic diseases was
asce ained o each coho by asking ‘has a doc o e e old you ha
you ha e ...’: (i) hea disease (hea a ack, co ona y hea disease,
angina, conges i e hea ailu e o o he hea p oblems), (ii) s oke
(s oke o ansien ischaemic a ack), (iii) ch onic lung disease
(ch onic b onchi is o emphysema o as hma), (i ) cance (cance
o a malignan umou o any kind excep skin cance ) and ( )
diabe es (diabe es o high blood suga ). Indi iduals we e de ined
as ha ing a ch onic disease i hey epo ed one o mo e o hese
condi ions. The p esence o ch onic disease a baseline ( i s obse -
a ion included in analysis) included any ch onic diseases epo ed
be o e he age o 50 om a ailable in o ma ion on esponden s. The
da a o Sweden on ch onic disease came om 2008 o 2014 wa es as
he 2006 wa e did no collec in o ma ion on all ch onic condi ions.
Mo ali y was asce ained om linked egis e da a wi h ollow-up
censo ed on 31 Decembe o he yea in which da a collec ion las
ook place o each coho .
Occupa ional posi ion
Occupa ional posi ion was coded in o h ee g oups acco ding o he
na ional occupa ional classi ica ion: highe , in e media e and lowe
occupa ional posi ions. In ELSA and SLOSH, occupa ional posi ion
was based on sel - epo ed job i le; in FPS and GAZEL occupa ional
posi ion was ob ained om he employe s’ eco ds.
S a is ical analyses
Cha ac e is ics o he pa icipa ing coho s a e p esen ed a he i s
obse a ion poin , which e e s o he da e each pa icipan is i s
included in his analysis.
Mul is a e li e able models we e used o es ima e heal hy li e
expec ancy and ch onic disease- ee li e expec ancy be ween he
ages o 50 and 75 (in o al 26 yea s). Fo bo h measu es, h ee
heal h s a es we e de ined: heal hy, unheal hy and dead. Fo sel -
a ed heal h, he e we e ou possible ansi ions be ween hese
heal h s a es, namely: heal hy o unheal hy (onse ), unheal hy o
heal hy ( eco e y), heal hy o dead, unheal hy o dead. Fo
ch onic disease, he e we e only h ee possible ansi ions as, by
de ini ion, eco e y was no possible. The analysis included all
s udy pa icipan s wi h a leas one measu e o heal h when hey
we e aged 50 yea s o olde . Fo each s udy, age-speci ic ansi ion
p obabili ies by sex and occupa ional posi ion we e es ima ed om
mul inomial logis ic models wi h age (in yea s), sex and occupa-
ional posi ion as co a ia es. Pa ial li e expec ancy, heal hy li e
expec ancy and ch onic disease- ee li e expec ancy om ages 50
o 75 we e hen calcula ed om hese es ima ed ansi ion
p obabili ies using a s ochas ic (mic o-simula ion) app oach.
21
Fo
each s udy, indi idual ajec o ies o a simula ed coho o 100 000
pe sons we e gene a ed wi h dis ibu ions o co a ia es a he
s a ing poin based on he obse ed s udy-speci ic p e alence by
i e yea age-g oup, sex and occupa ional posi ion. Heal h
expec ancies om age 50 o 75 we e calcula ed as he a e age
om hese ajec o ies o each occupa ional posi ion and sex.
Compu a ion o s anda d e o s and 95% con idence in e als
( om 2.5 h and 97.5 h pe cen iles) o hese mul is a e li e able
es ima es was pe o med using a boo s ap me hod wi h 500
eplica es o he whole analysis p ocess (mul inomial analysis and
simula ion s eps). Analyses we e conduc ed in SAS 9.2 using he
S ochas ic Popula ion Analysis o Complex E en s p og amme
(h p://www.cdc.go /nchs/da a_access/space.h m).
21
E hics app o al
In all coho s, pa icipan s ga e hei in o med consen o ake pa .
E hical app o al was gi en in each o he coun ies om ele an
e hical commi ees/boa ds.
