Uni e sidade No a de Lisboa
Ins i u o de Higiene e Medicina T opical
The ela ionship be ween con ex and heal h
inequali ies: Eu ope and Po ugal as case s udies
Inês Campos Ma os
DISSERTAÇÃO APRESENTADA PARA CUMPRIMENTO DOS REQUISITOS
NECESSÁRIOS À OBTENÇÃO DO GRAU DE DOUTOR EM SAÚDE
INTERNACIONAL, ESPECIALIDADE DE POLÍTICAS DE SAÚDE E
DESENVOLVIMENTO
JUNHO, 2017
Uni e sidade No a de Lisboa
Ins i u o de Higiene e Medicina T opical
The ela ionship be ween con ex and heal h
inequali ies: Eu ope and Po ugal as case s udies
Au o a: Inês Fe ei a Pi a de Campos Ma os
O ien ado : P o esso Giuliano Russo
Coo ien ado a: P o esso a Luzia Gonçal es
Disse ação ap esen ada pa a cump imen o dos equisi os necessá ios à ob enção do
g au de Dou o a em Saúde In e nacional, especialidade de Polí icas de Saúde e
Desen ol imen o, de aco do com o Regulamen o Ge al do 3.º Ciclo de Es udos
Supe io es Conducen es à Ob enção do G au de Dou o pelo Ins i u o de Higiene e
Medicina T opical/Uni e sidade No a de Lisboa (n.º 474/2012) publicado no Diá io da
República, 2.ª sé ie, n.º 223 de 19 de no emb o de 2012.
Apoio inancei o:
Subsídio pa a In e nos Dou o andos da Fundação pa a a Ciência e a Tecnologia
Re e ência SFRH/SINTD/94891/2013
Bolsa Fulb igh pa a In es igação em Saúde Pública da Comissão Fulb igh
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
i
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
ii
Publica ions om his disse a ion
Campos-Ma os I, Kawachi I. Social mobili y and heal h in Eu opean coun ies: does
wel a e egime ype ma e ? Social Science and Medicine. 2015;142:241-248 DOI:
h ps://doi.o g/10.1016/j.socscimed.2015.08.035
Campos-Ma os I, Sub amanian SV, Kawachi I. The ‘da k side’ o social capi al: us
and sel - a ed heal h in Eu opean coun ies. Eu opean Jou nal o Public Heal h.
2016;26(1):90-95. DOI: h p://dx.doi.o g/10.1093/eu pub/ck 089
Campos-Ma os I, Russo G, Pe elman J. Connec ing he do s on heal h inequali ies – a
sys ema ic e iew on he social de e minan s o heal h in Po ugal. In e na ional
Jou nal o Equi y in Heal h. 2016;15(1):15-26. DOI: 10.1186/s12939-016-0314-z
Campos-Ma os I, Russo G, Gonçal es L. Shi ing de e minan s o heal h inequali ies in
uns able imes: Po ugal as a case s udy. Eu opean Jou nal o Public Heal h. 2017.
DOI: h ps://dx.doi.o g/10.1093/eu pub/ckx080
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
iii

Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
i
Acknowledgemen s
To my supe iso s – Luzia Gonçal es and Giuliano Russo – and o he people who
supe ised my wo k on his hesis – I. Kawachi, S.V. Sub amanian, and J. Pe elman. To
colleagues and bosses in o he a eas o my li e – in pa icula Jo ge Nunes and Ped o
Se ano. To e e yone who ook he ime o each me, I am deeply g a e ul.
To my iends, who mos ly dis ac ed me om my wo k, and o whom I p o oundly
hank o ha .
To my amily, in pa icula o my pa en s, who ha e always suppo ed me. My lo e,
g a i ude, and admi a ion o you is ne e -ending.
To Ch is. Thanks o all he chocola e, pa ience, and an ing.
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
i
P e en ion is he hea o Public Heal h. Bu equi y is i s soul.
D Ma ga e Chan
Opening add ess a he Execu i e Boa d special session on WHO e o m, No embe
2011, when she was Di ec o -Gene al o he Wo ld Heal h O ganiza ion
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
ii
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
xi
3.1.4. Shi ing De e minan s o Heal h Inequali ies in Po ugal ...................... 107
3.1.5. Summa y ................................................................................................ 109
3.2. Limi a ions ....................................................................................................... 111
3.3. The E ec o Con ex on Heal h and on Heal h Dis ibu ion .......................... 113
3.3.1. Fi s Mechanism: Changes in O e all Heal h ........................................ 113
3.3.2. Second Mechanism: Changes in Heal h Dis ibu ion ............................ 115
3.3.3. Concep ual F amewo k .......................................................................... 117
3.3.4 Summa y ................................................................................................. 121
3.4. Applica ion o he Concep ual F amewo k ...................................................... 122
3.4.1. Social Capi al and Heal h in Eu opean Coun ies .................................. 122
3.4.2. Social Mobili y and Heal h in Eu opean Wel a e Regimes ................... 123
3.4.3. Shi ing De e minan s o Heal h Inequali ies in Po ugal ...................... 125
3.4.4. Summa y ................................................................................................ 127
3.5. Con ibu ion o Policy and Resea ch ............................................................... 129
3.5.1. Rega ding Heal h Inequali ies ................................................................ 129
3.5.2. Rega ding Eu ope .................................................................................. 130
3.5.3. Rega ding Po ugal ................................................................................ 131
3.5.4. Summa y ................................................................................................ 132
3.6. Conclusions ...................................................................................................... 133
3.7. Discussion and Conclusion Re e ences ........................................................... 135
4. APPENDICES ........................................................................................................ 141
4.1. Appendix 1: Illus a ion o how changes in heal h inequali ies a ec absolu e
and ela i e measu es – a hypo he ical example ..................................................... 143
4.2. Appendix 2: Online supplemen a y da a om he i s publica ion ................. 145
4.3. Appendix 3: Online supplemen a y da a om he hi d publica ion ............... 151
4.4. Appendix 4: Online supplemen a y da a om he ou h publica ion ............. 171

Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
x
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
x i
Lis o Figu es
Figu e 1. Wide De e minan s o Heal h Model, by Dahlg en and Whi ehead ........... 3
Figu e 2. WHO's Commission on he Social De e minan s o Heal h F amewo k ..... 5
Figu e 3. A amewo k o elucida ing he pa hways om he social con ex o heal h
ou comes and o in oducing policy in e en ions ..................................................... 29
Figu e 4. Flow o in o ma ion h ough he di e en phases o a sys ema ic e iew
p oposed by he PRISMA s a emen ............................................................................ 46
Figu e 5. Hypo he ical concen a ion cu e ................................................................. 49
Figu e 6. Fi s mechanism: changes in o e all heal h ................................................. 114
Figu e 7. Second mechanism: changes in heal h dis ibu ion ...................................... 116
Figu e 8. The impac o con ex on heal h and heal h dis ibu ion: concep ual amewo k
..................................................................................................................................... 117
Figu e 9. E ec o con ex ual social capi al on heal h dis ibu ion ............................. 123
Figu e 10. E ec s o wel a e egimes and social mobili y on heal h and heal h
dis ibu ion ................................................................................................................... 125
Figu e 11. E ec s o socioeconomic changes on heal h and heal h dis ibu ion ......... 127
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
x ii
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
x iii
Lis o Tables
Table 1. Wel a e egime ypes, hei main cha ac e is ics, and example coun ies ........ 23
Table 2. Summa y o disse a ion publica ions, de e minan s es ed, geog aphic con ex ,
ime pe iod, and main indings .................................................................................... 110
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
xix

Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
xx
Lis o Abb e ia ions
BMI Body Mass Index
CC Concen a ion Cu e
CHD Co ona y Hea Disease
CIx Concen a ion Index
CSDH Commission on he Social De e minan s o Heal h
DFID Depa men o In e na ional De elopmen
ESS Eu opean Social Su ey
EU Eu opean Union
EU-SILC Eu opean Su ey on Income and Li ing Condi ions
GDP G oss Domes ic P oduc
HI Heal h Inequali ies
HIV Human Immunode iciency Vi us
ILO In e na ional Labou O ganiza ion
OR Odds Ra io
PRISMA P e e ed Repo ing I ems o Sys ema ic Re iews and Me a-Analyses
RII Rela i e Index o Inequali y
SAH Sel Assessed Heal h
SDH Social De e minan s o Heal h
SES Socioeconomic S a us
SII Slope Index o Inequali y
SR Sys ema ic Re iew
UK Uni ed Kingdom
US Uni ed S a es
USSR Union o So ie Socialis Republics
WHO Wo ld Heal h O ganisa ion
Disse ação pa a a ob enção do g au de Dou o em Saúde In e nacional
Inês Campos Ma os, Junho de 2017
xxi
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
1
1. In oduc ion
1.1. The S udy o Heal h Inequali ies
1.1.1. Fou Decades o Poli ics and Resea ch on Heal h Inequali ies
Academia has a long adi ion o seeing i sel as an au onomous body, ee om
ideological hough and om social and poli ical con ex (1). Howe e , academic
wo k is no done in isola ion om i s su ounding wo ld. The s o y o how academic
knowledge abou heal h inequali ies (HI) and he social de e minan s o heal h (SDH)
has e ol ed in he pas ou decades is a e lec ion o how his knowledge and
dominan poli ical hough go hand in hand, some imes one pushing o wa d mo e
igo ously.
The connec ion be ween socioeconomic de e minan s and heal h has been known o
cen u ies, bu he e has been a g owing in e es in he subjec in he las ou decades.
In Eu ope, his poli ical and academic in e es was shaped by a ew landma k e en s.
The i s poli ical landma k was he publica ion o he Black Repo in he Uni ed
Kingdom (UK) in 1980 (2). This epo was se up by a Labou Sec e a y o S a e o
Heal h who was conce ned abou he dimension o mo ali y inequali ies be ween
social classes (3). The epo ound ha , despi e h ee decades o a Na ional Heal h
Se ice, HI s ill exis ed and could la gely be explained by di e ences in ma e ial
condi ions be ween social classes (2). Bu he inal documen was epo ed o a
Conse a i e go e nmen , elec ed in 1979 unde he commi men o educe public
spending, who dismissed i and ailed o p ope ly publish i (3). Despi e his, he
epo played a pi o al ole in se ing he esea ch agenda o wo decades a e i was
published (3, 4).
The indings o he Black Repo we e in o med by a body o academic wo k
published in he yea s be o e. P obably he mos no iceable indings came om he
Whi ehall s udy (3), a longi udinal s udy o ci il se an s wo king in London ha
s a ed in 1967 o analyse he ‘powe o isk ac o s and indica o s o co ona y hea
disease (CHD) o p edic mo ali y’ (p. 1165). The Whi ehall s udy showed a clea
in e se ela ionship be ween g ade o employmen and CHD mo ali y ha pe sis ed
e e a e con olling o a wide ange o ca dio ascula isk ac o s (5). As one o he
au ho s o he o iginal s udy, Si Michael Ma mo , la e commen ed in an in e iew,
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
2
his was unexpec ed, as a he ime common sense sugges ed ha hea disease was
mo e common in people wi h highe -g ade, mo e s ess ul jobs (6).
O e all, he Whi ehall s udies oge he wi h he Black Repo e lec ed a change in he
unde s anding o he de e minan s o heal h. The UK was seen as a success in e ms
o popula ion heal h imp o emen , wi h ex ensi e s a e-p o ided social suppo and
uni e sal heal hca e p o ision. These analyses showed ha , o e and abo e he di ec
e ec o he mos basic de e minan s o heal h – such as die o heal hca e – s ong
social o ces ope a ed o c ea e a heal h g adien , e en among people who had access
o all essen ial li ing condi ions.
The posi ioning o HI and SDH in he esea ch agenda led o a p oli e a ion o
empi ical analyses in he yea s ensuing he publica ion o he Black Repo . In 1991,
Dahlg en and Whi ehead p oposed he ‘Wide De e minan s o Heal h Model’,
p obably he mos widely known and used amewo k on his opic (7) ( igu e 1). This
model desc ibes he main in luences o heal h, buil in laye s, one on op o he o he .
The s uc u al en i onmen is he o e a ching laye , which includes ‘gene al socio-
economic, cul u al and en i onmen al condi ions’. This is ollowed by a laye o
li ing and wo king condi ions, which includes ac o s like employmen and
educa ion. The nex wo laye s e e o suppo om social ne wo ks and indi idual
li es yle, espec i ely. Finally, he cen al laye is made o indi idual unchangeable
ac o s, such as age and sex. This model p o ed use ul in p esen ing he main
de e minan s o heal h o b oad audiences, and was g oundb eaking when i was i s
published, as i highligh ed he impo ance o b oade socioeconomic ac o s in he
p oduc ion o heal h. I also emphasized he cumula i e na u e o he de e minan s o
heal h and p o ided a amewo k upon which o conside policy op ions, as each laye
can be ansla ed in o a le el o policy in e en ion.
In 1997, a e eigh een yea s o conse a i e ule, he UK elec ed a Labou
go e nmen . This new go e nmen was elec ed wi h a s ong commi men o educe
social inequali ies, and quickly commissioned a epo o ‘ e iew and summa ise
inequali ies in heal h in England and o iden i y p io i y a eas o he de elopmen o
policies o educe hem’ (8: p.5). This was ma e ialized wi h he publica ion o he
Acheson Inqui y in 1998, which e ealed a wide ange o HI h oughou he li e-
cou se, some e en inc easing o e he p e ious decades. The epo clea ly s a ed ha
HI we e a consequence o socioeconomic ac o s, and as such could only be ackled
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
9
esponsible o he dis ibu ion o indi idual de e minan s o heal h, such as income o
educa ion ( igu e 2) (12).
Bu how does con ex in luence heal h and i s dis ibu ion? The answe o his
ques ion has been unde mined by a numbe o challenges acing his body o
li e a u e. As an eme ging opic, one o he main issues has been he he e ogeneous
concep ualiza ion o con ex (14). Di e en au ho s ha e used di e en de ini ions o
wha con ex is, including a wha physical o geog aphical le el i ope a es –
household, neighbou hood o coun y will likely all in luence heal h, bu p obably in
di e en ways. Dide ichsen e al. (28), o example, de ine con ex as a ‘ca ch-all
ph ase used o e e o he spec um o ac o s in socie y ha canno be di ec ly
measu ed a he indi idual le el’ (p. 19), encompassing he ‘s uc u e, cul u e and
unc ion o a social sys em’ (p. 19). This ague de ini ion highligh s he di icul y o
homogenise he concep . Po a’s dic iona y o Epidemiology (33), on he o he hand,
de ines con ex as ‘ he loca ion o a pe son by ime and place’, e e ing o bo h
‘geog aphical loca ion and o g oup membe ship’ (p. 58). This de ini ion b ings o
ocus wo componen s o con ex ha a e commonly iden i ied: he physical and he
social en i onmen . These en i onmen s can no only in luence heal h, bu may also
in luence each o he (34). Fo example, he ex en o physical space a communi y has
a ailable will in luence how i s indi iduals in e ac . This dis inc ion b ings o ligh
ha , by being physical o social, con ex is no es ic ed o a geog aphical de ini ion;
con ex ual cha ac e is ics can be de ined wi hin a ne wo k o pee s who in e ac
exclusi ely online, bu who none heless sha e social no ms ha shape hei elemen s’
heal h.
Ano he common concep ualiza ion o con ex dis inguishes be ween ‘composi ional’
and ‘con ex ual’ e ec s. This dis inc ion a ose om geog aphical analyses ha hink
o place e ec s as a consequence o he cha ac e is ics o he people who eside in a
ce ain place (‘composi ional’) and o he cha ac e is ics o he place i sel
(‘con ex ual’). Howe e , as Macin y e e al. (14) and F ohlich e al. (21) a gue, his
dis inc ion is no necessa ily use ul no co ec . In ac , he e a e complex
in e dependencies be ween people (‘composi ion’) and places (‘con ex ’), as
indi iduals a e no placed a andom whe e hey li e o whe e hey wo k. As
Macin y e (35) pu i : ‘people make places and places make people’ (p. 12).