Resul s
A he i s obse a ion poin when pa icipan s we e included in he
analysis, he p e alence o poo sel - a ed heal h anged om 20% in
F ance o 36% in Finland. P e alence o sel - epo ed ch onic
diseases anged om 21% in F ance o 33% in England ( able 1).
Fo all coho s, he isk o mo ali y o poo sel - epo ed heal h
om a s a e o good sel - epo ed heal h was highe o people in he
lowes occupa ional g oup. Co espondingly, he likelihood o
eco e y om poo heal h was lowe o people in he lowes occu-
pa ional g oup (Supplemen a y able S1). Risk o mo ali y om a
268 Eu opean Jou nal o Public Heal h
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019
poo heal h s a e also showed a simila g adien by occupa ional
posi ion.
Socioeconomic di e ences in heal hy li e expec ancy
The e we e occupa ional posi ion di e ences o pa ial li e
expec ancy and heal hy li e expec ancy a age 50 in all coho s
such ha people in he lowe occupa ional posi ion could expec
o li e ewe o al yea s o age 75 and o spend ewe o hese yea s in
good heal h ( able 2). These di e ences we e mo e ma ked o
heal hy li e expec ancy han pa ial li e expec ancy o all coho s.
Fo example, women in England wi h a high occupa ional posi ion
could expec o spend 83% o hei li e om 50 o 75 in good heal h
whe eas he co esponding igu e o English women in a lowe
occupa ional posi ion was 68%. Ou es ima es o occupa ional
posi ion di e ences in heal hy yea s o li e los be ween he ages o
50–75 anged om 2.1 yea s o women in he Swedish coho o
6.8 yea s o men in he Finnish coho .
Socioeconomic di e ences in disease- ee li e
expec ancy
Resul s om he mul is a e models o ch onic diseases
(Supplemen a y able S2) showed socioeconomic di e ences in
isk o ansi ion o ch onic disease o mo ali y om a disease-
ee s a e as well as highe isk o mo ali y om ha ing a ch onic
disease. Fo each coho , ou es ima es o ch onic disease- ee li e
expec ancy ( able 3) showed ha , compa ed o hose in a highe
occupa ional posi ion, men and women in he lowe occupa ional
posi ion could expec o li e ewe yea s wi hou ch onic diseases
be ween ages 50 and 75. Fo example, women in England in he
highe occupa ional g ade could expec o spend 61% o hei li e
om 50 o 75 wi hou ch onic disease. The co esponding igu e o
women in he lowe occupa ional g ade in England was 56%. Ou
es ima es o occupa ional posi ion di e ences in ch onic disease- ee
yea s o li e los be ween he ages o 50–75 anged om 0.5 yea s o
women in he F ench coho o 2.3 yea s o men in he Finnish
coho .
Sensi i i y analyses
We conduc ed se e al sensi i i y analyses. Es ima es o heal hy and
ch onic disease- ee li e expec ancy we e e y simila o occupa-
ional socioeconomic posi ion eco ded a ei he age 35 o a age
50. We in es iga ed he po en ial impac o equency o ollow-up
on ou esul s by epea ing analyses o GAZEL wi h da a om e e y
ou h wa e a he han yea ly in e als and ound e y simila
esul s. As ch onic condi ions we e conside ed o be non-
e e sible, he es ima es o p e alence o ch onic disease by age 50
ended o be highe o pa icipan s wi h mo e wa es o da a be o e
age 50. The e o e, we epea ed analyses excluding in o ma ion on
ch onic condi ions collec ed be o e he age o 50. As expec ed, his
led o highe es ima es o yea s li ed wi hou ch onic disease bu
di e ences by socioeconomic posi ion emained simila o each
coho . Finally, we ob ained a simila pa e n o esul s om
analyses including in o ma ion on ch onic diseases om bo h
egis e and sel - epo da a o he Finnish and Swedish coho s.