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
10
Ano he ac o hampe ing he app op ia e analysis o con ex ual in luences on heal h
is he lack o clea heo izing abou he mechanisms by which con ex ope a es (14).
This leads o a sea ch o con ex ual de e minan s o heal h ha has no basis on a
s ong heo y o how hese de e minan s ope a e. As Mi chell e al. (36) commen ,
‘lack o heo y has o en esul ed in a choice o a iables wi h which o cha ac e ize
an a ea which is guided mo e by wha is a ailable ‘o he shel ’ han by ca e ul
heo e ical conside a ion’ (p. 68). As a esul o unclea de ini ion and heo izing,
measu emen o con ex has been, in he leas , he e ogeneous.
This has also dismissed impo an ques ions such as which spa ial o ime scales a e
app op ia e. Indeed, con ex ual in luences can be measu ed as cha ac e is ics o a
s ee , neighbou hood, ci y, o coun y, jus o name a ew. Pa hways ha link hese
cha ac e is ics o popula ion heal h and HI will di e acco ding o he geog aphical
scale hey a e being measu ed a . Fo example, while a uni e sal heal hca e policy
migh be a good measu e o heal hca e access, i does no ake in o accoun egional
inequali ies in he dis ibu ion o heal hca e se ices, which can only be de ec ed wi h
a smalle scale analysis. On he o he hand, ideological iews such as acism and
o he o ms o disc imina ion may no be de ec ed a such a small scale, bu
none heless be p ominen in he coun y and ha e an impo an e ec on he heal h o
ha popula ion (37). Time scales a e also o en imes dismissed: mos analyses
measu e con ex ual exposu e and heal h ou come a he same momen in ime, bu his
is o en implausible, as exposu es can ake ime o ha e an e ec . Fo example, ai
pollu ion may ake decades o ha e an impac on adul mo ali y, and his biological
plausibili y mus be aken in o accoun (14).
Finally, as desc ibed in he i s sec ion, con ex has also been pushed aside om
esea ch as a consequence o dominan poli ical iews. As Ma ga e Tha che
amously pu i , “ he e is no such hing as socie y” (38).
Unclea de ini ion, ope a ionaliza ion, and heo izing ha e conside ably complica ed
he cons uc ion o a cohe en heo y on he in luence o con ex on heal h and HI.
This has led o impo an c i iques, e en compa isons wi h medie al medical heo y –
Slogge and Joshi (39) called he con ex ual in luence a ‘social miasma’ (p. 1473).
Howe e , despi e he weaknesses in he cu en e idence on his opic, and he
c i iques o i s exis ence as an issue a all, esea che s would end o ag ee ha whe e
people li e ma e s o hei heal h (40). In a b ie e iew o he e idence, Macin y e
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
11
e al. (14) conclude ha ‘ a he han he e being one single, uni e sal ‘a ea e ec on
heal h’ he e appea o be some a ea e ec s on some heal h ou comes, in some
popula ion g oups, and in some ypes o a eas’ (p. 128). This conclusion sugges s ha
con ex can some imes ha e li le impac on o e all popula ion heal h, and a
emendous impac on i s dis ibu ion be ween popula ion g oups, and hence on HI.
Thus, con ex appea s o ha e an impo an in luence on heal h and i s dis ibu ion
wi hin a socie y, bu cu en knowledge is conside ably hampe ed by a numbe o
issues, such as he e ogeneous concep ualiza ion and measu emen . Addi ionally,
con ex can ha e di e en e ec s depending on he indi idual, place, and ime. This
di e en ial e ec sugges s ha no only i is necessa y o ake con ex in o accoun
when s udying HI, bu ha i may be key o hei unde s anding. An app op ia e
heo e ical amewo k ha summa izes hese ela ionships would p o ide he much
needed basis on which empi ical analysis could build e idence.
1.1.4. Summa y
Resea ch in HI and he SDH has g own exponen ially in Eu ope since he 1980s,
when poli ical in e es in he opic i s eme ged. Academic and poli ical in e es in HI
e ol ed h ough his pe iod o ime, wi h a clea shi o ocus om indi idual
de e minan s, indi idualis ic me hodology and ‘ isky beha iou s’, o social p ocesses
and con ex as de e minan s o HI. This shi is e lec ed no only on dominan
poli ical hough o indus ialized socie ies, bu also on dominan academic hough .
This pa allel cou se shows how he academic discou se is no sepa a e om he wo ld
ha su ounds i , as he iews o esea che s can be shaped by dominan no ma i e
iews o he socie y hey a e in.
The indi idualis app oach ha has domina ed esea ch on HI has cons ained
knowledge and hinde ed he c ea ion o policies ha e ec i ely educe hem. This has
happened because issues o he e ogeneous concep ualiza ion, measu emen , and
heo e ical de ini ion ha e unde mined he s udy o con ex ual de e minan s, which
has opened he ield o conside able c i ique. Howe e , con ex ual de e minan s a e
una oidable, pa icula ly when explo ing HI, namely o measu emen and concep ual
issues. Ul ima ely, i is likely ha con ex has a complex di e en ial e ec ha
in e ac s wi h indi idual cha ac e is ics. This makes con ex key in he unde s anding
o HI.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
12
The no ion ha no all de e minan s o heal h a e bes concep ualized a he indi idual
le el has been called he new pa adigm o public heal h. As a new idea, i is s ill in
de eloping s age, and would bene i g ea ly om a solid heo e ical basis ha would
allow build-up o knowledge and adequa e policy choices.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
13
1.2. Measu emen o Heal h Inequali ies
1.2.1. Which De e minan s?
Fo a long ime and ac oss many coun ies, he poo ha e had wo se heal h and
sho e li es han he weal hy (41). This consis ency is seen ega dless o he majo
causes o dea h in socie y: i is ue when communicable diseases a e he main kille s,
and i is ue o non-communicable diseases (42). I is seen h ough he li e cou se,
om ges a ion o bi h, childhood, adolescence, adul hood and old age (43). I is seen
using a numbe o di e en s a is ical me hods and measu emen s, and i is seen by a
numbe o indi idual SES measu es, like income, educa ion o occupa ion (44). Bu
he s udy o con ex ual de e minan s has aised bo h concep ual and me hodological
ques ions ha canno be answe ed by adi ional iews o indi idual cha ac e is ics
and adi ional s a is ical me hods alone.
Concep ually, ecognizing con ex ual e ec s on indi idual heal h implies a shi in he
unde s anding o how isk ac o s ope a e. In adi ional epidemiology, indi idual
cha ac e is ics a e iden i ied as causes o isk ac o s o ill heal h, implying ha
in e en ions should be ocused on he indi idual. Some c i iques o his app oach
claim ha an exclusi e ocus on he indi idual can lead o coun e p oduc i e
p ocesses o ic im blaming (45). The new pa adigm o con ex ual e ec s on heal h
ecognizes ha he con ex can be, in and o i sel , a de e minan o heal h.
Addi ionally, he acknowledgemen o con ex as ha ing an in luence on heal h also
implies ha indi idual cha ac e is ics mus be amed by he con ex hey a e in. In
his sense, ha ing a ce ain amoun o money ma e s no only in absolu e e ms, bu
also conside ing he a e age weal h o e e yone else, i.e., wha you ha e ma e s, bu
wha o he s a ound you ha e ma e s oo. This implies ha HI can be c ea ed, in pa ,
o psychosocial mechanisms. Indeed, a no iceable inding in HI esea ch has been ha
HI do no occu in a h eshold e ec . Sys ema ic di e ences a e no seen jus below a
ce ain poin o income, weal h o occupa ional ank. Ra he , he e is a socioeconomic
g adien in heal h, in which he poo es ha e wo se heal h han he ones who ea n an
a e age income, who in u n ha e wo se heal h han he iches (41). This g adien is
seen o all SES de e minan s: inancial esou ces, educa ion and occupa ion. This
g adien shows ha socioeconomic de e minan s do no in luence heal h only because
hey p o ide access o essen ial esou ces, o he wise he e would be no di e ence
once hose essen ial esou ces we e p esen (44). Ra he , he e a e o he e ec s a
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
14
wo k h ough he social spec um ha p oduce sys ema ic di e ences in heal h
ou comes. These e ec s a e a consequence o psychosocial mechanisms.
While ma e ial condi ions a e widely ecognized and impo an o he heal h
g adien , as hey a e di ec ly ela ed o access o heal h p omo ing esou ces, such as
quali y ood and housing, ex ensi e esea ch has shown ha ad e se social condi ions
can di ec ly lead o ad e se biological e ec s, ega dless o access o esou ces (46).
This iew helps unde s and he socioeconomic heal h g adien . Wi hin a s uc u e o a
socie y, e e yone is benea h someone else by some measu e – be i money, p es ige,
cul u al capi al, o o he s. This social hie a chy can c ea e s ess esponses, as being
o a lowe social s a us may lead o eelings o inadequacy and lack o con ol (47).
As such, since he e is always someone be e o , e e yone su e s he consequences
o his g adien , no only he poo es . These s ess esponses ha e a di ec impac on
biological unc ioning, and can lead o mo e heal h damaging beha iou , such as
d inking and smoking. This leads o a heal h g adien ha a ec s e e yone, e en a e
basic ma e ial condi ions o a heal hy li e a e sa is ied. No iceably, a he han being
opposi e explana ions, ma e ial and psychosocial mechanisms ac oge he o help
explain he socioeconomic heal h g adien (44).
Psychosocial explana ions highligh ha , u he han being a ques ion o absolu e
po e y, HI a e also a ques ion o ela i e dep i a ion. The in e na ional glossa y o
po e y de ines ela i e po e y as he ‘absence o inadequacy o hose die s,
ameni ies, s anda ds, se ices and ac i i ies which a e common o cus oma y in
socie y’ (48: p. 169). This no ion o po e y implies ha i can be a socially de ined
concep , measu ed wi hin he g oup he indi idual is in, as i depends on wha is
‘cus oma y in socie y’.
The heal h g adien and psychosocial explana ions show ha he c ea ion o HI is
complex and a p oduc o mul iple de e minan s ope a ing a he same ime.
Impo an ly, hese de e minan s occu a bo h he indi idual and con ex ual le el
simul aneously, and nei he should be igno ed.
a) Indi idual De e minan s
His o ically, income, educa ion, occupa ion and employmen ha e domina ed he
analysis o HI as de e minan s o indi idual SES. These de e minan s a e o en used

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
15
in e changeably, bu esea ch has shown ha his is no necessa ily co ec , as hey
can e lec di e en unde lying causal p ocesses (49).
i. Educa ion
Educa ion is p obably he mos commonly used measu e o SES and has ex ensi ely
been ela ed o a ious heal h ou comes. Compa a i e analyses be ween Eu opean
coun ies show s ong associa ions o educa ion wi h cause-speci ic mo ali y (50),
sel -assessed heal h (SAH) (51), limi ing long-s anding illness (52), ch onic
condi ions (53), smoking (54, 55) and obesi y (56). As a measu e o SES, educa ion
has a numbe o impo an ad an ages: i is easy o measu e, shows high esponse
a es in su eys, is ai ly compa able ac oss coun ies, applicable o bo h wo king and
non-wo king indi iduals, ends o emain s able h ough li e, and is no likely o be
a ec ed by e e se causa ion, since i is usually de e mined in young adul hood and
emains s able h oughou li e (57). Howe e , e e se causa ion canno be comple ely
excluded, since a heal hy li e expec ancy migh induce highe in es men s in
educa ion and ill child en migh be less able o comple e educa ion (58). None heless,
analyses o compulso y schooling laws in he US and Eu ope, which ‘ o ce’ mos
people in o educa ion, ega dless o hei heal h p ospec s, sugges ha educa ion
causes be e heal h, despi e he opposi e also being ue (59-61).
Se e al mechanisms explain he pa hway linking educa ion o heal h. The e ec
seems o be media ed in pa by income and occupa ion, al hough analyses show an
educa ional g adien e en a e con olling o hese ac o s (58). The emaining
heal h di e ences can be s ongly explained by beha iou al ac o s, which in u n
seem o be a consequence o be e in o ma ion and be e cogni i e abili ies, which
a ec s he abili y o p ocess in o ma ion ega ding heal hy beha iou s and disease
managemen (62). P e e ences also seem o play a pa , as hey a y sys ema ically
ac oss educa ional g oups (62). Finally, educa ion can also p o ide an indi idual wi h
a social ne wo k o simila ly educa ed pee s, which can ha e subs an ial heal h
bene i s (41).
ii. Financial esou ces
Financial esou ces, such as income o weal h, a e also s ongly co ela ed wi h
heal h, independen ly o educa ion. They can ha e an impac on heal h o he ex en
ha hey allow indi iduals o access heal h-p oducing esou ces, such as heal hca e o
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
16
be e li ing condi ions. Howe e , assessing causali y is di icul , since heal h
s ongly inc eases one’s abili y o ea n money and se e al con ounding ac o s – such
as educa ion o p e e ences – migh de e mine bo h be e heal h and be e income.
Whe eas he e ec o heal h on income has been ex ensi ely shown in a numbe o
socie ies, he e ec o income on heal h emains mo e o an ‘open ques ion’ (41, 58).
Unlike educa ion, inancial esou ces can be mo e di icul o measu e, as people may
no be as happy o sha e his in o ma ion, he e a e many componen s in ol ed
(income, weal h, sa ings, p ope y, e c.), and compa abili y can be hampe ed by
di e en cu encies and di e ences in cu ency alue, o example. None heless,
income- ela ed heal h g adien s ha e been shown o SAH (63, 64), unc ional
limi a ions (64, 65) and smoking (55) in Eu opean coun ies.
iii. Employmen
Employmen , o lack o i , is cen al o mos adul s’ li e and has been associa ed wi h
heal h in a numbe o se ings. This associa ion is no su p ising. Fi s o all,
employmen p o ides income, which is essen ial o access o basic goods.
Unemploymen can lead no only o inancial s ain, bu also unce ain y abou he
u u e. This leads o a second mechanism: s ess. Unemploymen , wi h he unce ain y
i b ings, leads o a eeling o lack o con ol, which has been ex ensi ely associa ed
wi h ad e se heal h ou comes (66). Employmen also p o ides psychological bene i s
– like p o iding a s uc u e o he day, sel -es eem, s a us and a sense o con ibu ion
o a collec i e cause – ha a e absen in unemploymen (67). Oppo uni ies o
socialize a e also mo e common when one is employed, and social suppo and
in eg a ion ha e ex ensi ely been linked o heal h (68). Finally, unemployed
indi iduals seem o ha e an inc eased isk o heal h-damaging beha iou s, such as
smoking and d inking (69). This migh occu because people who d ink and smoke
a e mo e likely o become unemployed, because people who become unemployed
d ink and smoke mo e o deal wi h hei s ess ul si ua ion, as a consequence o a
common causal ac o , o a combina ion o any o hese. Rega dless, employmen is a
majo ac o in HI among wo king-age adul s.
Analyses in Eu opean coun ies ha e shown ha unemploymen is associa ed wi h
wo se SAH (63), ch onic heal h condi ions (70) and mo ali y (71).
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
17
i . Occupa ion
The In e na ional Labou O ganisa ion (ILO) (72) de ines occupa ion as ‘a se o jobs
whose main asks and du ies a e cha ac e ised by a high deg ee o simila i y’ (p. 1).
Acco dingly, occupa ional g ades a e usually classi ied acco ding o asks and
esponsibili ies in ol ed. P obably he mos commonly used classi ica ion is ILO’s
In e na ional S anda d Classi ica ion o Occupa ions, which de ines en majo
occupa ional g oups, om elemen a y occupa ions o manage s, de ined in e ms o
skill le el and specialisa ion equi ed o each occupa ion (73). O he classi ica ions,
such as manual and non-manual (53), adminis a o s, p o essionals, execu i es and
cle ks (74) o whi e colla and blue colla (75) a e some imes used, bu all e lec
di e en deg ees o skills ha a e equi ed o he job.
Occupa ional g ade has been s ongly associa ed wi h heal h ou comes. In Eu opean
coun ies, i has been associa ed wi h o e all mo ali y (53, 76), s oke and ischaemic
hea disease mo ali y (77), in an mo ali y (78), child heal h (78), SAH, long e m
limi a ions and ch onic condi ions (79). The Whi ehall s udies a e one o he mos
impo an con ibu ions o he unde s anding o his ela ionship. These s udies
showed ha people in highe anks had a s onge sense o con ol o e hei heal h,
hei jobs and hei li es (74), which is s ongly associa ed wi h be e heal h (66).
Howe e , occupa ion can also e lec an indi idual’s place in socie y, o a g ea e
ex en han inancial esou ces, educa ion o employmen do. Ha ing an occupa ion
will usually g an adequa e ea nings and a ce ain deg ee o job secu i y, so he heal h
di e ences ha emain can also be explained by he e ec o social s anding ( ank)
and subjec i e eelings owa ds one’s posi ion in socie y (59). In ac , a numbe o
au ho s ha e used occupa ion as a ma ke o ‘social class’ (80-82), possibly based on
he unde s anding ha occupa ion e lec s mo e han jus skill le els and
specialisa ion.
. Social Mobili y
The SES o an indi idual can ha e an impac on hei heal h a any gi en ime and he
mo emen be ween di e en social s a a can oo. Social mobili y is he p ocess o
mo ing be ween social s a a, ei he be ween gene a ions (pa en s and child en) o
wi hin he li e-cou se o he indi idual (in e - and in agene a ional social mobili y,
espec i ely).
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
18
Social mobili y has been associa ed wi h heal h in a a ie y o con ex s (83, 84), bu
he way i ope a es can be di icul o ease ou . One possibili y is ha i is an
accumula ion e ec . Powe e al. (85), o example, showed ha in a bi h coho om
he UK, he accumula ion o un a ou able social ci cums ances was mo e impo an
han social mobili y pe se in de e mining adul heal h. This means ha , when
compa ing wo people wi h he same SES, one o whom ‘mo es down’ he social
ladde and ano he who emains s able, he i s will ha e wo se heal h ou comes.
Howe e , his is no a consequence o he downwa d social mo emen i sel , a he a
consequence o ha ing a lowe SES a a gi en momen in li e. Ano he explana ion
o he associa ion be ween social mobili y and heal h is an opposi e causal e ec :
ex ensi e empi ical analysis has shown a ‘heal h selec ion e ec ’ ha pushes people
who a e unheal hie down he social ladde (84, 86). This is no su p ising, as people
who a e ill can be less capable o s udy and wo k, hus educing hei po en ial
ea nings and social posi ion. Finally, some e idence also sugges s ha social
mo emen in and o i sel has some e ec on heal h (41, 83). Su p isingly, i is no
jus he downwa d mo emen ha seems o ha e a nega i e impac , bu also upwa d
mo emen , pa icula ly wi hin sho pe iods o ime, can also ha e a dele e ious e ec
(41). This unexpec ed e ec may be a consequence o an inc ease in unheal hy
beha iou (such as smoking mo e because one has mo e a ailable income), o o
physiological and beha iou al adap a ions o a di e en social se ing.
b) Con ex ual De e minan s
A mul i ude o con ex ual cha ac e is ics ha e been analysed in he li e a u e as
po en ial de e minan s o heal h. Building on he de ini ion o con ex ou lined in he
p e ious chap e , hese can be classi ied as physical o social de e minan s – wi h he
ca ea ha his is an o e simpli ied cha ac e isa ion, as physical and social con ex ual
cha ac e is ics o en in e ac wi h each o he .
Physical de e minan s can be hough o in e ms o na u al en i onmen – ai , wa e ,
noise, g een spaces – and o he buil en i onmen – houses, oads, in as uc u es,
and anspo sys ems. Ex ensi e esea ch has shown a s ong connec ion be ween he
buil en i onmen and heal h. Examples include impac s on men al heal h (87),
physical ac i i y (88), ea ing habi s (89), obesi y (90), and d inking (91). This
in luence can wo k h ough he a ailabili y o g een spaces ha allow people o ha e
a mo e physically ac i e li e, walkabili y o neighbou hoods ha help p o ide a sa e
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
25
s a us measu es and o he a iables ha impac mo ali y (124). Mo eo e , i s
capaci y o p edic mo ali y has become inc easingly mo e consis en be ween 1990
and 2002 in he US, possibly because o be e heal h ela ed in o ma ion (125). On
he o he hand, despi e i s s ong abili y o p edic mo ali y, his ela ion migh a y
be ween gende (126), age (127), and SES (128), possibly biasing esul s o HI
analyses. Howe e , SAH has alue on op o i s abili y o p edic mo ali y: i can be
seen as a mo e comp ehensi e measu e o heal h, which allows esponden s o weigh
di e en aspec s o hei own heal h and alue hem acco ding o hei own
p e e ences (129).
An inc easingly used al e na i e o SAH has been o use ‘heal h limi a ions’ as a
gene al measu e o heal h. This measu e is also based on a su ey ques ion ha asks
esponden s whe he hey a e limi ed in hei daily ac i i ies due o a heal h condi ion.
Possible answe s usually include ‘yes, se e ely limi ed’, ‘yes, somewha limi ed’, and
‘no’. Some au ho s conside his a ‘quasi-objec i e’ indica o , mo e accu a e han
SAH (65), and some ha e used i as a p oxy measu e o disabili y (130).
HI ha e been obse ed using bo h SAH and limi a ions as an ou come measu e (65).
These measu es ely on sel - epo ing, bu HI ha e also been obse ed in mo bidi y
indica o s ha a e based on objec i e measu emen s. Fo example, se e al cance s
(al hough no all) show a socioeconomic g adien , as does he su i al a e a e
cance diagnosis (78). Measu es o physical abili y also end o show a socioeconomic
g adien (131), as does Body Mass Index (BMI) (132) and he me abolic synd ome
(133), o men ion only a ew examples.
c) Heal h Rela ed Beha iou s
Acco ding o he Global Bu den o Disease S udy, he h ee isk ac o s ha mos
con ibu e o disease bu den in wes e n and cen al Eu ope a e high blood p essu e,
obacco smoking, and high BMI (134). These e lec he ou majo beha iou s
ela ed o non-communicable diseases: ea ing, d inking, smoking and exe cising.
These ou beha iou s ha e ex ensi ely shown a social pa e n, such ha people om
lowe SES end o show less compliance wi h die a y and exe cise ecommenda ions,
d ink mo e, and smoke mo e (135).
These heal h ela ed beha iou s occu wi hin social s uc u es and con ex s ha can
acili a e o hinde hem. Fo example, while ul ima ely smoking may be a ma e o