Discussion
This c oss-na ional mul i-coho s udy ound socioeconomic
inequali ies in bo h pa ial li e expec ancy and heal hy li e
expec ancy such ha socioeconomic di e ences in yea s o heal hy
li e los a e g ea e han di e ences in yea s o li e los be ween he
ages o 50 and 75. Compa ed o people in highe occupa ional
posi ions, hose in lowe occupa ional posi ions had a lowe
expec ancy o li e yea s un il age 75 and could expec o li e ewe
o hese yea s in good heal h. This pa e n was consis en ac oss he
ou coun ies and obse ed o bo h men and women. The e we e
also socioeconomic inequali ies in ch onic disease- ee li e
expec ancy, al hough es ima es o a e age yea s li ed wi hou
diseases we e lowe han o a e age yea s li ed in good pe cei ed
heal h.
Ou indings a e in line wi h p e ious s udies ha ha e used
educa ion as a measu e o socioeconomic inequali y ha show
ha hose wi h he leas le el o educa ion ha e he lowes heal h
expec ancies.
7–12
In ou s udy, men and women in he lowes
compa ed o he highes occupa ional posi ion could expec 2.1–
6.8 ewe yea s in good heal h and 0.5–2.3 ewe yea s li ed
wi hou ch onic disease be ween he ages o 50 and 75.
To ou knowledge, ew p e ious s udies ha e epo ed di e ences
in heal h expec ancies by occupa ional social posi ion. These s udies
also ound socioeconomic inequali ies in heal h expec ancies bu
he e we e some di e ences when compa ing inequali ies o men
and women. A p e ious s udy o socioeconomic inequali ies in
disabili y- ee li e expec ancy in he UK ound g ea e social class
inequali ies amongs women compa ed o men.
15
This is in line wi h
ou esul s o ch onic disease whe e he di e ence in ch onic
disease- ee li e expec ancy be ween he low and high occupa ional
class was 1.8 yea s o English women and 1.0 yea s o English men
bu no consis en wi h ou esul s o sel - a ed heal h whe e we
ound simila occupa ional class inequali ies in men and women. As
wi h he cu en s udy, an ea lie s udy in F ance ound simila
social class di e ences in heal hy li e expec ancy o men and
women.
14
Howe e , social class di e ences in ch onic disease- ee
li e expec ancy a age 50 we e wide o F ench women (4.4 yea s)
han o F ench men (3.0 yea s) in he ea lie s udy whe eas we
ound sligh ly wide social class di e ences o F ench men
(1.6 yea s) han o F ench women (0.5 yea s). This may e lec
me hodological di e ences as he p e ious s udy used in o ma ion
om a na ional su ey and hei measu e o ch onic disease was
based on a single ques ion.
Es ima es o absolu e socioeconomic di e ences in heal hy and
ch onic disease- ee li e expec ancy a ied ac oss he ou coho s
and we e mos ma ked in he Finnish coho , pa icula ly o sel -
a ed heal h. Al hough he e ha e been wo p e ious c oss-na ional
s udies o socioeconomic inequali ies in heal h expec ancies, hese
Table 1 P e alence (%) o sociodemog aphic and heal h cha ac-
e is ics a he i s obse a ion poin
a
ELSA FPS GAZEL SLOSH
Sample size 9213 42 978 18 263 8186
Sex (%)
Male 46.4 19.9 73.8 45.7
Female 53.6 80.1 26.2 54.3
Age-g oup (%)
50–54 21.9 70.5 94.5 41.8
55–59 24.2 24.9 3.9 24.2
60–64 18.7 4.5 1.3 23.5
65–69 18.9 0.1 0.3 10.2
70–74 16.3 – 0.1 0.3
Socioeconomic posi ion (%)
High g ade 30.0 29.9 14.1 19.3
Middle g ade 23.5 49.0 57.8 44.7
Low g ade 46.5 21.2 28.2 36.0
Sel - a ed heal h (%)
Good 75.0 63.4 79.9 77.9
Poo 25.0 36.0 20.1 22.1
Ch onic heal h condi ions
b
No 67.1 74.1 79.5 80.7
Yes 32.9 25.9 20.5 19.3
a: The i s obse a ion poin e e s o he da e each pa icipan is
o he i s ime included in he da ase .
b: P esence o ch onic heal h condi ions includes illness epo ed a
o be o e he i s obse a ion poin .
Socioeconomic di e ences in heal hy and disease- ee li e expec ancy be ween ages 50 and 75 269
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019
in es iga ed educa ional inequali ies in disabili y- ee li e
expec ancy.