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
26
indi idual choice, ac o s such as he economic abili y o buy ciga e es, ha ing a
mo e o less s ess ul li e, and li ing in an en i onmen whe e social no ms suppo o
shame smoking beha iou , can all exe an in luence on ha inal ‘indi idual’
decision.
Ano he pa icula o m o heal h ela ed beha iou is heal h ca e use. As wi h o he
beha iou s, heal h ca e use and quali y a e also socially pa e ned (136). This can be
ela ed o ac o s like economic esou ces, access o in o ma ion o geog aphical
accessibili y. Howe e , a leas in a Eu opean con ex , heal h ca e is a om being
he mos impo an cause o HI and p obably con ibu es only sligh ly o hese
inequali ies (137).
1.2.3. Which Measu es?
Despi e being known o cen u ies, HI ha e no always been a unanimously accep ed
ac . The Black Repo dedica ed some o i s pages in o showing ha HI we e no a
p oduc o ma hema ical a e ac (4), bu i was s ill c i icized o he measu es i used
(138). In ac , he measu e one chooses o analyse HI can de e mine he esul o he
analysis, possibly e en leading o con adic ing esul s (138).
One o he mos commonly used me hods a e ange measu es, p obably he mos
simple and easy o in e p e . They compa e he heal h s a us o wo g oups o he
socioeconomic dis ibu ion. This is done by calcula ing absolu e o ela i e
di e ences (i.e., a ios). Fo example, i hal (0.5) o he poo es quin ile and a i h
(0.2) o he iches quin ile ha e diabe es, he absolu e di e ence is 30 pe cen age
poin s (0.5-0.2=0.3) and he a io is 2.5 (0.5/0.2=2.5).
While inequali y is summa ized in one alue, ange measu es show an incomple e
pic u e, as hey only compa e wo g oups (usually op and bo om). Inequali y
be ween hese wo g oups may emain he same, while he dis ibu ion wi hin he
middle g oups d ama ically changes (138). Ano he sho coming o hese measu es is
ha hey igno e he sizes o he g oups. This is impo an especially in compa a i e
analyses – ac oss ime o space – as wo simila esul s may e lec wo e y di e en
dis ibu ions. This is also an issue when socioeconomic g oups do no ha e a ixed
size, such as hose de ined by occupa ion o educa ion. A ange measu e o inequali y
may emain s able o e ime, bu i he g oup wi h mo e yea s o educa ion inc eases,
while he numbe o less educa ed dec ease, he dis ibu ion is clea ly di e en .
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
27
To o e come some o hese sho comings, eg ession-based measu es can be used. An
example is he Slope Index o Inequali y (SII): he eg ession coe icien ( he slope) o
a eg ession model o he heal h ou come on SES, o de ed by SES. I can be
in e p e ed as he absolu e e ec on he heal h ou come o mo ing om one SES
ca ego y o alue o he nex (139). Despi e i s simplici y, his measu e is sensi i e o
he popula ion mean o he heal h ou come, which limi s is compa abili y ac oss
popula ions and ime. The Rela i e Index o Inequali y (RII) o e comes his issue, as
i is calcula ed by di iding he SII by he popula ion mean o he heal h ou come.
Measu es based on he concen a ion cu e (CC) a e also used. These measu es
o igina ed in he economic analysis o income dis ibu ion, amously summa ized in
he Gini Coe icien – a measu e o how income o weal h o a popula ion is
dis ibu ed among i s elemen s. Equally, when a heal h ou come measu e is used, i is
possible o summa ize in one numbe how heal h is dis ibu ed, e lec ing no only he
expe ience o wo g oups, bu o he whole popula ion. Addi ionally, unlike he SII
and he RII, measu es based on he CC do no assume a linea ela ionship be ween
independen and dependen a iables.
An impo an dis inc ion ha can be done in he me hods used o calcula e HI is he
di e ence be ween ela i e and absolu e measu es. I has been a gued ha he use o
only absolu e o ela i e measu es can be misleading, as i can in luence he eade s’
pe cep ion o he magni ude, signi icance and e en di ec ion o HI (140, 141). In ac ,
me hodological e iews o he epo ing o HI ha e shown ha absolu e and ela i e
measu es o he same e ec can yield opposi e esul s (141). Addi ionally, he choice
o ela i e o absolu e measu es can also e lec di e en equi y alue judgmen s
(142). When compa ing wo g oups ha ge heal hie a an equip opo iona e a e –
i.e., in bo h g oups’ heal h imp o es a an X% a e – hen a ela i e measu e o HI
will emain he same, while an absolu e measu e will change. On he o he hand, i
he g oups’ heal h imp o es in a uni o m way – i.e., bo h g oups see an o e all
imp o emen o X pe cen age poin s – hen an absolu e measu e o HI will no
change, while a ela i e measu e will change (142) 1.
1 This is be e illus a ed in he hypo he ical example ou lined in appendix 1.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
28
1.2.4. Which Mechanisms?
a) Pa hway Be ween Indi idual Socioeconomic S a us and Heal h
The s udy o he pa hways ha link SES o heal h is compa a i ely less common han
o i s de e minan s, ou comes o measu es. In 2001, Dide ichsen, E ans and
Whi ehead ou lined a amewo k o unde s anding he social o igins o HI (28) ha
summa izes hese pa hways ( igu e 3). A ha ime, li e a u e on he SDH was jus
escala ing and s a ing o ouch he opic o con ex as an impo an de e minan o
heal h, shying away om he adi ional concep ualiza ion o isk ac o s as indi idual
a ibu es. Dide ichsen’s amewo k ecognizes he ole o con ex in he c ea ion o
HI, al hough i mos ly ocuses on indi idual pa hways.
P obably he mos no iceable use o Dide ichsen’s amewo k was i s applica ion by
he WHO’s CSDH as a basis o hei own amewo k ( igu e 2) (12, 13).
Dide ichsen’s amewo k has also been used as a basis o o he amewo ks (143); as
a map o unde s and o he social phenomena, such as in ima e pa ne iolence (144)
and social consequences o disease (145); as a ame o p esen e idence o li e a u e
e iews on a ic inju ies in you h (146), in cys ic ib osis in he UK (147) and in
wo k- ela ed heal h (148); as a basis o policy compa isons be ween coun ies (145,
149, 150); and as a map o empi ical analysis o he heal h o lone mo he s (151,
152) and smoking in adolescen s (153).
Dide ichsen’s amewo k has no aised much academic discussion, and o he au ho s
who used i as a basis o hei own wo k did no explici ly c i ique i s applicabili y.
Despi e his, i p o ides a simple, ye comple e summa y o how SES in luences
heal h and ice- e sa. The amewo k desc ibes ou mechanisms: (I) social
s a i ica ion, (II) di e en ial exposu e, (III) di e en ial suscep ibili y and (IV)
di e en ial consequences; and ou en y poin s o policies o a ge hose
mechanisms: (A) in luencing social s a i ica ion, (B) dec easing exposu es, (C)
dec easing ulne abili y, and (D) p e en ing unequal consequences. Wi h his
amewo k, he au ho s aim o p o ide a model o unde s and he p ocess o c ea ion
o HI, a sys ema iza ion o wha kind o policies migh wo k in hei mi iga ion, and a
base o empi ically es which HI p oducing mechanisms a e mo e impo an in a
socie y.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
29
This amewo k p o ides a simple, ye exhaus i e, desc ip ion o how SES and heal h
ou comes a e in e ela ed. I can be easily applicable o any SES measu e and heal h
ou come, as he a ie y o analyses i has been used o clea ly shows. I also clea ly
iden i ies he s eps in which di e en ial e ec s occu , which in u n lead o HI.
Figu e 3. A amewo k o elucida ing he pa hways om he social con ex o heal h ou comes and o
in oducing policy in e en ions. Sou ce: Dide ichsen F, E ans T, Whi ehead M. The social basis o
dispa i ies in heal h: Challenging inequi ies in heal h: om e hics o ac ion. New Yo k: Ox o d
Uni e si y P ess; 2001.
i. Fi s Mechanism: Social S a i ica ion
Social s a i ica ion is he way con ex de e mines indi idual social posi ion ( igu e
3). ‘Con ex ’ is unde s ood by Dide ichsen (28) as ‘ he spec um o ac o s in socie y
ha canno be di ec ly measu ed a he indi idual le el’ (p. 19). As was la e
desc ibed by he au ho (154), he social s a i ica ion mechanism ac ually
encompasses wo mechanisms: ‘one ha gene a es and dis ibu es weal h and powe
o di e en social posi ions in socie y and one ha s a i ies indi iduals in o he social
posi ions’ (p. 59). Thus, he pa e n o heal h ac oss social posi ions is a esul o bo h
he cha ac e is ics o he posi ions and o he indi iduals occupying hem.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
30
Unde he au ho ’s unde s anding, he p ocess o social s a i ica ion occu s mos ly
du ing ea ly childhood de elopmen (155); a his s age, a ange o de e minan s
ope a es o de e mine social posi ion and heal h in adul li e. Fo example, in a sample
o he Po uguese popula ion, heigh (as a p oxy measu e o childhood social
ci cums ances) was s ongly associa ed wi h se e al heal h ou comes, such as as hma
and ch onic pain, such ha alle people (who ended o ha e had be e childhood
ci cums ances) we e heal hie (156). The exis ence o li e-long de e minan s
highligh s he impo ance o a li e-cou se app oach o he analysis o HI.
Policies ha in luence social s a i ica ion usually all ou o he ealm o ‘heal h
policies’. These a e policies ha dec ease social inequali ies, such as c ea ing equal
educa ional oppo uni ies o edis ibu ing weal h. Because social posi ion is
inex icably linked o heal h, educ ions in social inequali ies ha e he po en ial o
also educe HI. As po e y has also been ex ensi ely linked o heal h, hese policies
can be pa icula ly impo an in p o ec ing he mos ulne able – h ough, o
example, illness pensions.
ii. Second Mechanism: Di e en ial Exposu e
Di e en ial exposu e e e s o he unequal dis ibu ion in ype, amoun o du a ion o
exposu es ha impac heal h on di e en social g oups. Unequal exposu es comp ise
en i onmen al, biological o beha iou al isk ac o s, which a e commonly connec ed
o social posi ion. Fo example, an unskilled wo ke may ha e a low income ha does
no allow hem o choose a heal hy die ; a pe son li ing in an u ban en i onmen is
exposed o mo e ai pollu ion; and he s ess o ha ing a low income can make a
pe son mo e p one o smoke. Whils social s a i ica ion ope a es mainly in ea ly li e,
di e en ial exposu es can occu in childhood o adul li e, as a consequence o he
social s a i ica ion p ocess.
Many heal h policies ha employ isk educ ion s a egies – o example, media
campaigns o p omo e heal hy ea ing by in o ming people o wha cons i u es a
heal hy die – do no di e en ia e be ween social g oups. Howe e , heal h campaigns
each indi iduals di e en ly, as he mo e educa ed can be mo e exposed.
Acco dingly, analyses o hese s a egies ha e shown ha , despi e imp o ing some
people’s heal h, hey some imes inc ease HI, as hey disp opo iona ely imp o e he
heal h o hose ha a e heal hie o s a o wi h (157).

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
31
To educe exposu es, one mus ‘modi y he e ec o social posi ion on he occu ence
o he speci ic causes’ (154: p.60). This means s a egies should aim o educe he
disp opo iona e amoun o isk ac o s ha some people ace, ocusing on pa icula
g oups ha ace hem, o example, by p o ec ing hem om occupa ional isks, bad
housing, o inadequa e nu i ion. No iceably, di e en ial exposu es end o clus e , as
people in lowe social s a a a e mo e likely o be exposed o mul iple ad e se isk
ac o s. This clus e ing o ad e se exposu es h ough li e cons i u es he hi d
mechanism: di e en ial ulne abili y.
iii. Thi d Mechanism: Di e en ial Vulne abili y
The e ec o an exposu e on an indi idual is no exclusi ely a unc ion o he
exposu e, bu also o he indi idual him/he sel . Fo Dide ichsen e al. (155), he
added exposu es o mul iple isk ac o s inc ease he ulne abili y o people in lowe
social posi ions. These exposu es ac syne gis ically, making he indi idual mo e
suscep ible o he e ec o each o hem. This means ha e en i a gi en isk ac o is
dis ibu ed e enly ac oss social g oups, i s impac may be une enly dis ibu ed among
hose g oups due o di e en unde lying ulne abili y (28). Thus, he hi d mechanism
is ‘mos ly a ques ion o clus e ing o lowe socio-economic g oups o causes in he
same pa hway’ (154: p.60).
The success o policies ha aim o dec ease isky exposu es also depends on he
exis ence o o he exposu es and on he con ex in which he indi idual is in. Fo
example, a policy migh be pu in place o p o ide housing o e e yone, bu a
homeless indi idual wi h an incapaci a ing men al illness and no o he o m o
suppo will no be able o na iga e he adminis a i e p ocess o apply o he house
hey need. Reducing ulne abili y depends on jus ha : ackling in e ac ing exposu es
and no jus ocusing on a single one.
Howe e , ulne abili y is no only abou he addi i e o in e ac ing e ec o se e al
exposu es, i is also abou con ex ual e ec s. As he Dide ichsen poin s ou (158),
‘child en li ing in ex eme po e y ha e e y di e en mo ali y a es in di e en
coun ies, which shows ha he na ional policy con ex modi ies he e ec o po e y’
(p. 14). Fo example, a socie y wi h s ong social cohesion can po en ially educe he
e ec s o po e y o unemploymen in an indi idual’s heal h, by educing he s ess
associa ed wi h hese si ua ions.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
32
Thus, educing ulne abili y mus include comp ehensi e s a egies ha ackle
mul iple exposu es simul aneously, and p o iding people wi h an en i onmen ha
helps mi iga e he e ec s o indi idual isk ac o s.
i . Fou h Mechanism: Di e en ial Consequences
Heal h is a undamen al good, aluable no only o i s own sake, bu also o how i
allows indi iduals o ul il hei expec a ions abou hei li es. Illness can ha e social
and economic consequences, which can eed back in o he mechanism o social
s a i ica ion (mechanism I in igu e 3). Fo example, loss o a limb may lead o
ex ensi e heal hca e paymen s and also o los income due o inabili y o wo k; hese
di ec and indi ec cos s can ha e emendous consequences in a amily’s budge .
Howe e , his impac depends on o he de e minan s, such as how well o he amily
was be o e he heal h e en o he social suppo ha is a ailable – such as uni e sal
heal hca e o disabili y insu ance. Depending on hese, he consequences can be mo e
o less g a e, changing he likelihood o ‘ alling behind’.
The consequences o disease can also ha e an e ec on an agg ega e le el, as high
a es o illness can in luence a coun y’s social and economic de elopmen . An
example o his is he impac o HIV in Sou h A ica: in 2000, he Wo ld Bank
p ojec ed ha G oss Domes ic P oduc (GDP) would be 17% lowe in 2010 due o he
high a es o he disease (159).
Policies aimed a educing he social consequences o disease include hose ela ed o
he p o ision o heal hca e (including p ima y, seconda y and e ia y), and hose
aimed a mi iga ing he economic consequences o disease, such as p o iding wo k
oppo uni ies, bene i s o insu ance o people who a e ill. These policies, by p o iding
heal hca e and p omo ing ein eg a ion in he wo k o ce a e a pe son alls ill, ha e
he po en ial o b eak he cycle be ween SES and heal h.
b) Pa hways Be ween Con ex and Heal h and i s Dis ibu ion
Despi e ex ensi e e idence connec ing mul iple con ex ual cha ac e is ics wi h
indi idual and popula ion heal h, he pa hways be ween he wo ha e no been
pa icula ly explo ed (160). Macin y e e al. (14) called his he ‘black box o places’
(p. 131), an unknown in luence ha we can see bu canno explain.
The amewo k o he CSDH ies o ill his gap. Acco ding o his amewo k,
con ex c ea es social s a i ica ion and dis ibu es indi iduals h ough s a a, which
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
33
ha e an impac on hei heal h (13). In his iew, con ex is composed o go e nance,
policy (mac oeconomic, social and heal h), and cul u al and socie al no ms and alues
( igu e 2) (12). This amewo k ook some ideas om Dide ichsen’s amewo k,
whose desc ip ion o he pa hways be ween con ex and heal h was mo e de ailed. In
ac , Dide ichsen’s amewo k can be app oached om a pe spec i e o he indi idual
o o socie y (150). When iewed om he pe spec i e o socie y, Dide ichsen’s
amewo k iden i ies wo p ocesses h ough which socie y impac s heal h and HI:
social s a i ica ion and policy en y-poin s.
In Dide ichsen’s wo ds (28), ‘[ he p ocess o social s a i ica ion] bo h alloca es
powe and weal h o social posi ions and allows indi iduals in o di e en posi ions’
(p. 21). Dide ichsen a gues ha he p ocess o social s a i ica ion is ‘cen al’ o he
issue o HI, in so much as i helps us o dis inguish be ween ‘ ai ’ and ‘un ai ’
inequali ies (i.e., inequali ies e sus inequi ies). The au ho a gues ha , once one
unde s ands he p ocess o social s a i ica ion, one can judge i as o whe he i is
‘ ai ’ o ‘un ai ’. As such, i powe and weal h a e dis ibu ed ‘ ai ly’ among social
posi ions o a socie y and i indi iduals a e ‘ ai ly’ alloca ed o hose posi ions, HI
s emming om ha p ocess can also be deemed ‘ ai ’ (i.e., no inequi ies2). I
emains, none heless, ha classi ying some hing as ‘ ai ’ o ‘un ai ’ is a no ma i e
exe cise, and hus depends on iews o jus ice.
Con ex can also in luence heal h and heal h dis ibu ion o he ex en ha i c ea es
in o mal and o mal ules ha in luence heal h, such as policies. Fo example,
uni e sal access o heal h ca e can help equalize oppo uni ies, as i helps indi iduals
who ha e allen ill o eco e , e-en e he wo k o ce and a oid a all in social
posi ion. In o mal ules can also ha e a heal h impac , such as when expec a ions
abou a neighbou hood allow i o main ain high le els o iolence. Fu he mo e,
hese ules can impac di e en g oups di e en ly. This in luence o con ex is
acknowledged in Dide ichsen’s amewo k o a la ge ex en h ough he policy ‘en y
poin s’ he iden i ies (dec easing exposu es, dec easing ulne abili y and p e en ing
unequal consequences). By iden i ying hese en y-poin s, he au ho desc ibes how
2 The dis inc ion be ween heal h inequi ies and inequali ies is no uni e sally accep ed, bu i is
gene ally conside ed o be a ques ion o whe he a mo al judgemen is made o no : while inequali y
and equali y a e pu ely desc ip i e – simply desc ibe a di e ence be ween wo g oups – inequi y and
equi y encompass a mo al iew – no only he e is a di e ence, bu i is an unjus di e ence.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
34
policies, by shaping he ules wi hin a socie y, in luence i s heal h and heal h
dis ibu ion.
In summa y, acco ding o Dide ichsen’s amewo k, con ex in luences heal h and
heal h dis ibu ion h ough wo majo mechanisms: social s a i ica ion and policy
en y poin s. Social s a i ica ion ope a es h ough he c ea ion o hie a chies in a
socie y, whe eas policy en y poin s in luence he pa hways be ween social posi ion
and heal h. Howe e , as Whi ehead, Bu s öm and Dide ichsen (150) poin ou , policy
is only one componen o social con ex ha ‘may ha e an in luence on he pa hways
be ween social posi ion and heal h’ (p. 257). A comp ehensi e socie al pe spec i e
needs o ocus on how all con ex ual componen s – no jus policy – in luence heal h
and i s dis ibu ion.
I emains ha , despi e he gene al ag eemen ha con ex in luences heal h and HI,
lack o unde s anding o he mechanisms by which his happens unde mines he
e idence ha is p oduced (14, 160). This is u he ein o ced by he obse a ion ha
a ious public heal h policies seem o ha e ailed o educe HI, e en when his was
hei explici goal (25, 161). This shows a need o sys ema ize he mechanisms by
which con ex in luences heal h and HI, o be e unde s and and s udy hem and o
design policies ha a e app op ia e o hei ou lined goals.
1.2.5. Summa y
The choice o heal h ou come, SES indica o , o measu e o he analysis o HI, can
e lec di e en esea ch ques ions and di e en no ma i e iews, and can ul ima ely
lead o di e en conclusions. Despi e his, HI ha e been shown o a a ie y o heal h
ou comes – om mo ali y, o mo bidi y, and heal h beha iou s –, o SES indica o s –
bo h indi idual and con ex ual –, and using a a ie y o di e en measu es.
The mechanisms ha link indi idual SES o heal h ou comes can be desc ibed in ligh
o Dide ichsen’s amewo k, which desc ibes a pa hway ha s a s in social
s a i ica ion – he way ha con ex de e mines indi idual social posi ion. Social
posi ion hen goes on o de e mine exposu e o isk ac o s, ulne abili y o hose
ac o s, and he consequences o heal h back on social posi ion. Dide ichsen
summa ises how con ex in luences HI in wo mechanisms: social s a i ica ion and
policies. Howe e , a comp ehensi e pe spec i e on con ex mus include o he
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
41
Po uguese cons i u ion, in e es in heal h equi y has been limi ed. A WHO epo on
he Po uguese Na ional Heal h Plan a gued ha HI we e an ‘impo an policy gap’ in
his plan (202). In an in e es ing analysis o his seeming lack o in e es , Bago d’U a
a gues ha i is he absence o explici and e ec i e policies o ackle HI ha allows
hem o emain so p e alen (203).
1.3.3. Summa y
Despi e conside able imp o emen s in o e all heal h and in social suppo in
Eu opean coun ies, HI emain high, in some cases e en inc easing in he las yea s.
This has been called a ‘pa adox’, as HI emain an impo an public heal h challenge in
he con inen .
Po ugal is a pa icula ly in e es ing case s udy o HI, as i is one o he mos unequal
Eu opean coun ies, bo h in e ms o income and heal h dis ibu ion. Cu en e idence
sugges s he exis ence o signi ican p o-poo inequali y in mos heal h ou comes and
o mos socioeconomic de e minan s, wi h a ew no iceable excep ions, such as
alle gies and smoking. On op o his, Po ugal has gone h ough impo an changes in
he las decade, namely wi h he implemen a ion o aus e i y measu es ha seem o
ha e had an impac on he p o ision o social se ices, hus po en ially inc easing HI.
Despi e his, poli ical a en ion o he issue in Po ugal is s ill e y limi ed, which may
be exac ly why HI emain so high.