8,9
As s a ed by Pongiglione e al. in hei sys ema ic
li e a u e e iew o inequali ies in heal h expec ancies, es ima es o
yea s o heal hy li e los end o a y acco ding o he de ini ion o
heal h making i di icul o compa e esul s om di e en s udies.
7
Ou inding ha es ima es o a e age yea s li ed wi hou ch onic
disease we e lowe han o a e age yea s li ed in good heal h was
consis en wi h indings om s udies wi h mul iple heal h indica o s
included in he e iew by Pongligione e al.
7
S eng hs o ou s udy include he use o da a om p ospec i e
coho s om ou di e en coun ies, wi h se e al measu emen s o
heal h o e ime, long ollow-up, and high quali y ha monised da a.
Ou mul is a e li e able me hod o es ima ing heal hy li e
expec ancy and ch onic disease- ee li e expec ancy using longi u-
dinal da a p o ides in e nally consis en esul s o each coho .
Howe e , he e we e some limi a ions. No all coho s we e
na ionally ep esen a i e. Whe eas ELSA is a na ional su ey, he
o he s we e occupa ional coho s ha we e ei he designed o be
ep esen a i e o all employees (SLOSH) o speci ic employmen
sec o s (FPS and GAZEL). Pa icipan s in longi udinal s udies
end o be heal hie han he gene al popula ion and mo e
disad an aged g oups end o be unde ep esen ed, and addi ionally
he e we e socioeconomic di e ences in a i ion a es.
22–24
Howe e , pas esea ch om ou s udies ha e shown simila
socioeconomic di e en ials in heal h o hose seen na ionally
implying ha ou esul s a e gene alizable o a wide popula ion.
25
The coho s did no ha e en i ely consis en de ini ions o sel - a ed
heal h and his could con ibu e o some o he coho di e ences in
absolu e heal hy li e expec ancy. Howe e , ela i e di e ences by
socioeconomic g oup we e consis en ac oss he coho s. We used
i e ch onic diseases, namely hea disease, s oke, ch onic lung
disease, cance and diabe es, o es ima e ch onic disease- ee li e
expec ancy. Al hough musculoskele al diso de s we e measu ed in
each coho , i was no possible o de elop a ha monized measu e o
musculoskele al diso de s ac oss he coho s. As hese diso de s a e
e y common a olde ages and ela ed o poo e unc ioning and
quali y-o -li e, u u e s udies o ch onic disease- ee li e expec ancy
should include musculoskele al diso de s.
The e we e di e ences in equency o ollow-up in e als be ween
s udies anging om annual o ou yea ly wa es o da a collec ion.
To assess he possible impac o his, we epea ed analyses o he
GAZEL coho using da a om e e y ou h wa e and ob ained
Table 2 Pa ial li e expec ancy, heal hy li e expec ancy and p opo ion o li e spen in good sel - a ed heal h be ween he ages o 50 and 75
by socioeconomic posi ion
Pa ial li e expec ancy be ween he ages o 50 and 75
Li e expec ancy 95% CI Heal hy li e
expec ancy
95% CI Unheal hy li e
expec ancy
95% CI % spen in
good heal h
Men
ELSA
High g ade 24.1 24.0, 24.4 19.8 19.4, 20.3 4.3 4.0, 4.7 82.2
Middle g ade 23.7 23.4, 24.0 18.2 17.6, 18.9 5.5 4.9, 5.9 76.8