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
42
1.4. Objec i es
The o e all aim o his disse a ion is o con ibu e o he unde s anding o how
cha ac e is ics o he con ex can ha e an impac on popula ion heal h and heal h
dis ibu ion, using Po ugal and Eu ope as case s udies.
To accomplish his aim, he ollowing objec i es we e pu sued:
• To analyse and in e p e how changes in indi idual and con ex ual social
capi al, and he in e ac ion be ween he wo, we e associa ed wi h changes in
SAH in Eu opean coun ies be ween 2002 and 2012.
• To analyse and in e p e how he ela ionship be ween social mobili y and HI
a ied be ween six wel a e egime ypes in Eu opean coun ies be ween 2002
and 2012.
• To collec , summa ise, desc ibe, and in e p e a ailable e idence abou
socioeconomic HI in Po ugal.
• To ou line how social inequali ies in heal h limi a ions changed in Po ugal
be ween 2004 and 2014, conside ing di e en measu es o social s a us, and
in e p e he esul s in he ligh o mac oeconomic changes in he coun y in
ha pe iod.
• D awing on he indings om he p e ious analyses, o ou line a heo e ical
amewo k ha summa ises how con ex in luences popula ion heal h and
heal h dis ibu ion.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
43
1.5. Me hodological App oaches
This esea ch used seconda y da a collec ed be ween 2002 and 2014 in Eu opean
coun ies. Acco ding o he ules and egula ions o he E hical Council o he
Ins i u e o Hygiene and T opical Medicine o he No a Uni e si y o Lisbon, his
esea ch did no equi e e hical app o al om he Council (204).
1.5.1. Da a Sou ces
This wo k was based on da a om wo main da abases: he Eu opean Social Su ey
(ESS) and he Eu opean Union Su ey on Income and Li ing Condi ions (EU-SILC).
a) Eu opean Social Su ey
The ESS is a epea ed c oss-sec ional su ey ha collec s da a on a i udes, belie s
and beha iou pa e ns (205). The su ey is applied e e y wo yea s in mo e han
hi y Eu opean coun ies since 2002. Coun ies a e ee o pa icipa e o no e e y
yea he su ey is pe o med. The su ey is join ly unded by he EU and each
pa icipa ing coun y (206). Da a is made eely a ailable online o esea che s, upon
a simple egis a ion p ocess on he websi e (207).
ESS aims o achie e a ep esen a i e sample on each coun y o pe sons o e 15 yea s
old li ing in p i a e households. Each coun y is gi en he eedom o choose hei
p e e ed sampling design, conside ing cos s, expe ience and o he coun y speci ic
ac o s. Howe e , a ew equi emen s a e applied o e e y coun y o ensu e
compa abili y o he samples, such as he use o s ic andom p obabili y me hods a
e e y s age, a high esponse a e (minimum 70%), a ull co e age o he popula ion,
and a minimum e ec i e sample size (n=1,500 o n=800 in coun ies whe e he
popula ion is smalle han 2 million) (208).
The ESS ques ionnai e consis s o a co e sec ion and a o a ing sec ion. The co e
module ul ils ESS’s p ima y ole, o moni o ing change in alues and a i udes in
Eu ope h ough ime. The o a ing modules a e selec ed based on a call o p oposals
made in he O icial Jou nal o he EU. In 2014, he o a ing modules we e on ‘Social
inequali ies in heal h and hei de e minan s’ and on ‘a i udes owa ds immig a ion
and hei an eceden s’ (209). The ESS co e ques ionnai e includes wo heal h ela ed
ques ions: SAH (‘How is you heal h in gene al?’) and hampe ed in daily ac i i ies
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
44
(‘A e you hampe ed in you daily ac i i ies in any way by any longs anding illness, o
disabili y, in i mi y o men al heal h p oblem? I yes, is ha a lo o o some ex en ?’).
ESS esul s ha e been widely used in academic esea ch and in policy documen s.
The ESS bibliog aphy includes hund eds o publica ions ha show he ange o use
his su ey has had (210), including no only analyses o su ey esul s, bu also b oad
discussions on su ey me hodology.
b) Eu opean Union Su ey on Income and Li ing Condi ions
EU-SILC is an annual su ey ca ied ou in se e al Eu opean coun ies wi h a mixed
longi udinal and c oss-sec ional design. The su ey is egula ed by EU law as a way
o compiling compa able da a on income, po e y and social exclusion wi hin he
union, using ha monized me hods and de ini ions (211). As such, e e y EU coun y is
expec ed o pa icipa e, by se ing up i s own da a collec ion o using exis ing su eys
ha comply wi h EU-SILC equi emen s. The EU unds he i s ou yea s o da a
collec ion o each membe s a e. Da a is made a ailable o scien i ic pu poses gi en
he esea che s’ compliance wi h a numbe o p i acy equi emen s. To access his
da a, au ho isa ion was sough om and p o ided by Eu os a ’s Mic oda a Access
Team.
The su ey’s longi udinal componen ollows a simple o a ional design: in yea one,
a c oss-sec ional sample is selec ed; his sample is di ided in ou sub-samples, each
i sel ep esen a i e o he popula ion. In yea wo, one sub-sample is d opped, he
o he h ee a e ollowed up, and one new sub-sample is added. As such, excep o he
h ee i s sub-samples, e e y sub-sample is eques ed o be pa o he su ey o ou
consecu i e yea s. In any one pa icula yea , he o e all sample made up o ou sub-
samples, which make up he c oss-sec ional sample (212).
EU-SILC aims o in e iew a ep esen a i e sample o people aged 16 o o e li ing
in a p i a e household. Sample size depends on coun y size; o example, minimum
e ec i e sample size o he c oss-sec ional sample in Po ugal is 10,500 and 7,500
o he longi udinal componen . This adds up o a minimum sample size o 166,000
pe yea in he c oss-sec ional sample, when all coun ies a e combined (211). Jus
like he ESS, sampling p ocedu es can be de ined wi hin each coun y, as long as hey
ollow ce ain equi emen s ha make he samples compa able, such as he use o
p obabili y sampling and o households as he basic uni o sampling (213).
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
45
Mos o he EU-SILC ques ionnai e conce ns li ing condi ions, po e y, and income.
Th ee heal h ela ed ques ions a e included:
• SAH: ‘How is you heal h in gene al?’ allowing ollowing answe s:
o Ve y good / Good / Fai / Bad / Ve y bad / Don’ know / Re usal o
answe
• Ch onic condi ions: ‘Do you ha e any long-s anding illness o heal h
p oblem?’ allowing ollowing answe s:
o Yes / No / Don’ know / Re usal o answe
• Limi a ions by heal h condi ions: ‘Fo a leas he pas six mon hs, o wha
ex en ha e you been limi ed because o a heal h p oblem in ac i i ies people
usually do?’ allowing ollowing answe s:
o Se e ely limi ed / Limi ed bu no se e ely / No limi ed a all / Don’
know / Re usal o answe
Eu os a uses his las a iable – limi a ions by heal h condi ions – as a measu e o
disabili ies (212).
EU-SILC esul s a e a undamen al sou ce o da a o he wo k o he Eu opean
Commission. They ha e been ex ensi ely used in he p oduc ion o books, scien i ic
publica ions, poli ical s a emen s, and s a is ical wo king pape s, among o he s (214).
1.5.2. Me hods
a) Sys ema ic Re iew o he Li e a u e
A sys ema ic e iew (SR) can be de ined as a ‘ e iew o he scien i ic e idence which
applies s a egies ha limi bias in he assembly, c i ical app aisal, and syn hesis o all
ele an s udies on he speci ic opic’ (33: p.276). SR a e impo an gi en he
ou s anding numbe o scien i ic publica ions and he amoun o exis ing e idence,
some imes con adic o y, on a pa icula opic (215). Thus, he alue o SR has long
been es ablished, especially o clinical p ac ice, as hey p o ide quick and easy
access o a summa y o he a ailable e idence o busy clinicians (216).
Public heal h in e en ions a e usually mo e complex han he adi ional andomized
clinical ial. They can ac on di e en le els simul aneously – di ec ly wi h he
indi idual o on a b oade , con ex ual le el – and h ough mul iple mechanisms.
Some imes, andomized in e en ions a e no easible o p ac ical o e hical easons.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
46
This means ha he adi ional me hodological guidance o SR can be insu icien o
ace his complexi y (215), and public heal h esea che s ha e o be c ea i e when
doing a SR. This includes eso ing o al e na i e s udy designs, such as quali a i e o
obse a ional s udies (217). This p esen s i s own issues, as obse a ional s udies can
ind spu ious associa ions, and, in his si ua ion, he explo a ion o he e ogenei y
migh yield be e insigh s han a emp s o look o an o e all measu e o e ec
(218).
To help o e come limi a ions associa ed wi h SR, he PRISMA guidelines sys ema ise
o e a ching p inciples ha should be applied (PRISMA s ands o P e e ed
Repo ing I ems o Sys ema ic Re iews and Me a-Analyses). These guidelines (216)
consis o a 27-i em checklis and a low diag am ou lined by expe s and, as much as
possible, based on a ailable e idence, ha aim o ‘imp o e he epo ing o SR and
me a-analysis’ (p. 2). The p oposed diag am ( igu e 4) sugges s ha all s eps ha lead
o he inclusion o a ce ain numbe o s udies should be eco ded and explained in
de ail, including sou ces and easons o exclusion. The 27-i em checklis includes
guidance on all elemen s o he SR, om he i le, abs ac , in oduc ion, me hods,
esul s, discussion and unding.
Figu e 4. Flow o in o ma ion h ough he phases o a sys ema ic e iew p oposed by he PRISMA
s a emen . Sou ce: Mohe D e al. P e e ed epo ing i ems o sys ema ic e iews and me a-analyses:
he PRISMA s a emen . BMJ 2009;339(7716):332.

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
47
b) S a is ical Me hods
i. Logis ic Reg ession
Reg ession analyses a e an impo an pa o esea ch in HI. Bina y logis ic eg ession
analyses a e a pa icula ype o eg ession analyses ha conce n hemsel es wi h
bina y ou comes, i.e., when he heal h a iable only akes wo dis inc alues.
Al hough he gene al p inciples o he logis ic eg ession a e simila o hose o he
linea eg ession, he e a e di e ences in he o m o he model and i s assump ions
(219). The goal o any eg ession analysis is o ind he bes i ing, mos
pa simonious, and in e p e able model o a ela ionship be ween a se o explana o y
a iables and an ou come a iable (219). Once i ed, he selec ed model yields an
es ima e o each coe icien o each o he independen a iables included. This
coe icien can be in e p e ed as a a e o change – o ‘a unc ion o he dependen
a iable pe uni o change in he independen a iable’ (220: p.49), holding all o he
independen a iables cons an . In p ac ice, he coe icien is he di e ence be ween
he log o he odds o wo di e en alues o he independen a iable.
In he analysis o HI, he independen a iables a e usually he SES a iables o
in e es , such as income o educa ion. Changes in he scales o he log-odds a e no
easy o in e p e ; howe e , logis ic eg ession coe icien s can be easily con e ed o
OR. These, on he o he hand, a e easily in e p e ed as he a io o he odds o he
ou come a iable be ween wo g oups de ined by he independen a iable (220). Fo
example, i he heal h ou come is mo ali y (y=1 o dea h, y=0 o su i al), and
educa ion is he independen a iable (x=0 o less han high educa ion, x=1 o high
educa ion), an OR=2 means ha he odds o dying among he less educa ed a e wo
imes highe han he odds o dying among he mo e educa ed. This is also applicable
o con inuous independen a iables, conside ing a 1-uni o an x-uni inc ease (220).
Using he same example, bu i ins ead o educa ion we use income, an OR=2 may
mean ha o e e y inc ease in uni o income (1 €, 1.000 £, 100 $, depending on he
scale used), he odd o dea h dec ease by hal .
ii. Mul ile el Models
Ecological a iables ha e been dismissed by epidemiology o a long ime,
pa icula ly since he iden i ica ion o he ecologic allacy, which s a es ha
indi idual ela ionships canno be in e ed om ecologic ela ionships (15).
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
48
Con e sely, in he o he ex eme, comple ely igno ing he e ec s o con ex ual
cha ac e is ics can lead o he indi idualis ic (o a omis ic) allacy – ‘an e oneous
in e ence abou causal ela ionships in g oups o people made on he basis o
ela ionships obse ed in indi iduals’ (33: p. 11). To answe his need, s a is ical
me hods mus include bo h indi idual and g oup le el cha ac e is ics, i.e., hey mus
be mul ile el. Mul ile el me hods a e eg ession-based models ha ake in o accoun
he hie a chical s uc u e o he da a, o example, a he indi idual and
neighbou hood le el.
The concep o heal h de e minan s as bo h indi idual and con ex ual na u ally leads
o a mul ile el pe spec i e. Se e al o he easons also wa an he use o mul ile el
analyses when s udying SDH. Fi s , mul ile el models allow o he simul aneous
es ima ion o eg ession coe icien s o a iables a mul iple le els. This is essen ial,
as he esul o a single le el analysis migh in ac be an a e ac , e lec ing a
ela ionship ha uly only exis s a a di e en le el (18). Second, hey can
disen angle complex ques ions o con ex ual and indi idual he e ogenei y. While
con ex s can ha e an o e all e ec on popula ion heal h, hey migh do so by a ec ing
only a pa icula g oup (con ex ual he e ogenei y); on he o he hand, he a iabili y o
a heal h ou come wi hin a con ex may be e y di e en among di e en g oups
(indi idual he e ogenei y). Thi d, mul ile el models allow o he analysis o
in e ac ions be ween a iables a di e en le els (indi idual-con ex ual in e ac ions).
These a e impo an as he same con ex may some imes ha e opposi e heal h e ec s
on di e en g oups (221). Fou h, hey can sequen ially include mul iple le els o
hie a chical clus e ing, om indi idual, o households, communi ies, and egions, o
example (222). They also allow o mo e complex da a s uc u es, such as c oss-
classi ied and mul iple membe ship, which allow indi iduals o be assigned o
mul iple g oups simul aneously (223). Finally, mul ile el models can also ake in o
accoun he ime dimension, such as when obse a ions a e nes ed wi hin ime
a iables (such as yea ), which a e hen nes ed wi hin egion, o example (18).
iii. Concen a ion Index
The CIx is a measu e o HI based on he CC. The CC is he esul o plo ing he
cumula i e pe cen age o indi iduals, anked by income, wi h he cumula i e
pe cen age o he heal h a iable. Figu e 5 shows a hypo he ical example o a CC. In
plo ing a CC, pe ec equali y is ep esen ed by a diagonal line, showing a pe ec ly
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
49
equal dis ibu ion o heal h among he sample popula ion, ega dless o income. The
CIx is calcula ed as wice he a ea be ween he CC and he line o pe ec equali y.
When he e is no inequali y, heal h is dis ibu ed equally wi hin he popula ion, he
CC coincides wi h he diagonal and he CIx is ze o. In he o he ex eme, i heal h is
concen a ed in one pe son, he CC is shaped like an in e ed ‘L’, and he CIx is 1 (o
-1).
Figu e 5. Hypo he ical concen a ion cu e. The x-axis ep esen s he cumula i e p opo ion o he
popula ion, anked om poo es o iches . The y-axis ep esen s he cumula i e p opo ion o he
heal h a iable. In his example, heal h is disp opo iona ely concen a ed among he iches , as, o
example, he poo es hal o he popula ion only ha e abou 25% o he heal h a iable. Sou ce:
au ho ’s own elabo a ion.
In he pa icula case o bina y heal h ou comes, he CIx is no limi ed by he (-1,1)
ange, bu depends on he mean o he ou come a iable in he popula ion. As his
limi s compa abili y be ween di e en popula ions (ac oss ime o be ween a eas, o
example), Wags a (2005) p oposed a ‘no malisa ion’ o he CIx, by which i is
di ided by 1 minus he mean, making i compa able (224).
Wags a , an Doo slae , and Wa anabe (2003) ha e shown ha he CIx can be
decomposed in o he con ibu ions o indi idual ac o s o he income- ela ed HI
(225). These ac o s a e usually demog aphic a iables, such as age and sex, o o he
SES, such as educa ion o occupa ion. The con ibu ion o each ac o depends on wo
cha ac e is ics: he elas ici y o ha ac o wi h espec o he heal h a iable, and he
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
50
deg ee o income- ela ed inequali y o ha ac o (i.e., he CIx o ha ac o ). The
con ibu ion o each ac o o he o e all CIx is he p oduc o he heal h elas ici y and
he CIx o ha ac o .
1.5.3. Summa y
The exis ence o a social g adien in heal h shows how heal h is no de e mined only
by indi idual SES, bu depends on he con ex he pe son is in. This concep ual
app oach canno be explo ed using only adi ional s a is ical me hods, bu needs
mo e ad anced s a is ical models o conside he complexi y and di e en dimensions.
Mul ile el analysis allows o he desc ip ion o bo h indi idual and con ex ual
cha ac e is ics and o he quan i ica ion o hei impac on heal h. While no
excluding o he s a is ical me hods, such as simple eg ession analyses and me hods
based on he CC, mul ile el models can be e e lec he hie a chical na u e o heal h
de e minan s.
The CIx, on he o he hand, allows o a di e en analysis o HI. While no including
con ex ual de e minan s, i e lec s he o e all expe ience o he popula ion.
This esea ch used wo eg ession-based me hods: single le el and mul ile el logis ic
eg essions. I was also used measu es based on he CC, namely he CIx and i s
decomposi ion. Fu he mo e, a SR o he li e a u e was pe o med, ollowing he
PRISMA guidelines.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
57
113. Johansson E. A no e on he impac o hou s wo ked on mo ali y in OECD
coun ies. The Eu opean Jou nal o Heal h Economics. 2004;5(4):335-40.
114. Ruhm CJ. A e ecessions good o you heal h? The Qua e ly jou nal o
economics .115.2 (2000): 617-650.
115. Ca alano R, Goldman-Mello S, Sax on K, Ma ge ison-Zilko C, Subba aman M,
LeWinn K, e al. The heal h e ec s o economic decline. Annual e iew o public
heal h. 2011;32.
116. Ka asek RA, Theo ell T. Heal hy wo k: s ess, p oduc i i y, and he
econs uc ion o wo king li e. Basic books; 1992.
117. B uckne TA. Economic an eceden s o p one in an sleep placemen among
black mo he s. Annals o epidemiology. 2008;18(9):678-81.
118. McC one A, Ushe E, Sonn ag-O’B ien V, Moslene U, G üning C. Global
ends in enewable ene gy in es men 2012. F ank u School UNEP Collabo a ing
Cen e o Clima e and Sus ainable Ene gy Finance. 2012.
119. Buonoco e J. Mi iga ion and heal h: Clima e policy no so cos ly. Na u e Clima e
Change. 2014;4(10):861-2.
120. Bacigalupe A, Escola -Pujola A. The impac o economic c ises on social
inequali ies in heal h: wha do we know so a ? In e na ional jou nal o equi y in
heal h. 2014;13(1):52.
121. Chen L, Be lingue G, E ans T, Whi ehead M, Dide ichsen F, Bhuiya A, e al.
Heal h equi y in a globalizing wo ld. In: E ans T, edi o . Challenging inequi ies in
heal h: om e hics o ac ion. New Yo k: Ox o d Uni e si y P ess; 2001. p. 34-44.
122. Wilkinson RG. Socioeconomic de e minan s o heal h. Heal h inequali ies:
ela i e o absolu e ma e ial s anda ds? BMJ. 1997;314(7080):591.
123. Zajaco a A, Dowd JB. Reliabili y o sel - a ed heal h in US adul s. Ame ican
jou nal o epidemiology. 2011;174(8):977-83.
124. Idle EL, Benyamini Y. Sel - a ed heal h and mo ali y: a e iew o wen y-se en
communi y s udies. Jou nal o heal h and social beha io . 1997:21-37.
125. Schni ke J, Bacak V. The inc easing p edic i e alidi y o sel - a ed heal h.
PloS one. 2014;9(1):e84933.
126. Singh-Manoux A, Guéguen A, Ma ikainen P, Fe ie J, Ma mo M, Shipley M.
Sel - a ed heal h and mo ali y: sho -and long- e m associa ions in he Whi ehall II
s udy. Psychosoma ic medicine. 2007;69(2):138-43.
127. F anks P, Gold MR, Fiscella K. Sociodemog aphics, sel - a ed heal h, and
mo ali y in he US. Social science & medicine. 2003;56(12):2505-14.
128. Bu s öm B, F edlund P. Sel a ed heal h: Is i as good a p edic o o
subsequen mo ali y among adul s in lowe as well as in highe social classes?
Jou nal o epidemiology and communi y heal h. 2001;55(11):836-40.
129. Bjo ne JB, Faye s P, Idle E. Sel - a ed heal h. Assessing quali y o li e.
2005:309-23.
130. Eu os a . Disabili y s a is ics in oduced 2015. A ailable om:
h p://ec.eu opa.eu/eu os a /s a is ics-
explained/index.php/Disabili y_s a is ics_in oduced. [Accessed June 1s 2017]
131. A endano M, A o AR, Mackenbach JP. Socio-economic dispa i ies in physical
heal h in 10 Eu opean coun ies. Mannheim Resea ch Ins i u e o he Economics o
Aging; 2005.
132. Rennie K, Jebb S. P e alence o obesi y in G ea B i ain. Obesi y e iews.
2005;6(1):11-2.

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
58
133. B unne E, Ma mo M, Nanchahal K, Shipley M, S ans eld S, Juneja M, e al.
Social inequali y in co ona y isk: cen al obesi y and he me abolic synd ome.
E idence om he Whi ehall II s udy. Diabe ologia. 1997;40(11):1341-9.
134. Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adai -Rohani H, e al. A
compa a i e isk assessmen o bu den o disease and inju y a ibu able o 67 isk
ac o s and isk ac o clus e s in 21 egions, 1990–2010: a sys ema ic analysis o he
Global Bu den o Disease S udy 2010. The Lance . 2013;380(9859):2224-60.
135. Okechukwu C, Da ison K, Emmons K. Changing heal h beha io s in a social
con ex . In: Be kman LF, Kawachi I, Glymou MM, edi o s. Social epidemiology.
Second edi ion. New Yo k: Ox o d Uni e si y P ess; 2014. p. 365-395.
136. Van Doo slae E, Masse ia C, Koolman X, G oup OHER. Inequali ies in access
o medical ca e by income in de eloped coun ies. Canadian medical associa ion
jou nal. 2006;174(2):177-83.
137. S i bu I. Inequali ies in heal h, does heal h ca e ma e ? Social inequali ies in
mo ali y in Eu ope, wi h a special ocus on he ole o he heal h ca e sys em.
Ro e dam: E asmus MC; 2008.
138. Wags a A, Paci P, Van Doo slae E. On he measu emen o inequali ies in
heal h. Social science & medicine. 1991;33(5):545-57.
139. Regido E. Measu es o heal h inequali ies: pa 1. Jou nal o epidemiology and
communi y heal h. 2004;58(10):858.
140. King NB, Ha pe S, Young ME. Use o ela i e and absolu e e ec measu es in
epo ing heal h inequali ies: s uc u ed e iew. BMJ. 2012;345:e5774.
141. Kelly MP, san é CdlOsldsdl, Bonne oy J. The social de e minan s o heal h:
de eloping an e idence base o poli ical ac ion. Na ional Ins i i u e o Clinical
Excellence; 2007.
142. Allanson P, Pe ie D. On The Choice O Heal h Inequali y Measu e Fo The
Longi udinal Analysis O Income‐Rela ed Heal h Inequali ies. Heal h economics.
2013;22(3):353-65.
143. Due P, K ølne R, Rasmussen M, Ande sen A, Damsgaa d MT, G aham H, e al.
Pa hways and mechanisms in adolescence con ibu e o adul heal h inequali ies.
Scandina ian Jou nal o Public Heal h. 2011;39(6 suppl):62-78.
144. La sen MM. Heal h Inequi ies Rela ed o In ima e Pa ne Violence Agains
Women. The Role o Social Policy in he Uni ed S a es, Ge many, and No way.
Swi ze land: 2016.
145. Whi ehead M, Han a y B, Bu s om B. 3.1 Unequal consequences o ill heal h:
esea ching he ole o public policy. In: G aham H, edi o . Unde s anding Heal h
Inequali ies. 2009. p.181-198.
146. La lamme L, Dide ichsen F. Social di e ences in a ic inju y isks in childhood
and you h—a li e a u e e iew and a esea ch agenda. Inju y P e en ion.
2000;6(4):293-8.
147. Taylo -Robinson D, Diggle P, Smy h R, Whi ehead M. Heal h T ajec o ies in
People wi h Cys ic Fib osis in he UK: Explo ing he E ec o Social Dep i a ion. In:
Bu on-Jeang os C, Culla i S, Sacke A, Blane D, edi o s. A Li e Cou se Pe spec i e
on Heal h T ajec o ies and T ansi ions. London: Sp inge ; 2015. p. 85-110.
148. Toi anen S. Wo k- ela ed inequali ies in heal h. Cen e o Heal h Equi y
S udies. 2007.
149. Phipps S. Heal h ou comes o child en in Can ada, England, No way and he
Uni ed S a es. Social indica o s esea ch. 2007;80(1):179-221.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
59
150. Whi ehead M, Bu s öm B, Dide ichsen F. Social policies and he pa hways o
inequali ies in heal h: a compa a i e analysis o lone mo he s in B i ain and Sweden.
Social science & medicine. 2000;50(2):255-70.
151. F i zell S, Wei o GR, F i zell J, Bu s öm B. F om mac o o mic o: he heal h
o Swedish lone mo he s du ing changing economic and social ci cums ances. Social
science & medicine. 2007;65(12):2474-88.
152. Bu s om B, Whi ehead M, Clay on S, F i zell S, Vannoni F, Cos a G. Heal h
inequali ies be ween lone and couple mo he s and policy unde di e en wel a e
egimes– he example o I aly, Sweden and B i ain. Social science & medicine.
2010;70(6):912-20.
153. Schnoh CW, K eine S, Rasmussen M, Due P, Dide ichsen F. School- ela ed
media o s in social inequali ies in smoking: a compa a i e c oss-sec ional s udy o
20,399 adolescen s. In e na ional jou nal o equi y in heal h. 2009;8(1):17.
154. Dide ichsen F. Policies o heal h equi y: an e hical and epidemiological
amewo k and a ge s o a new na ional heal h policy in Sweden. Heal h Ca e
P io i y Se ing. 2003:57-62.
155. Dide ichsen F, Ande sen I, Manuel C, Ande sen A-MN, Bach E, Baadsgaa d M,
e al. Heal h Inequali y-de e minan s and policies. Scandina ian jou nal o public
heal h. 2012;40(8 suppl):12-105.
156. Pe elman J. A e ch onic diseases ela ed o heigh ? Resul s om he Po uguese
Na ional Heal h In e iew Su ey. Economics & Human Biology. 2014;15:56-66.
157. F eimu h VS. The ch onically unin o med: Closing he knowledge gap in heal h.
In: Ray EB, Donohew L, edi o s. Communica ion and heal h: Sys ems and
applica ions. Hillsdale, New Je sey: Law ence E lbaum Associa es; 1990:171-86.
158. Dide ichsen F. Resou ce alloca ion o heal h equi y: issues and me hods. Wo ld
Bank. 2004.
159. A nd C, Lewis JD. The mac o implica ions o HIV/AIDS in Sou h A ica: a
p elimina y assessmen . Sou h A ican Jou nal o Economics. 2000;68(5):380-92.
160. Galea S. In eg a i e chap e : mac osocial de e minan s o popula ion heal h. In:
Galea S, edi o . Mac osocial De e minan s o Popula ion Heal h. New Yo k:
Sp inge ; 2007. p. 295-7.
161. Mackenbach JP. The pe sis ence o heal h inequali ies in mode n wel a e s a es:
he explana ion o a pa adox. Social science & medicine. 2012;75(4):761-9.
162. Wo ld Heal h O ganisa ion. Eu opean Heal h Fo All Da abase. A ailable om:
h p://da a.eu o.who.in /h adb/. [Accessed on June 1s 2017].
163. Gla ze M. Wel a e s a e g ow h and he cu en c isis in Po ugal: Social
spending and i s challenges. Ins i u e o Eu opean S udies. 2012.
164. Guillén A, Ál a ez S, Sil a PAE. Redesigning he Spanish and Po uguese
wel a e s a es: he impac o accession in o he Eu opean Union. Sou h Eu opean
Socie y and Poli ics. 2003;8(1-2):231-68.
165. Van Ke m P, Alpe in MNP. Inequali y, g ow h and mobili y: The in e empo al
dis ibu ion o income in Eu opean coun ies 2003–2007. Economic Modelling.
2013;35:931-9.
166. Hespanha P, Po ugal S. Wel a e Cu s and Insecu i y unde he Rule o
Aus e i y: he Impac o he C isis on Po uguese Social Se ices. Oña i Socio-Legal
Se ies. 2015;5(4).
167. OECD. Social and Wel a e Issues: Inequali y 2014. A ailable om:
h p://www.oecd.o g/social/inequali y.h m. [Accesses June 1s 2017]
168. Ca ilho MJ. O declínio ecen e da mo alidade in an il em Po ugal. Re is a do
Cen o de Es udos Demog á icos. 1985;27:155-9.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
60
169. Lucas JS. Iniquidade social pe an e a doença e a mo e em Po ugal — 1985.
Re is a do Cen o de Es udos Demog á icos. 1986;n.º 28:29.
170. Pe ei a JACG. Equi y, heal h and heal hca e: an economic s udy wi h e e ence
o Po ugal. PhD disse a ion. Uni e si y o Yo k; 1995.
171. Go sens M, Ma í-Dell'Olmo M, Pé ez K, Palència L, Ma inez-Benei o M-A,
Rod íguez-Sanz M, e al. Socioeconomic inequali ies in inju y mo ali y in small
a eas o 15 Eu opean ci ies. Heal h & place. 2013;24:165-72.
172. Simões A, Paque e A, A aújo M. Equidade Ho izon al no sis ema de saúde
po uguês: sec o público s sec o p i ado. Associação Po uguesa de Economia da
Saúde. 2006.
173. Simões N, C espo N, Mo ei a SB, Va um CA. Measu emen and de e minan s o
heal h po e y and ichness: e idence om Po ugal. Empi ical Economics.
2016;50(4):1331–1358.
174. Ribei o S. Desigualdades Socioeconómicas na Doença Ca dio ascula em
Po ugal: Es udo baseado no Inqué i o Nacional de Saúde 2005/06. Mas e s Thesis.
No a Uni e si y o Lisbon; 2010.
175. Ribei o O, Teixei a L, A aújo L, A onso RM, Pachana N. P edic o s o anxie y
in cen ena ians: heal h, economic ac o s, and loneliness. In e na ional
Psychoge ia ics. 2014:1-10.
176. F aga S, Ma ques-Vidal P, Vollenweide P, Waebe G, Guessous I, Paccaud F, e
al. Associa ion o socioeconomic s a us wi h in lamma o y ma ke s: A wo coho
compa ison. P e en i e medicine. 2014.
177. Lawlo DA, Ha o M, Wedde kopp N, Ande sen LB, Sa dinha LB, Riddoch CJ,
e al. Associa ion o socioeconomic posi ion wi h insulin esis ance among child en
om Denma k, Es onia, and Po ugal: c oss sec ional s udy. BMJ.
2005;331(7510):183.
178. Sil a M. Heal h se ice usage in he Ribei a de Pena concelho o Po ugal. PhD
Disse a ion. Uni e si y o Exe e ; 1988.
179. Al es J, Kuns AE, Pe elman J. E olu ion o socioeconomic inequali ies in
smoking: esul s om he Po uguese na ional heal h in e iew su eys. BMC public
heal h. 2015;15(1):1.
180. Lopez AD, Collishaw NE, Piha T. A desc ip i e model o he ciga e e epidemic
in de eloped coun ies. Tobacco con ol. 1994;3(3):242.
181. Mo ei a PA, Pad ão PD. Educa ional and economic de e minan s o ood in ake
in Po uguese adul s: a c oss-sec ional su ey. BMC Public Heal h. 2004;4(1):58.
182. San os A-C, Ba os H. P e alence and de e minan s o obesi y in an u ban
sample o Po uguese adul s. Public Heal h. 2003;117(6):430-7.
183. San os AC, Eb ahim S, Ba os H. Gende , socio-economic s a us and me abolic
synd ome in middle-aged and old adul s. BMC Public Heal h. 2008;8(1):1.
184. Noguei a H, San ana P, San os R. Saúde: Vulne abilidade e Opo unidade na
Á ea Me opoli ana de Lisboa. In: San ana P, edi o . A Cidade e a Saúde. Coimb a:
Edições Almedina. 2007.
185. Nunes B, Sil a RD, C uz VT, Ro iz JM, Pais J, Sil a MC. P e alence and
pa e n o cogni i e impai men in u al and u ban popula ions om No he n
Po ugal. BMC neu ology. 2010;10(1):1.
186. Al es L, Aze edo A, Sil a S, Ba os H. Socioeconomic inequali ies in he
p e alence o nine es ablished ca dio ascula isk ac o s in a sou he n Eu opean
popula ion. PLoS One. 2012;7(5):e37158.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
61
187. Pe ei a M, Lune N, Paulo C, Se e o M, Aze edo A, Ba os H. Incidence o
hype ension in a p ospec i e coho s udy o adul s om Po o, Po ugal. BMC
ca dio ascula diso de s. 2012;12(1):1.
188. Al es E, Hen iques A, Co eia S, San os AC, Aze edo A, Ba os H.
Ca dio ascula isk p o ile o mo he s o a Po uguese bi h coho : A su ey 4yea s
a e deli e y. P e en i e medicine. 2013;57(5):494-9.
189. Sandes AR, Nascimen o C, Figuei a J, Gou eia R, Valen e S, Ma ins S, e al.
Alei amen o Ma e no: p e alência e ac o es condicionan es. Ac a Médica
Po uguesa. 2007;20:193-200.
190. Camões M, Lopes C, Oli ei a A, San os AC, Ba os H. O e all and cen al
obesi y incidence in an u ban Po uguese popula ion. P e en i e medicine.
2010;50(1):50-5.
191. San os MR, Vale MS, Mi anda L, Mo a J. Socio-demog aphic and pe cei ed
en i onmen al co ela es o walking in Po uguese adul s—A mul ile el analysis.
Heal h & place. 2009;15(4):1094-9.
192. Mo ei a P, San os S, Pad ão P, Co dei o T, Bessa M, Valen e H, e al. Food
pa e ns acco ding o sociodemog aphics, physical ac i i y, sleeping and obesi y in
Po uguese child en. In e na ional jou nal o en i onmen al esea ch and public
heal h. 2010;7(3):1121-38.
193. Ca elaa s AE, Kuns AE, Geu s JJ, C ialesi R, G ö ed L, Helme U, e al.
Educa ional di e ences in smoking: in e na ional compa ison. BMJ.
2000;320(7242):1102-7.
194. Veiga PA. Income- ela ed heal h inequali y in Po ugal. Uni e si y o Minho.
2005.
195. Ribei o S, Fu ado C, Pe ei a J. Associa ion be ween ca dio ascula disease and
socioeconomic le el in Po ugal. Re is a Po uguesa de Ca diologia (English
Edi ion). 2013;32(11):847-54.
196. Vilela S, Oli ei a A, Pin o E, Mo ei a P, Ba os H, Lopes C. The in luence o
socioeconomic ac o s and amily con ex on ene gy-dense ood consump ion among
2-yea -old child en. Eu opean jou nal o clinical nu i ion. 2015;69(1):47-54.
197. Kuns AE, Bos V, Ande sen O, Ca dano M, Cos a G, Ha ding S, e al.
Moni o ing o ends in socioeconomic inequali ies in mo ali y: expe iences om a
Eu opean p ojec . Demog aphic esea ch. 2004;2:229-54.
198. Oli ei a A, Maia B, Lopes C. De e minan s o inadequa e ui and ege able
consump ion amongs Po uguese adul s. Jou nal o Human Nu i ion and Die e ics.
2014;27(s2):194-203.
199. Sco KM, Al-Hamzawi AO, And ade LH, Bo ges G, Caldas-de-Almeida JM,
Fies as F, e al. Associa ions be ween subjec i e social s a us and DSM-IV men al
diso de s: esul s om he wo ld men al heal h su eys. JAMA psychia y.
2014;71(12):1400-8.
200. Eu opean Commission. Resea ch Findings – Social Si ua ion Moni o - Income
inequali y in EU coun ies. A ailable om:
h p://ec.eu opa.eu/social/main.jsp?ca Id=1050&in PageId=1870&langId=en.
Accessed on June 1s 2017.
201. Pe ei a J, Fu ado C. Equidade em saúde: do conhecimen o à acção. In:
DGS/ENSP, edi o . A No a Saúde Pública: A saúde Pública na E a do
Conhecimen o. Lisbon: G adi a; 2011.
202. Wo ld Heal h O ganisa ion. WHO e alua ion o he na ional heal h plan o
Po ugal (2004-2010). Copenhagen: WHO Regional O ice o Eu ope. 2010.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
62
203. Bago d'U a T. Equidade no sec o da saúde em Po ugal. In: Simões J, edi o . 30
Anos do Se iço Nacional de Saúde - um pe cu so comen ado. Lisbon: Simões, Jo ge;
2010.
204. Conselho de É ica do Ins i u o de Higiene e Medicina T opical da Uni e sidade
No a de Lisboa. Pedido de ap eciação é ica pa a p oje os e p o ocolos de
in es igação. Lisbon: CEIHMT-UNL; 2017.
205. ESS eam. Abou he Eu opean Social Su ey Eu opean Resea ch In as uc u e
– ESS ERIC. A ailable om: h p://www.eu opeansocialsu ey.o g/abou /. [Accessed
June 1s 2017]
206. ESS eam. Abou ESS - Funding. A ailable om:
h p://www.eu opeansocialsu ey.o g/abou / unding.h ml. [Accessed June 1s 2017]
207. Eu opean Social Su ey. F equen ly Asked Ques ions. A ailable om:
h p://www.eu opeansocialsu ey.o g/abou / aq.h ml. [Accessed June 1s 2017]
208. The ESS Sampling Expe Panel. Sampling o he Eu opean Social Su ey
Round VII: P inciples and Requi emen s. London: Cen e o Compa a i e Social
Su eys, Ci y Uni e si y London; 2014.
209. ESS eam. Ro a ing Modules. A ailable om:
h p://www.eu opeansocialsu ey.o g/me hodology/ques ionnai e/ o a ing_ques ionna
i e.h ml. [Accessed June 1s 2017]
210. ESS eam. ESS Bibliog aphy. A ailable om:
h p://www.eu opeansocialsu ey.o g/bibliog aphy/. [Accessed June 1s 2017]
211. Eu opean Commission. Regula ion (EC) No 1177/2003 o he Eu opean
Pa liamen and o he Council o 16 o June 2003 conce ning Communi y s a is ics on
income and li ing condi ions (EU-SILC). B ussels; 2003.
212. Eu os a . Heal h a iables in SILC – me hodology. A ailable om:
h p://ec.eu opa.eu/eu os a /s a is ics-
explained/index.php/Heal h_ a iables_in_SILC_-_me hodology-
Main_concep s_and_de ini ions. [Accessed June 1s 2017]
213. Eu opean Commission. Commission Regula ion (EC) No 1982/2003 o 21
Oc obe 2003 implemen ing Regula ion (EC) No 1177/2003 o he Eu opean
Pa liamen and o he Council conce ning Communi y s a is ics on income and li ing
condi ions (EU-SILC) as ega ds he sampling and acing ules, 1982/2003.
B ussels; 2003.
214. Eu os a . S a is ics on Income and Li ing Condi ions – Publica ions. A ailable
om: h p://ec.eu opa.eu/eu os a /web/income-and-li ing-condi ions/publica ions.
[Accessed June 1s 2017]
215. Khan KS, Te Rie G, Glan ille J, Sowden AJ, Kleijnen J. Sys ema ic e iews:
CRD's guidance o unde aking e iews in heal h ca e. 3 d edi ion. London:NHS
Cen e o Re iews and Dissemina ion; 2009.
216. Mohe D, Libe a i A, Te zla J, Al man DG. P e e ed epo ing i ems o
sys ema ic e iews and me a-analyses: he PRISMA s a emen . PLoS med.
2009;6(7):1-6.
217. Pe ic ew M. Sys ema ic e iews in public heal h: old ches nu s and new
challenges. Bulle in o he Wo ld Heal h O ganiza ion. 2009;87(3):163-A.
218. Egge M, Smi h GD, Schneide M. Sys ema ic e iews o obse a ional s udies.
In: Egge M, Smi h GD, Al man DG, edi o s. Sys ema ic Re iews in Heal h Ca e:
Me a-Analysis in Con ex , Second Edi ion. 2001:211-27.
219. Hosme DW, Lemeshow S. In oduc ion o he logis ic eg ession model. In:
Hosme DW, Lemeshow S, S u di an RX ,edi o s. Applied Logis ic Reg ession,
Thi d Edi ion. 2013:1-33.