Low g ade 22.9 22.6, 23.3 15.5 14.9, 16.1 7.4 7.0, 7.9 67.7
Highlow 1.2 4.3 3.1 14.5
FPS
High g ade 24.5 24.4, 24.7 18.5 18.2, 19.0 6.0 5.6, 6.3 75.4
Middle g ade 24.1 23.9, 24.4 15.4 15.1, 16.0 8.7 8.2, 9.1 63.8
Low g ade 23.4 23.2, 23.9 11.7 11.3, 12.3 11.7 11.2, 12.2 49.9
Highlow 1.1 6.8 5.7 25.5
GAZEL
High g ade 24.8 24.6, 25.0 21.6 21.3, 21.8 3.2 3.0, 3.4 87.1
Middle g ade 24.3 24.2, 24.4 20.5 20.3, 20.7 3.9 3.7, 4.0 84.2
Low g ade 23.8 23.6, 24.0 18.8 18.6, 19.1 5.3 5.0, 5.6 79.1
Highlow 1.0 2.8 2.1 8.0
SLOSH
High g ade 25.6 25.4, 25.9 20.4 19.8, 21.1 5.1 4.5, 5.8 79.9
Middle g ade 25.2 24.9, 25.5 19.7 19.2, 20.4 5.5 4.9, 5.9 78.2
Low g ade 25.3 24.9, 25.5 18.1 17.4, 18.6 7.2 6.6, 7.8 71.6
Highlow 0.3 2.3 2.1 8.3
Women
ELSA
High g ade 24.8 24.6, 25.0 20.6 20.0, 21.0 4.2 3.9, 4.8 83.0
Middle g ade 24.5 24.2, 24.7 19.1 18.6, 19.6 5.4 4.9, 5.9 78.1
Low g ade 23.9 23.6, 24.2 16.1 15.5, 16.7 7.8 7.3, 8.3 67.5
Highlow 0.9 4.5 3.6 15.5
FPS
High g ade 25.1 25.0, 25.3 18.9 18.7, 19.2 6.2 6.0, 6.5 75.3
Middle g ade 24.9 24.8, 25.0 16.3 16.1, 16.6 8.6 8.4, 8.8 65.5
Low g ade 24.6 24.4, 24.8 12.4 12.1, 12.8 12.2 11.8, 12.5 50.5
Highlow 0.5 6.5 6.0 24.8
GAZEL
High g ade 25.3 25.1, 25.5 21.6 21.2, 21.9 3.7 3.4, 4.1 85.4
Middle g ade 25.0 24.8, 25.2 20.5 19.9, 20.4 4.5 4.3, 4.7 82.0
Low g ade 24.7 24.5, 24.9 18.9 18.0, 18.8 5.7 5.6, 6.0 76.7
Highlow 0.6 2.7 2.0 8.7
SLOSH
High g ade 25.7 25.5, 25.9 21.3 20.7, 21.9 4.4 3.9, 5.0 82.8
Middle g ade 25.4 25.3, 25.6 20.8 20.3, 21.1 4.7 4.4, 5.0 81.7
Low g ade 25.5 25.1, 25.8 19.2 18.6, 19.6 6.3 5.8, 6.8 75.3
Highlow 0.2 2.1 1.9 7.5
270 Eu opean Jou nal o Public Heal h
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019
simila esul s. Ou s udy p o ides es ima es o heal hy li e
expec ancy and ch onic disease- ee li e expec ancy om age 50 o
75 so u u e s udies a e needed o in es iga e socioeconomic di e -
ences in yea s o li e spen in good heal h and wi hou ch onic
disease a olde ages. Finally, we we e no able o analyze changes
o e ime in heal h expec ancies by socioeconomic posi ion. Resul s
om a s udy compa ing heal h expec ancies in 1991 and 2011 in
England showed a ela i e comp ession o mo bidi y o sel -
pe cei ed heal h, wi h an inc ease in he p opo ion o li e spen
in good heal h.
26
Fu u e esea ch is needed o in es iga e
socioeconomic di e ences in changes in heal h expec ancies o e
ime.
Conclusions
A be e unde s anding o he u u e heal h o olde people is o
c ucial policy impo ance as i a ec s public expendi u e on
income, heal h and long- e m ca e needs o ageing popula ions.
27
Fu he mo e, heal hy and disease- ee li e expec ancy is also
impo an o wo k pa icipa ion, gi en he ac ions by mos
Eu opean go e nmen s o ex end wo king li es.
28
Ou s udy
highligh s he impo ance o educing social class di e ences in
heal h expec ancies as pa o e o s o achie e ha a ge .
Supplemen a y da a
Supplemen a y da a a e a ailable a EURPUB online.
Funding
This s udy was suppo ed by he Swedish Resea ch Council o Heal h
o Wo king Li e and Wel a e (2007-1762, 2009-1758 and 2012-1661);
he Academy o Finland (264944); and he UK Economic and Social
Resea ch Council [ES/K01336X/1] unde he ERA-AGE2 ini ia i e.
J.H. was pa ly suppo ed by he ESRC (ES/L0028921/1). S.S. is
suppo ed by he Academy o Finland (286294 and 294154). M.K.
was suppo ed by he UK Medical Resea ch Council (MR/K013351/
1), he Economic and Social Resea ch Council and No dFo sk, he
No dic P og amme on Heal h and Wel a e.
Table 3 Pa ial li e expec ancy, ch onic disease- ee li e expec ancy and p opo ion o li e spen wi hou ch onic disease be ween he ages
o 50 and 75 by socioeconomic posi ion
Pa ial li e expec ancy be ween he ages o 50 and 75
Li e expec ancy
a
95% CI Ch onic disease- ee
li e expec ancy
95% CI Li e expec ancy wi h
ch onic diseases
95% CI % spen wi hou
ch onic disease
Men
ELSA
High g ade 24.1 23.9, 24.3 13.8 12.7, 14.9 10.3 9.3, 11.3 57.4
Middle g ade 23.6 23.3, 24.0 13.2 11.6, 14.4 10.4 9.4, 11.8 56.1
Low g ade 23.1 22.8, 23.4 12.8 11.5, 13.8 10.3 9.4, 11.5 55.2
Highlow 1.0 1.0 0.0 2.2
FPS
High g ade 24.4 24.2, 24.7 13.7 12.9, 13.9 10.8 10.5, 11.6 55.9
Middle g ade 24.1 23.6, 24.3 12.4 11.7, 12.9 11.7 11.1, 12.3 51.3
Low g ade 23.5 23.2, 23.8 11.4 11.0, 11.9 12.0 11.4, 12.6 48.7
Highlow 0.9 2.3 1.2 7.2
GAZEL
High g ade 24.8 24.6, 25.0 15.5 15.1, 16.1 9.3 8.8, 9.7 62.5
Middle g ade 24.3 24.2, 24.4 14.9 14.6, 15.2 9.4 9.1, 9.7 61.3
Low g ade 23.8 23.6, 24.0 13.9 13.6, 14.4 9.9 9.4, 10.2 58.4
Highlow 1.0 1.6 0.6 4.1
SLOSH
High g ade 25.6 25.4, 25.9 15.0 13.6, 16.5 10.7 9.1, 12.1 58.4
Middle g ade 25.1 24.7, 25.4 14.5 13.3, 15.6 10.6 9.6, 11.8 57.8
Low g ade 25.1 24.7, 25.4 13.6 12.6, 14.5 11.6 10.6, 12.5 53.9
Highlow 0.5 1.4 0.9 4.5
Women
ELSA
High g ade 24.8 24.6, 24.9 15.2 13.8, 16.5 9.6 8.4, 10.9 61.3
Middle g ade 24.5 24.3, 24.7 15.2 13.6, 16.4 9.2 8.1, 10.8 62.2
Low g ade 24.0 23.8, 24.3 13.4 12.2, 14.6 10.6 9.4, 11.9
16
55.9
Highlow 0.8 1.8 1.0 5.4
FPS
High g ade 25.1 25.0, 25.2 13.3 12.9, 13.7 11.9 11.5, 12.2 52.8
Middle g ade 24.9 24.8, 25.0 13.5 13.1, 13.6 11.4 11.3, 11.8 54.0
Low g ade 24.5 24.4, 24.8 12.6 12.2, 13.1 11.9 11.5, 12.5 51.5
Highlow 0.6 0.7 0.0 1.3
GAZEL
High g ade 25.3 25.1, 25.5 15.7 14.8, 16.9 9.7 8.3, 10.5 61.9
Middle g ade 25.0 24.8, 25.1 15.6 15.2, 16.0 9.4 9.0, 9.8 62.4
Low g ade 24.7 24.5, 24.9 15.2 14.8, 15.7 9.5 9.0, 9.9 61.6
Highlow 0.6 0.5 0.2 0.3
SLOSH
High g ade 25.8 25.6, 26.0 16.7 15.5, 18.2 9.1 7.5, 10.3 64.9
Middle g ade 25.4 25.1, 25.6 17.0 16.0, 17.8 8.4 7.5, 9.2 66.9
Low g ade 25.4 25.0, 25.6 15.3 14.3, 16.5 10.1 8.9, 11.0 60.3
Highlow 0.4 1.4 1.0 4.6
a: Due o he mic o-simula ion app oach, he pa ial LE p esen ed in his column may di e sligh ly om hose in able2.