The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
63
220. Hosme DW, Lemeshow S, S u di an RX. In e p e a ion o he Fi ed
Logis ic Reg ession Model. In: Hosme DW, Lemeshow S, S u di an RX, edi o s.
Applied Logis ic Reg ession, Thi d Edi ion. 2013:49-88.
221. Sub amanian S, Kim DJ, Kawachi I. Social us and sel - a ed heal h in US
communi ies: a mul ile el analysis. Jou nal o U ban Heal h. 2002;79(1):S21-S34.
222. Sub amanian S, Delgado I, Jadue L, Vega J, Kawachi I. Income inequali y and
heal h: mul ile el analysis o Chilean communi ies. Jou nal o epidemiology and
communi y heal h. 2003;57(11):844-8.
223. Fielding A, Golds ein H. C oss-classi ied and mul iple membe ship s uc u es in
mul ile el models: An in oduc ion and e iew. Uni e si y o B is ol. Resea ch
Repo RR791, 2006.
224. Wags a A. The bounds o he concen a ion index when he a iable o in e es
is bina y, wi h an applica ion o immuniza ion inequali y. Heal h economics.
2005;14(4):429-32.
225. Wags a A, Van Doo slae E, Wa anabe N. On decomposing he causes o
heal h sec o inequali ies wi h an applica ion o malnu i ion inequali ies in Vie nam.
Jou nal o econome ics. 2003;112(1):207-23.
The ela ionship be ween con ex and heal h inequali ies Chap e 1. In oduc ion
64
The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
65
2. Resul s
2.1. The ‘da k side’ o social capi al: us and sel - a ed heal h in
Eu opean coun ies
Re e ence:
Campos-Ma os I, Sub amanian SV, Kawachi I. The ‘da k side’ o social capi al: us
and sel - a ed heal h in Eu opean coun ies. Eu opean Jou nal o Public Heal h.
2016;26(1):90-95. DOI: h p://dx.doi.o g/10.1093/eu pub/ck 089
Online supplemen a y da a o his publica ion is in Appendix 2.
The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
66
The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
73
2.2. Social mobili y and heal h in Eu opean coun ies: does wel a e
egime ype ma e ?
Re e ence:
Campos-Ma os I, Kawachi I. Social mobili y and heal h in Eu opean coun ies: does
wel a e egime ype ma e ? Social Science and Medicine. 2015;142:241-248 DOI:
h ps://doi.o g/10.1016/j.socscimed.2015.08.035

The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
74
Sho communica ion
Social mobili y and heal h in Eu opean coun ies: Does wel a e egime
ype ma e ?
In^
es Campos-Ma os
a
,
b
,
*
, Ichi o Kawachi
b
a
Depa men o In e na ional Public Heal h and Bios a is ics, No a Uni e si y o Lisbon, Lisbon, Po ugal
b
Depa men o Social and Beha io al Sciences, Ha a d School o Public Heal h, Bos on, MA, USA
a icle in o
A icle his o y:
Recei ed 7 Oc obe 2014
Recei ed in e ised o m
10 Augus 2015
Accep ed 19 Augus 2015
A ailable online 22 Augus 2015
Keywo ds:
Eu ope
Wel a e egimes
Social mobili y
Sel - a ed heal h
abs ac
Heal h inequali ies pose an impo an public heal h challenge in Eu opean coun ies, o which inc eased
social mobili y has been sugges ed as a cause. We sough o desc ibe how he ela ionship be ween
heal h inequali ies and social mobili y a ies among wel a e egime ypes in he Eu opean egion. Da a
om six ounds o he Eu opean Social Su ey was analyzed using mul ile el s a is ical echniques,
s a ified by wel a e egime ype, including 237,535 indi iduals om 136 coun ies. Social mobili y
among indi iduals was defined acco ding o he disc epancy be ween pa en al and o sp ing educa ional
a ainmen . Fo each wel a e egime ype, he associa ion be ween social mobili y and sel - a ed heal h
was examined using odds a ios and isk di e ences, con olling o pa en al educa ion. Upwa dly mobile
indi iduals had be ween 23 and 44% lowe odds o epo ing bad o e y bad sel - a ed heal h when
compa ed o hose who emained s able. On an absolu e scale, o me USSR coun ies showed he bigges
and only significan di e ences o upwa d mo emen , while Scandina ian coun ies showed he
smalles . Downwa d social mobili y ended o be associa ed wi h wo se heal h, bu he esul s we e less
consis en . Upwa d social mobili y is associa ed wi h wo se heal h in all Eu opean wel a e egime ypes.
Howe e , in Scandina ian coun ies he associa ion o upwa d mobili y was smalle , sugges ing ha he
No dic model is mo e e ec i e in mi iga ing he impac o social mobili y on heal h and/o o heal h on
mobili y.
©2015 Else ie L d. All igh s ese ed.
1. In oduc ion
Despi e sus ained e o s pu in e ec ac oss Eu opean coun-
ies, heal h inequali ies pe sis as an impo an public heal h
challenge (Mackenbach, 2012). A ange o policy solu ions has been
ied, bu so a wi h ela i ely li le impac . Social mobili y has
been iden ified as an impo an d i e o heal h inequali ies. Social
mobili y can occu ei he be ween gene a ions (pa en s and chil-
d en) as well as wi hin he li e-cou se o he indi idual. T unca ed
in e gene a ional social mobili y is o pa icula conce n because i
can esul in he c ys alliza ion o weal h inequali y as well as
heal h inequali ies. Ill heal h is a po en cause o bo h in a-
indi idual and in e -gene a ional mobili y es ic ion. Fo
example, childhood illness has been shown o ad e sely a ec
educa ional a ainmen (Case and Paxson, 2008), which will
subsequen ly a ec an indi idual's success in he labo ma ke . Ill
heal h in midli e can a ec labo o ce pa icipa ion ( educed
wo king hou s, job loss), esul ing in downwa d income mobili y.
Res ic ed social mobili y can be mani es in mul iple dimensions e
educa ional achie emen , occupa ional s a us o ea nings and in-
come. Fu he mo e heal h selec ion can be bo h di ec (e.g.
dep essi e illness di ec ly esul ing in unca ed educa ional
achie emen ) as well as indi ec ee.g. dep essi e illness esul ing
in educed social mobili y ia in e media y ac o s such as s igma
and disc imina ion (Wes , 1991). Acco dingly, social p o ec ions
such as uni e sal access o heal h ca e o an i-disc imina ion
legisla ion ep esen impo an policies o p omo e bo h in a-
indi idual and in e -gene a ional social mobili y.
An indi idual's socioeconomic posi ion is a obus de e minan
o his/he heal h, bo h in e ms o hei cu en (o achie ed) so-
cioeconomic posi ion, bu also hei li e ime ajec o y (Ma mo
and Macmillan, 2004; Ma mo and Wilkinson, 2005). This can
eflec p ocesses o accumula ion o a di ec impac o social
mobili y (Hallq is e al., 2004). S udies on he e ec o social
mobili y on heal h ha e no always p oduced clea -cu esul s, wi h
*Co esponding au ho . Depa men o In e na ional Public Heal h and Bios a-
is ics, No a Uni e si y o Lisbon, Lisbon, Po ugal.
E-mail add ess: [email p o ec ed] (I. Campos-Ma os).
Con en s lis s a ailable a ScienceDi ec
Social Science & Medicine
jou nal homepage: www.else ie .com/loca e/socscimed
h p://dx.doi.o g/10.1016/j.socscimed.2015.08.035
0277-9536/©2015 Else ie L d. All igh s ese ed.
Social Science & Medicine 142 (2015) 241e248
some seeming o indica e ha upwa d social mobili y can be jus as
dele e ious o heal h as downwa d social mobili y (Hemmingsson
e al., 1999; Libe a os e al., 1988). These mixed esul s migh ,
howe e , be a consequence o he inconsis en ways in which social
mobili y has been ope a ionalized in he empi ical li e a u e
(Singhamme and Mi elma k, 2010). The use o di e en indica o s
o cha ac e ize social g oups can also be o consequence, since
di e en indica o s, such as educa ion, occupa ion o income, as
well as he in e gene a ional mo emen be ween hem, can ha e
di e en meanings (Galoba des e al., 2006).
O e all, he e a e bo h heo e ical and empi ical g ounds o
sugges ha he causal ela ionship be ween heal h and social
mobili y is bidi ec ional: indi iduals ha e mo e o less oppo u-
ni ies o social mobili y depending on hei heal h endowmen
and hei heal h achie emen is a ec ed by ansi ions be ween
social s a a.
The ex en o social mobili y a ies subs an ially be ween
coun ies (Belle and Hou , 2006). Go e nmen ac ions, such as
expanding access o schooling o in es ing in he heal h o child en
(e.g. ia imp o ed nu i ion o accina ion p og ams) ha e he
po en ial capaci y o b eak he in e -gene a ional ansmission o
social disad an age. Conside ing he s ong ela ionship be ween
social mobili y and heal h, hese go e nmen al ac ions, sys ema-
ized in Fig. 1, can ha e an impo an impac on heal h inequali ies.
Wel a e egime ypes, o en used o ca ego ize Eu opean coun ies,
sha e common policies such as he ones ou lined in Fig. 1.
In his c oss na ional compa a i e s udy, we sough o examine
he ela ion be ween social mobili y and popula ion heal h among
di e en ypes o wel a e egimes in he Eu opean egion, in o de
o unde s and how he wel a e s a e migh mode a e he link be-
ween mobili y and heal h.
2. Me hods
2.1. Da a sou ces and a iables
Indi idual da a was collec ed om six ounds o he Eu opean
Social Su ey (ESS), be ween 2002 and 2012, om hi y selec ed
coun ies. The ESS is a epea ed c oss-sec ional su ey ha collec s
compa able da a on indi idual socioeconomic cha ac e is ics and
heal h s a us o se e al Eu opean coun ies (ESS ERIC, 2014). Da a is
a ailable online a www.eu opeansocialsu ey.o g.
The ou come a iable, sel - a ed heal h, was based on he su -
ey pa icipan 's esponse o he ques ion ‘How is you heal h in
gene al?’, dicho omized so ha 1 included ‘bad’o ‘ e y bad’(o he
possible answe s we e ‘ ai ’,‘good’o ‘ e y good’).
Social mobili y was measu ed in ela ion o mo he and a he 's
achie ed le el o educa ion acco ding o he In e na ional S anda d
Classifica ion o Educa ion (ISCED) le els. Al hough social mobili y
is usually measu ed on he basis o he a he s' social s anding, he
inc easing pa icipa ion o women in he wo k o ce and he
impo ance o he mo he s' cha ac e is ics on child en's heal h
beha io s (Fa a o and San onas aso, 1995) suppo he impo ance
o conside ing mo he s' s a us in social mobili y s udies; he e o e,
his analysis was done sepa a ely. Social mobili y was classified in
h ee possible ca ego ies: ‘down’,‘s able’and ‘up’, acco ding o
whe he he esponden had eached, espec i ely, a lowe , he
same, o highe educa ional le el han his o he pa en . Ou
measu e o mobili y con olled o he pa en 's educa ional
achie emen when he esponden was 14 ( he same a iable used
o assess mobili y). Failing o ake in o accoun he ‘social g oup o
o igin’has been a common pi all in p e ious s udies o in e -
gene a ional social mobili y and heal h (Singhamme and
Mi elma k, 2010). Con olling o pa en 's educa ional achie e-
men yields mobili y coe ficien s ha can be in e p e ed as inde-
penden om social g oup o o igin.
O he indi idual-le el a iables included age ( es ic ed o 25
yea s and up), gende , ma i al s a us, belonging o an e hnic mi-
no i y g oup, sel -pe cei ed income, domicile and main occupa-
ional ac i i y. Responden s who we e in ull- ime educa ion we e
excluded, since no ha ing comple ed educa ion did no pe mi
compa ison o pa en s' achie emen . Fo all hese a iables a base
ca ego y wi h con as ing indica o a iables was specified, excep
age, which was cen e ed a ound i 's g and mean.
Fig. 1. En y poin s o educing and elimina ing heal h dispa i ies.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248242
To educe he possibili y o con ounding by economic de elop-
men , G oss Domes ic P oduc (GDP) pe capi a, con e ed o in-
e na ional dolla s using pu chasing powe pa i y, was e ie ed
om he Wo ld Bank da abase (Wo ld Bank, 2014) and included as
a coun y-le el a iable (specified pe coun y, pe yea ).
Addi ionally, he Gini coe ficien , which migh also con ound he
associa ion be ween social mobili y and heal h, was e ie ed om
he Eu os a da abase (Eu os a , 2014) and used as a le el 2 a iable
(coun y-yea specific). Howe e , his was only used as a sensi i i y
analysis, since he Gini coe ficien was missing o many coun ies
o se e al yea s.
2.2. Wel a e egimes
Coun ies we e g ouped by wel a e egime ype and analyzed
sepa a ely. Wel a e egime classifica ion is a much-deba ed opic,
no only wi h dispu ed ypologies, bu also ega ding which cha -
ac e is ics should be used o o hei classifica ion (Bamb a, 2007).
None heless, we s a ed wi h a widely used ypology ha di ides
Eu opean coun ies in o ou egime ypes:
(i) Scandina ian, cha ac e ized by uni e sal and gene ous
benefi s and a s ong edis ibu i e social secu i y sys em
(Eikemo e al., 2008a; Fenge , 2007);
(ii) Anglo-Saxon, wi h a low le el o go e nmen spending on
social p o ec ion, modes benefi s, usually means- es ed
(Eikemo e al., 2008a; Fenge , 2007);
(iii) Bisma ckian, wi h benefi s ied o employmen , financed
mainly by employe and employee, and minimal edis ibu-
ion (Eikemo e al., 2008a);
(i ) Sou he n, wi h a dualis sys em o wel a e p o ision, which
s ongly p o ec s pa o he popula ion while unde -
p o ec ing ano he (19).
This classifica ion is p ima ily based on Esping-Ande sen's e al.
(1990) g oundb eaking wo k, which ope a ionalized h ee p inci-
ples: decommodifica ion, social s a ifica ion and he public-
p i a e mix, o classi y he fi s h ee ypologies (Eikemo e al.,
2008a; Espig Ande sen e al., 1990). Fe e a (1996) la e added
he Sou he n ype, basing his classifica ion on he co e age o social
p o ec ion schemes (Fe e a, 1996). This ypology has been epli-
ca ed in o he a emp s o define wel a e egime ypes (Bamb a,
2007; Bonoli, 1997) and has been used p e iously in he heal h
li e a u e (Eikemo e al., 2008a, 2008b).
The conside a ion o Cen al and Eas e n Eu opean coun ies o
he Eu opean Union adds u he complexi y o his classifica ion.
His o ically, he ajec o ies o hese coun ies' wel a e ans-
o ma ion can be sepa a ed in wo, depending on he ex en o
which he wel a e e o collapsed in he 1990's (Cook and P ess,
2010). This ypology sepa a es Cen al and Eas e n Eu opean
coun ies (including Poland, Hunga y, Czech Republic and he Bal ic
s a es, among o he s) om he emaining o me USSR s a es (such
as Russia and Cen al Asian coun ies). Howe e , using a hie a -
chical clus e analysis, Fenge (2007) showed ha , based on simi-
la i ies on go e nmen spending, social si ua ion and poli ical
pa icipa ion, hese coun ies could be di ided in he ollowing
way:
(i) Fo me USSR, wi h gene ally low go e nmen al spending on
social p og ams, mos ly financed h ough social
con ibu ions;
(ii) Pos -Communis Eu opean, e y simila o he fi s ype, bu
wi h highe le els o economic g ow h, infla ion, social well-
being and egali a ianism (Fenge , 2007).
I is impo an o no e ha hese cha ac e is ics do no neces-
sa ily desc ibe he coun ies in absolu e; in ac , mos coun ies
ha e a mix o di e en wel a e egimes, bu none heless ha e
p edominan cha ac e is ics o one ype.
2.3. S a is ical analysis
Da a was analyzed using mul ile el logis ic models based on a
logi -link unc ion wi h a fi s o de quasi-likelihood es ima ion
p ocedu e. The models we e un using MLwiN p og am e sion
2.28 (Rasbash e al., 2013). The da a was analyzed conside ing i s
hie a chical s uc u e in h ee le els: indi iduals (le el 1), nes ed
wi hin yea s (le el 2), nes ed wi hin coun ies (le el 3). O e all
odds, odds a ios, o e all p obabili ies and a e di e ences we e
calcula ed o each wel a e egime. This was based in pa in he
me hods used by Hemmingsson e al. (1999), al hough we applied a
mul ile el modeling echnique.
The use o mul ile el s a is ical echniques allows o he anal-
ysis o he e ec o bo h indi idual and con ex ual a iables on he
ou come o in e es . In his analysis, he con ex ual a iable was
coun y- and yea -specific, making i manda o y o include bo h as
le els. Addi ionally, hese models ake in o accoun he hie a chical
nes ing o indi idual obse a ions wi hin a yea and wi hin a
coun y, co ec ing o he wise unde es ima ed s anda d e o s and
allow he modeling o a iabili y a each le el o analysis
(Sub amanian e al., 2003).
As a sensi i i y analysis, he same models we e un using
di e en es ima ion p ocedu es.
3. Resul s
Table 1 ou lines he dis ibu ion o he a iables among wel a e
egime ypes. Each ype includes be ween wo and se en coun ies,
anging in sample size om 23,310 o 62,509 indi iduals. Coun ies
wi h a Scandina ian wel a e egime had he lowes p opo ion o
people epo ing bad o e y bad heal h (5.4%), ollowed by Anglo-
Saxon (5.8%), Bisma ckian (6.9%), Sou he n (11.9%), Pos -
Communis Eu opean (14.5%) and Fo me USSR (19.6%).
Mos esponden s had achie ed a highe educa ional le el han
hei pa en s: be ween 49.3 and 63.2% had imp o ed in ela ion o
hei mo he , and be ween 45.6 and 61.7% in ela ion o hei a he .
Be ween one hi d and hal o esponden s emained in he same
educa ional le el as hei pa en s and a smalle p opo ion ‘mo ed
down’eonly 1.6% in sou he n coun ies in ela ion o mo he 's
achie emen , up o 12.1% in Bisma ckian coun ies, in ela ion o
a he 's.
Table 2 shows he coe ficien s o he mobili y a iable (s able,
upwa d, downwa d) om he mul ile el models. The ull models
a e a ailable in Tables 3 and 4 as an online supplemen . The gene al
pa e n o associa ion be ween social mobili y and heal h was
simila ac oss all egime ypes, i.e. upwa d mobili y was p o ec i e,
while downwa d mobili y was de imen al o sel - a ed heal h.
Figs. 2 and 3 show he p obabili ies and isk di e ences in he
di e en mobili y g oups in each wel a e egime ype. As was no ed
in Table 1, he o e all p obabili y o bad sel - a ed heal h di e s
significan ly be ween wel a e egime ypes. The benefi s o upwa d
mobili y a e p esen in all egime ypes, bu isk di e ences a e
pa icula ly high o Sou he n, Pos Communis Eu opean and
Fo me USSR egimes and significan ly di e en only in he la e .
Scandina ian coun ies showed he lowes isk di e ences be-
ween upwa dly mobile and s able indi iduals when conside ing
ei he mo he s' o a he s' educa ion.
Downwa d mobili y is gene ally de imen al. Pos Communis
Eu opean coun ies show he bigges absolu e di e ences, wi h
downwa d mobili y leading o an absolu e di e ence o 2.9 and
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248 243
3.6% in he isk o bad heal h measu ed on he basis o ma e nal and
pa e nal achie emen , espec i ely. In Scandina ian, Anglo-Saxon,
Bisma ckian and o me USSR coun ies, his isk di e ence
anges om 0.1 o 1.9%. O e all, Scandina ian coun ies showed he
smalles absolu e di e ence in heal h be ween s able and down-
wa d mobili y when measu ed by he a he 's achie emen , and
o me USSR coun ies when measu ed by he mo he 's.
As a sensi i i y analysis, he same models we e un using
di e en es ima ion p ocedu es (second o de ma ginal quasi-
likelihood, fi s and second o de p edic i e quasi-likelihood and
Ma ko Chain Mon e Ca lo). All esul ed in models wi h he same
e ec o social mobili y in he di e en wel a e egimes. A model
was also un wi h he Gini coe ficien as a le el 2 a iable; his had
e y li le e ec in he odds a ios o isk di e ences and no e ec in
he s a is ical significance o he esul s.
4. Discussion
This s udy sough o desc ibe di e ences in he ela ionship
be ween social mobili y and heal h wi hin di e en wel a e e-
gimes o be e unde s and he e ec ha di e en egime ypes
migh ha e. Di e en wel a e egimes had a subs an ially di e en
p opo ion o indi iduals wi h bad o e y bad sel - a ed heal h,
lowes in coun ies in he Scandina ian egime (5.4%), ollowed by
Anglo-Saxon (5.8%), Bisma ckian (6.9%), Sou he n (11.9%), Pos -
Communis Eu opean (14.5%) and finally Fo me USSR (19.6%).
Table 1
Composi ion o each wel a e ype.
Scandina ian Anglo-Saxon Bisma ckian Sou he n Pos -Communis Eu opean Fo me USSR
Coun ies Denma k
Finland
No way
Sweden
Iceland
Uni ed Kingdom
I eland
Aus ia
Belgium
Swi ze land
Ge many
F ance
Luxembou g
Ne he lands
Spain
G eece
I aly
Po ugal
Czech Republic
Hunga y
Poland
Slo enia
Slo akia
C oa ia
Bulga ia
Es onia
La ia
Li huania Russia
Uk aine
n (le el 2) 23 12 33 18 34 16
n (le el 1) 37,975 23,310 62,509 31,789 51,698 30,072
Le el 1 a iables
% bad o e y bad heal h 5.4 5.8 6.9 11.9 14.5 19.6
Mobili y, mo he (%)
Down 6.3 9.3 4.7 1.6 4.0 9.9
S able 30.4 35.2 32.9 49.0 38.0 29.2
Up 63.2 55.5 62.4 49.3 58.0 60.9
Mobili y, a he (%)
Down 10.1 10.2 12.1 3.4 6.8 9.4
S able 34.6 36.7 42.3 50.9 46.1 28.9
Up 55.4 53.1 45.6 45.6 47.0 61.7
Women (%) 50.0 55.0 52.9 56.6 55.1 61.7
Mean age (yea s) 51.7 51.7 51.6 52.4 51.8 52.7
U ban (%) 64.3 66.0 56.9 65.0 61.3 70.2
Main ac i i y (%)
Paid wo k 63.5 49.7 53.9 48.0 48.9 50.6
No ac i i y 31.3 35.3 32.9 35.5 42.0 42.0
O he 5.2 15.0 13.3 16.5 9.1 7.3
Feeling abou income (%)
Li ing com o ably 49.5 35.4 38.1 15.3 11.3 4.8
Coping 41.4 45.3 46.3 45.0 45.3 37.7
Di ficul 7.0 14.3 12.3 27.7 29.0 37.0
Ve y di ficul 2.0 5.0 3.3 12.1 14.5 20.5
Mino i y (%) 2.4 5.0 4.6 3.0 5.7 12.7
Ma i al s a us (%)
Ma ied 57.5 55.0 59.8 63.7 61.9 52.9
Sepa a ed/di o ced 12.2 11.2 11.9 6.5 9.5 14.7
Widow 7.0 10.9 9.3 12.2 14.2 19.8
Single 23.3 22.9 19.1 17.6 14.3 12.7
Le el 2 a iables
GDP pe capi a (in e na ional dolla s) 39,459 36,705 37,075 26,396 19,924 16,599
No e: ISCED In e na ional S anda d Classifica ion o Educa ion.
Table 2
Mobili y odds a ios and 95% confidence in e als om he mul ile el models in each wel a e egime ype.
Scandina ian Anglo-Saxon Bisma ckian Sou he n Pos -Communis Eu opean Fo me USSR
Mo he n (le el 1) 30,458 19,752 53,644 29,030 46,556 24,959
S able 11111 1
Down 1.39 (1.06,1.83) 1.40 (1.11,1.78) 1.09 (0.89,1.31) 0.72 (0.43,1.21) 1.35 (1.13,1.60) 1.01 (0.84,1.22)
Up 0.77 (0.68,0.86) 0.70 (0.60,0.81) 0.77 (0.71,0.84) 0.56 (0.49,0.62) 0.69 (0.65,0.75) 0.67 (0.61,0.74)
Fa he n (le el 1) 29,837 19,184 52,326 28,417 45,500 23,036
S able 11111 1
Down 1.18 (0.95,1.45) 1.51 (1.19,1.91) 1.20 (1.06,1.37) 1.24 (0.92,1.68) 1.45 (1.28,1.65) 1.17 (0.98,1.39)
Up 0.76 (0.67,0.86) 0.68 (0.59,0.79) 0.76 (0.69,0.82) 0.57 (0.51,0.64) 0.73 (0.68,0.78) 0.68 (0.62,0.75)
No e: bold indica es OR significan a p <0.05.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248244