Socioeconomic di e ences in heal hy and disease- ee li e expec ancy be ween ages 50 and 75 271
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019

Disclaime
The sponso s had no ole in he design and conduc o he s udy;
collec ion, managemen , analysis and in e p e a ion o he da a; and
p epa a ion, e iew o app o al o his manusc ip .
Con lic s o in e es : None decla ed.
Key poin s
Few s udies ha e examined socioeconomic di e ences in
heal h expec ancies and mos o hese ha e used educa ion
as a measu e o socioeconomic posi ion.
Occupa ional socioeconomic posi ion may be a mo e
ele an measu e, pa icula ly in he con ex o ex ending
wo king li es.
The e we e inequali ies in heal hy li e expec ancy and
ch onic disease- ee li e expec ancy be ween ages 50 and 75
acco ding o occupa ional posi ion.
Men and women in low occupa ional posi ions had a lowe
expec ancy o li e yea s un il age 75 and could expec o li e
ewe o hese yea s in good heal h o wi hou ch onic
disease.
Reducing socioeconomic di e ences in heal h expec ancies
should be pa o policies aimed a ex ending he wo king
li es o people aged 50 and o e .
Re e ences
1 OECD. Heal h a a Glance: Eu ope 2014. Pa is: OECD Publishing, 2014.
2 C immins EM, Sai o Y. T ends in heal hy li e expec ancy in he Uni ed S a es,
1970-1990: gende , acial, and educa ional di e ences. Soc Sci Med 2001;52:1629–41.
3 C immins EM, Haywa d MD, Hagedo n A, e al. Change in disabili y- ee li e
expec ancy o Ame icans 70-yea s-old and olde . Demog aphy 2009;46:627–46.
4 Robine JM, Michel JP. Looking o wa d o a gene al heo y on popula ion aging.
J Ge on ol A Biol Sci Med Sci 2004;59:M590–7.
5 Jagge C, Robine JM. Heal hy li e expec ancy. In: Roge s GM, C immins EM,
edi o s. In e na ional Handbook o Adul Mo ali y. The Ne he lands: Sp inge , 2011,
551–68.
6 C immins EM, Cambois E. Social inequali ies in heal h expec ancy. In: Robine J,
Jagge C, Ma he s CD, e al., edi o s. De e mining Heal h Expec ancies. Chiches e :
John Wiley & Sons, L d, 2003.
7 Pongiglione B, De S a ola BL, Ploubidis GB. A Sys ema ic Li e a u e Re iew o
S udies Analyzing Inequali ies in Heal h Expec ancy among he Olde Popula ion.
PLoS One 2015;10:e0130747.
8 Maje IM, Nusselde WJ, Mackenbach JP, Kuns AE. Socioeconomic inequali ies in
li e and heal h expec ancies a ound o icial e i emen age in 10 Wes e n-Eu opean
coun ies. J Epidemiol Communi y Heal h 2011;65:972–9.
9 Maki N, Ma ikainen P, Eikemo T, e al. Educa ional di e ences in disabili y- ee
li e expec ancy: a compa a i e s udy o long-s anding ac i i y limi a ion in eigh
Eu opean coun ies. Soc Sci Med 2013;94:1–8.
10 Yong V, Sai o Y. A e he e educa ion di e en ials in disabili y and mo ali y
ansi ions and ac i e li e expec ancy among Japanese olde adul s? Findings
om a 10-yea p ospec i e coho s udy. JGe on olBPsycholSciSocSci2012;
67B:343–53.
11 Bel an-Sanchez H, And ade FCD. Educa ional and sex di e en ials in li e
expec ancies and disabili y- ee li e expec ancies in Sao Paulo, B azil, and u ban
a eas in Mexico. J Aging Heal h 2013;25:815–38.