This di e ence was ep oduced in he mul ile el models, which
con olled o se e al socioeconomic indi idual and coun y
cha ac e is ics.
Bo h on a ela i e and on an absolu e scale, upwa d mobili y was
associa ed wi h be e heal h, ega dless o wel a e egime ype.
Howe e , on he ela i e scale, hese we e significan o all egime
ypes, whe eas on he absolu e scale he e was only a ue di e -
ence in coun ies om he Fo me USSR. Downwa d mobili y was
gene ally associa ed wi h wo se heal h, bu o di e ing ex en s and
ollowing a less clea pa e n.
I has been a gued be o e ha he use o only absolu e o ela i e
measu es can be misleading, and ou findings ei e a e his a gu-
men (King e al., 2012; Kelly e al., 2007). In ac , ela i e measu es
o heal h inequali ies a e insensi i e o equip opo iona e changes,
while absolu e measu es a e insensi i e o uni o m changes, which
eflec s di e en equi y alue judgmen s implied in he empi ical
analysis (Allanson and Pe ie, 2013). Thus, di e en esul s be ween
ela i e and absolu e scales migh be a consequence o di e en
le els o o e all ill-heal h: o me USSR coun ies had he highes
p e alence o bad o e y bad sel - a ed heal h, making absolu e
di e ences mo e likely o eme ge.
This s udy is sensible o a numbe o limi a ions. The ou come
measu e, sel - a ed heal h, is e y cul u ally-sensi i e, compli-
ca ing c oss-na ional compa isons (Jylh€
a e al., 1998). Ne e heless,
i is an impo an p edic o o mo ali y in e e y socie y whe e i
has been examined (Idle and Benyamini, 1997), making i a much
used and alued heal h measu e. Addi ionally, social mobili y
doesn' ha e a unanimously accep ed ope a ionaliza ion. Di e -
ences in educa ional achie emen a e no necessa ily a eflec ion o
di e en socie al p es ige o access o di e en social esou ces.
Indeed, occupa ional mobili y is o en p e e ed (Belle and Hou ,
2006), bu he occupa ional measu es a ailable in he ESS we e
c ude and di ficul o compa e be ween esponden s and hei
pa en s. Also impo an ly, al hough he analyses con olled o
pa en al educa ion, his ope a ionaliza ion o social mobili y migh
be measu ing p ocesses o accumula ion. The wel a e egime
classifica ion is also deba able. Al hough mos o he egime ypes
ha we e used in his s udy ha e been ex ensi ely used be o e, and
despi e bo h le el 2 and le el 3 a iabili y being e y low and non-
significan (hin ing o a high homogenei y be ween coun ies and
coun y-yea s), hey migh no eflec he cha ac e is ics o wel a e
egimes ha ha e an impac on he ela ionship be ween social
mobili y and heal h. Finally, i is no possible o assess causali y
be ween social mobili y and heal h, conside ing ha ou analyses
a e based on c oss-sec ional da a. Indeed, i is possible ha he
heal h o pa icipan s in ou sample was al eady a ec ed by
mobili y in he p e ious ime pe iod. The associa ion be ween
heal h and socioeconomic s a us is likely dynamic and bi-
di ec ional ac oss he li e cou se.
O e all, i is in e es ing o no e ha al hough all wel a e egime
ypes show ela i e di e ences in bad sel - a ed heal h o up-
wa dly mobile indi iduals, on an absolu e scale he Scandina ian
egime shows he smalles di e ences and he o me USSR g oup
he la ges . Al hough wel a e egime ype seems o accoun o an
impo an pa o he a ia ion in sel -pe cei ed heal h among
Eu opean coun ies (Eikemo e al., 2008b), he ex en o which i
impac s heal h inequali ies has been ques ioned. Mackenbach e al.
(2008) epo ed a su p isingly high deg ee o heal h inequali ies in
no he n Eu opean coun ies, showing ha , despi e egali a ian
policies, li es yle- ela ed isk ac o s emain an impo an cause o
mo ali y inequali ies (Mackenbach e al., 2008). Eikemo e al.
(2008a) also showed a clea g adien o heal h inequali ies
Table 3
Mul ile el models: mobili y om mo he 's educa ion.
Scandina ian Anglo-Saxon Bisma ckian Sou he n Pos communis Fo me USSR
n (le el 1) 30,458 19,752 53,644 29,030 46,556 24,959
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Fixed Pa ame e s
In e cep 0.008 (0.005,0.012) 0.009 (0.006,0.027) 0.011 (0.007,0.018) 0.019 (0.009,0.037) 0.016 (0.011,0.022) 0.036 (0.026,0.050)
Indi idual Le el
Mobili y ( e e ence: s able)
Down 1.39 (1.06,1.83) 1.40 (1.11,1.78) 1.09 (0.89,1.31)* 0.72 (0.43,1.21)* 1.35 (1.13,1.60) 1.01 (0.84,1.22)*
Up 0.77 (0.68,0.86) 0.70 (0.60,0.81) 0.77 (0.71,0.84) 0.56 (0.49,0.62) 0.69 (0.65,0.75) 0.67 (0.61,0.74)
Gende ( e : male) 1.03 (0.92,1.15)* 0.89 (0.77,1.02)* 1.07 (0.98,1.15)* 1.59 (1.45,1.75) 1.06 (0.99,1.13)* 1.21 (1.12,1.32)
Age 0.99 (0.99,1.00)* 1.00 (0.99,1.01)* 1.01 (1.01,1.01) 1.04 (1.04,1.05) 1.04 (1.04,1.04) 1.05 (1.04,1.05)
Domicile ( e : u ban) 0.94 (0.84,1.05)* 0.85 (0.74,0.98) 0.88 (0.82,0.95) 1.14 (1.05,1.24) 1.11 (1.04,1.18) 0.97 (0.89,1.05)*
Main ac i i y ( e : paid wo k)
No ac i i y 6.29 (5.39,7.35) 6.17 (5.00,7.61) 4.35 (3.90,4.84) 2.54 (2.22,2.89) 3.35 (3.06,3.67) 2.63 (2.35,2.93)
O he 2.83 (2.21,3.62) 2.55 (1.96,3.31) 1.82 (1.59,2.09) 1.91 (1.64,2.21) 1.97 (1.74,2.24) 1.56 (1.31,1.86)
Income ( e : li ing com o ably)
Coping 1.40 (1.24,1.59) 1.49 (1.26,1.78) 1.52 (1.38,1.67) 1.38 (1.15,1.64) 1.55 (1.33,1.81) 1.10 (0.84,1.45)*
Di ficul 2.97 (2.49,3.54) 2.61 (2.12,3.21) 3.29 (2.94,3.68) 2.28 (1.90,2.73) 3.00 (2.57,3.51) 1.76 (1.35,2.29)
Ve y di ficul 4.57 (3.56,5.86) 4.09 (3.18,5.27) 5.21 (4.47,6.06) 3.97 (3.28,4.82) 5.37 (4.55,6.32) 2.86 (2.19,3.75)
Mino i y ( e : no) 1.44 (1.05,1.97) 0.86 (0.61,1.21)* 1.27 (1.08,1.49) 1.03 (0.79,1.34)* 0.97 (0.86,1.09)* 1.02 (0.91,1.15)*
Ma i al s a us ( e : ma ied)
Sepa a ed/di o ced 1.37 (1.18,1.59) 2.02 (1.67,2.43) 1.28 (1.14,1.42) 1.23 (1.03,1.47) 1.03 (0.93,1.14)* 1.28 (1.14,1.43)
Widow 1.07 (0.89,1.27)* 1.13 (0.92,1.37)* 1.08 (0.97,1.21)* 1.06 (0.95,1.17)* 1.06 (0.98,1.14)* 1.18 (1.08,1.30)
Single 1.02 (0.87,1.19)* 1.19 (0.99,1.43)* 1.17 (1.05,1.31) 1.29 (1.12,1.51) 1.09 (0.97,1.21)* 1.31 (1.13,1.52)
Mo he 's educa ion ( e : ISCED V/VI)
ISCED I 2.39 (1.76,3.26) 1.63 (1.18,2.25) 2.14 (1.66,2.77) 1.23 (0.79,1.92)* 2.48 (1.93,3.21) 2.02 (1.68,2.43)
ISCED II 2.00 (1.49,2.71) 1.25 (0.91,1.72)* 1.72 (1.34,2.21) 0.67 (0.39,1.13)* 1.89 (1.48,2.41) 1.76 (1.45,2.12)
ISCED III 1.67 (1.25,2.23) 1.20 (0.82,1.77)* 1.34 (1.04,1.71) 1.03 (0.62,1.71)* 1.26 (0.99,1.59)* 1.25 (1.04,1.49)
ISCED IV 1.64 (1.04,2.59) 0.37 (0.17,0.79) 1.58 (1.09,2.27) 0.53 (0.15,1.88)* 1.54 (1.05,2.26) 1.12 (0.91,1.38)*
Le el 2
GDP 0.99 (0.99,1.00)* 0.99 (0.99,1.00)* 1.00 (0.99,1.00)* 0.99 (0.99,0.99) 0.99 (0.99,0.99) 0.99 (0.99,0.99)
Random Pa ame e s
Le el 3 a iance (
s
o
) 0.08 (0.06)* 0.19 (0.19)* 0.22 (0.12)* 0.23 (0.17)* 0.07 (0.04)* 0.03 (0.02)*
Le el 2 a iance (
s
uo
) 0.01 (0.01)* 0 (0) 0 (0) 0.01 (0.01)* 0.03 (0.01) 0.01 (0.01)*
No es: * no significan a p <0.05. ISCED In e na ional S anda d Classifica ion o Educa ion. OR Odds a io. CI Confidence in e al.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248 245
be ween Eu opean wel a e s a es, om Sou he n (wi h he highes
inequali ies) o Bisma ckian (wi h he lowes ) (Eikemo e al.,
2008a). This ‘pa adox’was examined by Mackenbach (2012), who
pos ula ed ha social mobili y migh be one o he d i e s o heal h
inequali ies in Wes e n Eu opean wel a e s a es. Some s udies ha e
indeed shown ha inc eased social mobili y is associa ed wi h
s onge heal h inequali ies (Simons e al., 2013;

Asgei sd
o i and
Ragna sd
o i , 2013; Els ad, 2001). Howe e , B ekke, G un eld and
K e ndokk (2014), showed ha highe heal h inequali ies in mo e
egali a ian coun ies migh be solely a consequence o a mo e equal
heal h dis ibu ion, since he concen a ion index is mo e sensi i e
o heal h-con ingen income ans e s han o income-con ingen
heal h ans e s (B ekke e al., 2012).
Ou findings sugges ha some wel a e s a es a e in ac mo e
Table 4
Mul ile el models: mobili y om a he 's educa ion.
Scandina ian Anglo-Saxon Bisma ckian Sou he n Pos communis Fo me USSR
Le el 1 n 29,837 19,184 52,326 28,417 45,500 23,036
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Fixed Pa ame e s
In e cep 0.009 (0.007,0.014) 0.009 (0.005,0.019) 0.013 (0.009,0.018) 0.014 (0.008,0.024) 0.017 (0.013,0.023) 0.033 (0.023,0.047)
Indi idual Le el
Mobili y ( e e ence: s able)
Down 1.18 (0.95,1.45)* 1.51 (1.19,1.91) 1.20 (1.06,1.37) 1.24 (0.92,1.68)* 1.45 (1.28,1.65) 1.17 (0.98,1.39)*
Up 0.76 (0.67,0.86) 0.68 (0.59,0.79) 0.76 (0.69,0.82) 0.57 (0.51,0.64) 0.73 (0.68,0.78) 0.68 (0.62,0.75)
Gende ( e : male) 1.03 (0.92,1.15)* 0.89 (0.77,1.03)* 1.05 (0.97,1.14)* 1.56 (1.42,1.72) 1.07 (1.00,1.14) 1.26 (1.15,1.37)
Age 0.99 (0.99,1.00)* 1.00 (0.99,1.01)* 1.01 (1.01,1.01) 1.04 (1.04,1.05) 1.04 (1.04,1.04) 1.05 (1.04,1.05)
Domicile ( e : u ban) 0.95 (0.85,1.06)* 0.88 (0.76,1.02)* 0.88 (0.82,0.95) 1.14 (1.04,1.24) 1.11 (1.04,1.18) 0.95 (0.87,1.03)*
Main ac i i y ( e : paid wo k)
No ac i i y 6.11 (5.23,7.14) 5.87 (4.76,7.24) 4.44 (3.97,4.95) 2.54 (2.22,2.91) 3.29 (2.99,3.60) 2.70 (2.41,3.04)
O he 2.67 (2.07,3.45) 2.47 (1.89,3.22) 1.19 (1.64,2.17) 1.94 (1.67,2.26) 1.91 (1.68,2.17) 1.58 (1.32,1.90)
Income ( e : li ing com o ably)
Coping 1.41 (1.24,1.61) 1.45 (1.21,1.72) 1.50 (1.36,1.65) 1.35 (1.13,1.61) 1.55 (1.32,1.81) 1.08 (0.82,1.43)*
Di ficul 2.86 (2.39,3.42) 2.54 (2.06,3.13) 3.25 (2.90,3.65) 2.24 (1.87,2.69) 2.94 (2.51,3.45) 1.71 (1.29,2.26)
Ve y di ficul 4.66 (3.62,6.01) 3.74 (2.89,4.85) 5.05 (4.33,5.90) 3.94 (3.24,4.78) 5.31 (4.49,6.27) 2.83 (2.13,3.76)
Mino i y ( e : no) 1.46 (1.06,2.01) 0.97 (0.69,1.35)* 1.33 (1.13,1.57) 1.03 (0.79,1.35)* 0.97 (0.86,1.09)* 1.01 (0.89,1.15)*
Ma i al s a us ( e : ma ied)
Sepa a ed/di o ced 1.42 (1.34,1.96) 2.06 (1.69,2.49) 1.26 (1.13,1.41) 1.23 (1.03,1.48) 1.04 (0.93,1.16)* 1.25 (1.11,1.42)
Widow 1.08 (0.84,1.34)* 1.17 (0.96,1.43)* 1.06 (0.95,1.19)* 1.06 (0.95,1.18)* 1.07 (0.99,1.16)* 1.18 (1.07,1.30)
Single 1.03 (0.88,1.22)* 1.19 (0.99,1.44)* 1.15 (1.03,1.29) 1.31 (1.13,1.53) 1.08 (0.96,1.21)* 1.15 (0.98,1.35)*
Educa ion ( e : ISCED V/VI)
ISCED I 2.18 (1.70,2.79) 1.59 (1.19,2.12) 1.92 (1.61,2.28) 1.70 (1.24,2.33) 2.27 (1.87,2.74) 2.12 (1.77,2.54)
ISCED II 1.67 (1.29,2.16) 1.09 (0.82,1.45)* 1.59 (1.34,1.88) 1.33 (0.94,1.89)* 1.72 (1.44,2.06) 1.79 (1.49,2.15)
ISCED III 1.42 (1.13,1.78) 1.16 (0.82,1.64)* 1.36 (1.17,1.57) 0.99 (0.68,1.47)* 1.21 (1.02,1.43) 1.39 (1.17,1.67)
ISCED IV 1.67 (1.21,2.29) 0.93 (0.54,1.62)* 1.26 (0.98,1.61)* 1.11 (0.52,2.38)* 0.83 (0.58,1.18)* 1.17 (0.94,1.45)*
Le el 2
GDP 0.99 (0.99,1.00)* 1.00 (0.99,1.00)* 1.00 (0.99,1.00)* 0.99 (0.99,0.99) 0.99 (0.99,0.99) 0.99 (0.99,0.99)
Random Pa ame e s
Le el 3 a iance (
s
o
) 0.09 (0.07)* 0.19 (0.19)* 0.21 (0.12)* 0.24 (0.17)* 0.08 (0.05)* 0.03 (0.02)*
Le el 2 a iance (
s
uo
) 0.01 (0.01)* 0 (0) 0 (0) 0.003 (0.004)* 0.03 (0.01) 0.01 (0.01)*
No es:* no significan a p <0.05. ISCED In e na ional S anda d Classifica ion o Educa ion. OR Odds a io. CI Confidence in e al.
Fig. 2. P obabili y o ‘Bad’o ‘Ve y Bad’sel - a ed heal h pe o mobili y g oup, defined om mo he 's educa ional achie emen , pe wel a e ype (e o ba s a e 95% confidence
in e als) and isk di e ence.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248246
e ec i e in sepa a ing social mobili y om heal h, namely Scan-
dina ian coun ies exhibi smalle di e ences while o me USSR
socie ies he la ges . This is no su p ising conside ing ha , o
example, comp ehensi e social policies seem o be associa ed wi h
ewe inequali ies in ‘sickness’in Eu opean coun ies, as well as
lowe a es o non-employmen ( an de Wel e al., 2011). I is
unde s andable hen, ha Scandina ian coun ies will manage o
disassocia e social mobili y om heal h mo e e ec i ely, leading o
he small isk di e ences we ound o upwa d mobili y in hese
coun ies.
Impo an ly oo, Cen al and Eas e n Eu opean coun ies un-
de wen conside able ansi ions in he las decades, wi h impo -
an consequences o hei social s uc u es (Saa e al., 2012). Ou
esul s o he pos -Communis Eu opean and o me USSR wel a e
egimes, which ended o show he la ges absolu e di e ences in
heal h, migh eflec , a leas pa ially, hese impo an s uc u al
changes and no jus ela i e social mobili y.
Ou findings also ein o ce he need o assess heal h inequali ies
using bo h ela i e and absolu e measu es, since he use o only one
migh be e y misleading.
In e es ingly oo, when compa ing he associa ion o downwa d
mobili y wi h heal h as assessed in e e ence o pa e nal e sus
ma e nal achie emen , he o me was la ge in e e y wel a e
egime excep he Scandina ian egion. Conside ing ha Scandi-
na ian coun ies ha e he bes indica o s o gende equali y
(Eu opean Ins i u e o Gende Equali y, 2013), a possible expla-
na ion o his is ha in o he , less gende -egali a ian coun ies, he
a he 's s a us is mo e decisi e in de e mining he amily's socio-
economic s a us, and he e o e a downwa d mobili y om his so-
cial posi ion has a g ea e impac .
To he bes o ou knowledge, his is he fi s esea ch in o he
mode a ing e ec o wel a e egimes on he ela ionship be ween
social mobili y and heal h, and o measu e social mobili y sepa a ely
based on ma e nal and pa e nal achie emen . Fu he explo a ion o
ou findings would benefi om measu ing occupa ional social
mobili y in addi ion o educa ional mobili y. I would also benefi
om a sepa a e analysis o each gende , since he e ec migh be
di e en o women and men and migh help explain he di e ences
in he associa ion o downwa d mobili y when measu ed on he
basis o ma e nal and pa e nal achie emen .
P e ious s udies ha e ques ioned he con ibu ion o he
wel a e egime in mi iga ing he ex en o heal h inequali ies and
iden ified inc eased social mobili y as a possible cause o his
(Mackenbach, 2012). Howe e , he p esen esul s show ha
impo an sys ema ic di e ences exis be ween egime ypes wi h
ega d o upwa d mobili y and heal h, wi h a no ably a enua ed
associa ion on he absolu e scale in Scandina ian coun ies and
s onge associa ion in he o me USSR egimes. This sugges s ha
social mobili y is no a cause o high heal h inequali ies ound in
Scandina ian coun ies in p e ious analyses.
Acknowledgmen s
I. Campos Ma os is suppo ed by a Fulb igh g an o Public
Heal h esea ch, awa ded by he U.S. Depa men o S a e and he
Po uguese Go e nmen , and by a s uden ship om he Po uguese
Founda ion o Science and Technology.
Appendix A. Supplemen a y da a
Supplemen a y da a ela ed o his a icle can be ound a h p://
dx.doi.o g/10.1016/j.socscimed.2015.08.035.
Re e ences
Allanson, P., Pe ie, D., 2013. On he choice o heal h inequali y measu e o he
longi udinal analysis o income- ela ed heal h inequali ies. Heal h Econ. 22 (3),
353e365.