12 Chiu CT, Haywa d M, Sai o Y. A Compa ison o Educa ional Di e ences on
Physical Heal h, Mo ali y, and Heal hy Li e Expec ancy in Japan and he Uni ed
S a es. J Aging Heal h 2016;28:1256–78.
13 Galoba des B, Shaw M, Lawlo DA, e al. Indica o s o socioeconomic posi ion
(pa 1). J Epidemiol Communi y Heal h 2006;60:7–12.
14 Cambois E, Labo de C, Romieu I, Robine J-M. Occupa ional inequali ies in heal h
expec ancies in F ance in he ea ly 2000s: unequal chances o eaching and li ing
e i emen in good heal h. Demog Res 2011;25:407–35.
15 Ma hews RJ, Jagge C, Hancock RM. Does socio-econo nic ad an age lead o a
longe , heal hie old age? Soc Sci Med 2006;62:2489–99.
16 S enholm S, Head J, Ki imaki M, e al. Smoking, physical inac i i y and obesi y as
p edic o s o heal hy and disease- ee li e expec ancy be ween ages 50 and 75: a
mul icoho s udy. In J Epidemiol 2016;45:1260–70.
17 S ep oe A, B eeze E, Banks J, Naz oo J. Coho p o ile: he English longi udinal
s udy o ageing. In J Epidemiol 2013;42:1640–8.
18 Goldbe g M, Lecle c A, Bonen an S, e al. Coho p o ile: he GAZEL Coho S udy.
In J Epidemiol 2007;36:32–9.
19 Magnusson Hanson LL, Theo ell T, Oxens ie na G, e al. Demand, con ol and
social clima e as p edic o s o emo ional exhaus ion symp oms in wo king Swedish
men and women. Scand J Public Heal h 2008;36:737–43.
20 Wes e lund H, Ki imaki M, Singh-Manoux A, e al. Sel - a ed heal h be o e and
a e e i emen in F ance (GAZEL): a coho s udy. Lance 2009;374:1889–96.
21 Cai LM, Haywa d MD, Sai o Y, e al. Es ima ion o mul i-s a e li e able unc ions
and hei a iabili y om complex su ey da a using he SPACE P og am. Demog
Res 2010;22:129–57.
22 Goldbe g M, Chas ang JF, Lecle c A, e al. Socioeconomic, demog aphic, occupa-
ional, and heal h ac o s associa ed wi h pa icipa ion in a long- e m epidemiologic
su ey: a p ospec i e s udy o he F ench GAZEL coho and i s a ge popula ion.
Am J Epidemiol 2001;154:373–84.
23 Banks J, Mu iel A, Smi h JP. A i ion and heal h in ageing s udies: e idence om
ELSA and HRS. Longi Li e Cou se S ud 2011;2:101–26.
24 Magnusson Hanson LL, Leinewebe C, Pe sson V, e al. Coho P o ile: he
Swedish Longi udinal Occupa ional Su ey o Heal h (SLOSH). In J Epidemiol
2018;47:691–92.
25 Melchio M, Be kman LF, Kawachi I, e al. Li elong socioeconomic ajec o y and
p ema u e mo ali y (35-65 yea s) in F ance: indings om he GAZEL Coho
S udy. J Epidemiol Communi y Heal h 2006;60:937–44.
26 Jagge C, Ma hews FE, Wohland P, e al. A compa ison o heal h expec ancies o e
wo decades in England: esul s o he Cogni i e Func ion and Ageing S udy I and II.
Lance 2016;387:779–86.
27 Cai L. The cos o an addi ional disabili y- ee li e yea o olde Ame icans: 1992-
2005. Heal h Se Res 2013;48:218–35.
28 Walke A. Commen a y: he eme gence and applica ion o ac i e aging in Eu ope.
J Aging Soc Policy 2009;21:75–93.
272 Eu opean Jou nal o Public Heal h
Downloaded om h ps://academic.oup.com/eu pub/a icle-abs ac /29/2/267/5126423 by Tampe e Uni e si y Lib a y use on 02 Ap il 2019