Asgei sd
o i , T.L., Ragna sd
o i , D.

O., 2013. De e minan s o ela i e and absolu e
concen a ion indices: e idence om 26 Eu opean coun ies. In . J. Equi y
Heal h 12, 53.
Bamb a, C., 2007. ‘Si ing he whea om he Cha ’: a wo-dimensional disc imi-
nan analysis o wel a e s a e egime heo y. Soc. Policy &Adm. 41 (1), 1e28.
Belle , E., Hou , M., 2006. Wel a e s a es and social mobili y: how educa ional and
social policy may a ec c oss-na ional di e ences in he associa ion be ween
occupa ional o igins and des ina ions. Res. Soc. S a i . Mobil. 24 (4), 353e365.
Bonoli, G., 1997. Classi ying wel a e s a es: a wo-dimension app oach. J. Soc. policy
26 (03), 351e372.
B ekke, K.A., G un eld, L., K e ndokk, S., 2012. Asymme ic Impac s o T ans e s o
he Concen a ion Index: Explaining he Heal h Equali y Pa adox. HERO. Uni-
e si y o Oslo.
Case, A., Paxson, C., 2008. Heigh , heal h, and cogni i e unc ion a olde ages. Am.
Econ. Re . 98 (2), 463.
Cook, L., 2010. In: P ess, O.U. (Ed.), Eas e n Eu ope and Russia. The Ox o d Hand-
book o The Wel a e S a e, New Yo k.
Eikemo, T.A., Huisman, M., Bamb a, C., Kuns , A.E., 2008. Heal h inequali ies ac-
co ding o educa ional le el in di e en wel a e egimes: a compa ison o 23
Fig. 3. P obabili y o ‘Bad’o ‘Ve y Bad’sel - a ed heal h pe o mobili y g oup, defined om a he 's educa ional achie emen , pe wel a e ype (e o ba s a e 95% confidence
in e als) and isk di e ence.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248 247
Eu opean coun ies. Sociol. Heal h &Illn. 30 (4), 565e582.
Eikemo, T.A., Bamb a, C., Judge, K., Ringdal, K., 2008. Wel a e s a e egimes and
di e ences in sel -pe cei ed heal h in Eu ope: a mul ile el analysis. Soc. Sci.
Med. 66 (11), 2281e2295.
Els ad, J.I., 2001. Heal h- ela ed mobili y, heal h inequali ies and g adien cons ain
discussion and esul s om a No wegian s udy. Eu . J. Public Heal h 11 (2),
135e140.
Espig Ande sen, G., 1990. In: Dona i, P., P andini, R. (Eds.), The Th ee Wo lds o
Wel a e Capi alism, Ci .
ESS ERIC, 2014. Abou he Eu opean Social Su ey Eu opean Resea ch In as uc u e
eESS ERIC [ci ed 2014 Ap il 22nd]. A ailable om: h p://www.
eu opeansocialsu ey.o g/abou /index.h ml.
Eu opean Ins i u e o Gende Equali y, 2013. Gende Equali y Index: Coun y
P ofiles.
Eu os a , 2014. Gini Coe ficien o Equi alised Disposable Income [ci ed 2014 Ma ch
30]. A ailable om: h p://ec.eu opa.eu/eu os a / gm/ able.do?
ab¼ able&ini ¼1&language¼en&pcode¼ essi190&plugin¼1.
Fa a o, A., San onas aso, P., 1995. E ec s o pa en s' psychological cha ac e is ics
and ea ing beha iou on childhood obesi y and die a y compliance.
J. Psychosom. Res. 39 (2), 145e151.
Fenge , H., 2007. Wel a e egimes in cen al and Eas e n Eu ope: inco po a ing
pos -communis coun ies in a wel a e egime ypology. Con emp. Issues Ideas
Soc. Sci. 3 (2).
Fe e a, M., 1996. The'Sou he n model'o wel a e in social Eu ope. J. Eu . Soc. policy
6 (1), 17e37.
Galoba des, B., Shaw, M., Lawlo , D.A., Lynch, J.W., Smi h, G.D., 2006. Indica o s o
socioeconomic posi ion (pa 1). J. Epidemiol. Commun. Heal h 60 (1), 7e12.
Hallq is , J., Lynch, J., Ba ley, M., Lang, T., Blane, D., 2004. Can we disen angle li e
cou se p ocesses o accumula ion, c i ical pe iod and social mobili y? an
analysis o disad an aged socio-economic posi ions and myoca dial in a c ion
in he S ockholm Hea Epidemiology P og am. Soc. Sci. Med. 58 (8),
1555e1562.
Hemmingsson, T., Lundbe g, I., Dide ichsen, F., 1999. The oles o social class o
o igin, achie ed social class and in e gene a ional social mobili y in explaining
social-class inequali ies in alcoholism among young men. Soc. Sci. Med. 49 (8),
1051e1059.
Idle , E.L., Benyamini, Y., 1997. Sel - a ed heal h and mo ali y: a e iew o wen y-
se en communi y s udies. J. Heal h Soc. Beha . 21e37.
Jylh€
a, M., Gu alnik, J.M., Fe ucci, L., Jokela, J., Heikkinen, E., 1998. Is sel - a ed heal h
compa able ac oss cul u es and gende s? Jou nals Ge on ology Se . B Psychol.
Sci. Soc. Sci. 53 (3), S144eS152.
Kelly M.P., Wo ld Heal h O ganiza ion Commission on he Social De e minan s o
Heal h and Bonne oy J., The Social De e minan s o Heal h: De eloping an E -
idence Base o Poli ical Ac ion, 2007, Uni e sidad del Desa ollo, Chile, and
Na ional Ins i u e o Heal h and Clinical Excellence, Uni ed Kingdom.
King, N.B., Ha pe , S., Young, M.E., 2012. Use o ela i e and absolu e e ec measu es
in epo ing heal h inequali ies: s uc u ed e iew. BMJ 345 e5774.
Libe a os, P., Link, B.G., Kelsey, J.L., 1988. The measu emen o social class in
epidemiology. Epidemiol. Re . 10 (1), 87e121.
Mackenbach, J.P., S i bu, I., Roskam, A.J.R., Schaap, M.M., Men ielle, G., Leinsalu, M.,
Kuns , A.E., 2008. Socioeconomic inequali ies in heal h in 22 Eu opean coun-
ies. N. Engl. J. Med. 358 (23), 2468e2481.
Mackenbach, J.P., 2012. The pe sis ence o heal h inequali ies in mode n wel a e
s a es: he explana ion o a pa adox. Soc. Sci. Med. 75 (4), 761e769.
Ma mo , M., 2004. Men and women beha ing badly? In: Macmillan (Ed.), The
S a us Synd ome: How Social S anding A ec s Ou Heal h and Longe i y. New
Yo k.
Ma mo , M., Wilkinson, R., 2005. Social De e minan s o Heal h. Ox o d Uni e si y
P ess.
Rasbash, J., Cha l on, C., B owne, W.J., Healy, M., Came on, B., 2013. MLwiN Ve sion
2.28. Cen e o Mul ile el Modelling, Uni e si y o B is ol.
Saa , E., Un , M., 2012. Labou ma ke e u ns o e ia y educa ion in pos -socialis
coun ies. In: Lambe , P. (Ed.), Social S a ifica ion: T ends and P ocesses.
Ashga e Publishing, L d, England.
Simons, A.M., G o en, D.A., Bosma, H., 2013. Socio-economic inequali ies in all-
cause mo ali y in Eu ope: an explo a ion o he ole o heigh ened social
mobili y. Eu . J. Public Heal h ck 145.
Singhamme , J., Mi elma k, M.B., 2010. S anda d measu es o in e -gene a ional
social mobili y dis o ac ual pa e ns o mobili y and heal h beha iou : e i-
dence o a be e me hodology. C i . Public Heal h 20 (2), 223e232.
Sub amanian, S., Jones, K., Duncan, C., 2003. Mul ile el Me hods o Public Heal h
Resea ch: Neighbo hoods and Heal h. Ox o d Uni e si y P ess, New Yo k.
an de Wel, K.A., Dahl, E., Thielen, K., 2011. Social inequali ies in ‘sickness’: Eu o-
pean wel a e s a es and non-employmen among he ch onically ill. Soc. Sci.
Med. 73 (11), 1608e1617.
Wes , P., 1991. Re hinking he heal h selec ion explana ion o heal h inequali ies.
Soc. Sci. Med. 32 (4), 373e384.
Wo ld Bank, 2014. GDP pe capi a, PPP (Cu en In e na ional Dolla s) [ci ed 2014
Ma ch 30 h]. A ailable om: h p://da a.wo ldbank.o g/indica o /NY.GDP.PCAP.
PP.CD.
I. Campos-Ma os, I. Kawachi / Social Science & Medicine 142 (2015) 241e248248
Table 1 Desc ip ion o main esul s o eligible publica ions, acco ding o SES a iable and heal h ou come used
Heal h ou come
Medical indica o s Func ional indica o s Subjec i e heal h
Socioeconomic
de e minan s
Place o esidence Physical heal h ended o be
be e among u al adolescen s
(Machado-Rod igues, 2012,
Machado-Rod igues, 2011) and
less dep i ed neighbo hoods
(Bas os, 2013). Pa en al
pe cep ions o be e
neighbo hood en i onmen s also
ended o show an associa ion
wi h be e physical (Noguei a,
2013a, Machado-Rod igues, 2014)
bu wo se men al heal h
(Ca alho, 2014) among child en.
The only s udy (Nunes, 2010)
showed no associa ion
be ween place o esidence
and cogni i e abili y.
One s udy (Humbold , 2014)
showed ha li e sa is ac ion was
be e in u al a eas.
Race/e hnici y/
cul u e/language
Mig an s showed highe
mo ali y (Ha ding, 2008,
Williamson, 2009), wo se o al
heal h (Pe ei a, 2013) and a
highe pe cen age o small
p e e m bi hs (Ha ding, 2006b).
On he o he hand, mig an
adolescen s had less men al
heal h p oblems (Ne o, 2009
and Ne o, 2010) and be e
ca dio espi a o y i ness
(San os, 2011).
The e we e di e ences in SRH
among na ionali ies in one s udy
(Dias, 2013), bu all o he s udies
showed no associa ion be ween
mig a ion, e hnici y o na ionali y
and subjec i e heal h (Malmusi,
2014 and Humbold , 2014).
Occupa ion Mos s udies showed a s ong
associa ion be ween
unemploymen o less
di e en ia ed occupa ions and
wo se heal h (see, o example,
F aga, 2014 o San os, 2008),
al hough some ound no
associa ion ( o example Al es,
2012 o Bas os, 2013). None
ound an opposi e esul .
One s udy (Aze edo, 2012) ound
people who we e unemployed
o e i ed we e mo e likely o
su e om ch onic pain.
Sil a (2014) showed s ong
associa ions be ween
employmen and mo e
di e en ia ed occupa ions wi h
SRH. On he o he hand,
Humbold (2014) ound no
associa ion be ween
employmen and li e sa is ac ion.
Gende /sex Almos all s udies showed an
associa ion be ween being
emale and wo se heal h (see,
o example, San os, 2011 o
Bulhões, 2013). Some s udies
ound no gende di e ences
(see, o example, Bas os, 2013 o
Ne o, 2010) and wo ound he
opposi e associa ion (Pe elman,
2012 and S ewa -Knox, 2012).
Women we e mo e likely o ake
sickness absence (Mas e kaasa,
2014 and Pe elman, 2012) and
epo ch onic pain (Aze edo,
2012 and Pe elman, 2012), and
one s udy showed men epo ed
mo e bed days (Pe elman, 2012).
Cogni i e abili ies di e ed
be ween gende s, depending on
he es used (Ma ins, 2012,
San os, 2014a).
Almos e e y s udy showed
women had wo se subjec i e
heal h ou comes (see, o
example, Bamb a, 2009, Dias,
2013 o Pe ei a, 2011).
Religion One s udy showed no
associa ion be ween eligion o
spi i uali y and he onse o
majo dep ession (Leu en , 2013).
One s udy showed eligious
people showed highe li e
sa is ac ion (Humbold , 2014),
and ano he showed no
associa ion be ween eligion o
spi i uali y and quali y o li e o
well-being (Vilhena, 2014).
Educa ion Lowe educa ion ended o show
a s ong associa ion wi h wo se
heal h in almos all s udies (see,
o example, Bas os, 2013 o
San os, 2010). The e we e wo
excep ions: Lawlo , 2005, who
showed ha insulin esis ance
was mo e common in child en
o mo e educa ed pa en s and
Cos a, 2008, who showed gi ls
whose pa en s we e mo e
Educa ion was s ongly
associa ed wi h cogni i e abili y
(Ma ins, 2012, Nunes, 2010 and
San os, 2014a), ch onic pain
(Aze edo, 2012) and unc ional
limi a ions (Eikemo, 2008,
Knesebeck, 2006).
Be e SRH was associa ed wi h
highe educa ion in all s udies
(see, o example, Knesebeck,
2006 o Sil a, 2014) excep one,
ha showed he opposi e
(Humbold , 2014).
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 5 o 10

Table 1 Desc ip ion o main esul s o eligible publica ions, acco ding o SES a iable and heal h ou come used (Con inued)
educa ed had mo e ea ing
diso de symp oma ology.
Socioeconomic
s a us
Ma ied indi iduals ended o
show be e heal h ou comes
(see, o example, Ha ding, 2008
o Williamson, 2009), bu had
highe odds o being obese
(Al es, 2012 and Goulão, 2015).
Income, dep i a ion and inancial
di icul ies showed con lic ing
esul s: while mos s udies
ended o show wo se heal h
ou comes o mo e dep i ed
people (see, o example, Pe ei a,
2013 o Al es, 2012) o no
associa ion a all (see, o
example, Co eia, 2014 o
Pimen a, 2011), he e we e some
excep ions ha showed, o
example, lowe p e alence o
obesi y among homeless people
(Oli ei a, 2012) o mo e insulin
esis ance among child en wi h
iche pa en s (Lawlo , 2005).
One s udy (Aze edo, 2012) ound
no associa ion be ween ma i al
s a us and ch onic pain. Ea ly li e
SES, as measu ed by heigh , was
s ongly associa ed wi h ch onic
pain in women (Pe elman, 2014).
Objec i e income (Humbold ,
2014, Sil a, 2014) and pe cei ed
income (Dias, 2013) we e ound
o be associa ed wi h subjec i e
heal h, bu no ma i al s a us
(Humbold , 2014) o heigh , as a
measu e o ea ly li e SES
(Pe elman, 2014).
Social capi al One s udy (Fe ei a-Valen e, 2014)
showed ha social suppo was
associa ed wi h be e
psychological unc ioning.
One s udy (Fe ei a-Valen e, 2014)
showed ha social suppo had a
s ong associa ion wi h physical
unc ioning, bu no pain
in ensi y.
Numbe o ac i i ies ou side he
home was he only social capi al
indica o ha showed an
associa ion wi h SRH (Sil a, 2014).
O he analyses showed no
associa ion (Vilhena, 2014,
Sil a, 2014).
No e: no eligible publica ion explo ed he ela ionship be ween ‘ ace/e hnici y/cul u e/language’o ‘ eligion’and unc ional indica o s
Legend: SRH Sel Ra ed Heal h. SES Socioeconomic S a us
Fig. 2 Diag am ep esen ing main esul s o he associa ions ound in he eligible publica ions. The isual aspec o he diag am, bu no he ules o
i s cons uc ion, was based on he diag am buil by Ashley EA e al., “Clinical assessmen inco po a ing a pe sonal genome”The Lance 375(2010):
1525-35. No e: Fon size o heal h ou comes and ci cle size o socioeconomic de e minan s a e p opo ional o he numbe o eligible publica ions in
which hey ea u ed. Black a ows ep esen s ong e idence o an associa ion be ween socioeconomic indica o and heal h ou come; g ey a ows
ep esen weak e idence and dashed a ows ep esen e idence o he “nega i e”associa ions. In he esul s ob ained, “nega i e”includes mig an
popula ions ha ing be e men al heal h and ma ied indi iduals ha ing highe p e alence o obesi y. E idence o all o he associa ions had a “posi i e”
di ec ion, i.e., ill heal h was associa ed wi h lowe educa ion, lowe income, emale gende , unemploymen , dep i a ion, ha ing less di e en ia ed
occupa ions and li ing in an un a ou able o u ban a ea. De ails on how his diag am was cons uc ed a e in he online Addi ional ile 3
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 6 o 10
Educa ion was he mos equen ly s udied de e minan
o heal h and o which mos e idence exis s o heal h in-
equali ies. E idence o educa ional inequali ies in obesi y
was pa icula ly common, especially o women, as he
wo s udies ha s a i ied he analysis by gende ound
only women showed signi ican inequali ies [43, 44]. This
sugges s educa ional inequali ies in o e weigh /obesi y a e
ound mos ly o exclusi ely in women. This is no
unique o Po ugal: Roskam e al. (2010) ound ha
o he sou he n Eu opean coun ies also show high
educa ion inequali ies in o e weigh and obesi y only
o women [57]. In his analysis, Po ugal had he highes
educa ional inequali ies in o e weigh and obesi y among
women in all he coun ies analysed. This can be a conse-
quence o a ious ac o s, such as inequali ies in physical
ac i i y, die a y pa e ns o pa i y. Howe e , bo h men and
women seem o show he same ex en o educa ional in-
equali ies in physical ac i i y and die in Po ugal [58, 59],
which makes hem unlikely ac o s in explaining inequal-
i ies in obesi y seen mos ly in women. On he o he hand,
women wi h lowe educa ion in Po ugal ha e a highe
e ili y index [60], and since highe pa i y is s ongly
associa ed wi h obesi y [61], his migh be he mos
sui able explana ion o he high educa ional inequali ies
in o e weigh and obesi y seen o women in Po ugal.
Educa ion was also s ongly associa ed wi h SRH
[28, 39, 47–49], which is consis en wi h o he in e -
na ional analyses [49, 62]. In e es ingly a Eu opean com-
pa ison among 22 coun ies ound ha Po uguese men
showed he highes educa ion inequali ies in SRH when
compa ed o o he coun ies [49]. Howe e , educa ional in-
equali ies in SRH should be in e p e ed wi h cau ion. As
Huisman, Len he and Mackenbach (2007) poin ed ou , he
p edic i e abili y o SRH o mo ali y a ies signi ican ly
among educa ional g oups o men [63]. This p obably e-
lec s educa ional di e ences in men’s heal h pe cep ion,
biasing he answe s o ques ions on subjec i e heal h.
Ou e iew also sugges ed s ong gende inequali ies in
bo h SRH and men al heal h symp oms. Gende - ela ed
heal h inequali ies is a b oad and complex opic. Despi e
he p e ailing no ion ha men ha e highe mo ali y and
women highe mo bidi y [64], his has been challenged in
he li e a u e, and con adic o y pa e ns con inue o ap-
pea [65, 66]. Addi ionally, gende inequali ies in heal h
a e p obably a esul o mul iple ac o s, including bio-
logical and social [67], which aises ques ions o whe he
hey should be conside ed as un ai o as una oidable.
Despi e his, almos e e y publica ion ha explo ed gen-
de di e ences in ou e iew showed s ongly a ou able
esul s o men, pa icula ly o men al heal h symp oms
and SRH [32]. No iceably, no publica ion explo ed gende
di e ences in mo ali y.
Academic a en ion o heal h inequali ies in Po ugal
has ended o ocus on speci ic opics. Gende and
educa ion a e by a he mos commonly used SES indi-
ca o s, possibly because hey a e he mos easily measu -
able, commonly used in su eys wi h high esponse a es
and high alidi y o answe s and a e less a ec ed by
e e se causa ion. Twel e publica ions also looked a
heal h inequali ies be ween mig an s and Po uguese
na i es; his is su p ising conside ing Po ugal is one o
he Eu opean coun ies wi h he lowes p opo ion o
mig an popula ion among i s esiden s [68]. This could
be impu ed o bo h he ease o measu emen o his
a iable and he p esence o esea ch g oups in he
coun y in es iga ing his subjec .
O he SES indica o s appea o ha e been o e looked.
Fo example, despi e he g owing li e a u e on he e ec
o place in heal h, only a ew publica ions explo ed his
opic, mos o which ocused on u al/u ban di e ences.
The e was also a no able de iciency o s udies o social
capi al and po e y, despi e Po ugal’s high income in-
equali y [6] and conside able isk o po e y and social ex-
clusion [69]. Addi ionally, despi e he g owing ecogni ion
o he ime dimension in he building o heal h inequal-
i ies [70], no publica ion ook a li e cou se app oach o
how SES indica o s migh a ec heal h. This, coupled wi h
he sca ci y o longi udinal s udies, subs an ially p ecludes
he possibili y o assessing causal ela ionships. This also
speaks o a e y sca ce ocus on he elde ly - o he 71 eli-
gible publica ions, only 7 ocused on olde people, which
is su p ising in a coun y whe e he old-age dependency
a io was he i h highes in Eu ope in 2014 [71].
In 2013, he majo causes o dea h in Po ugal we e dis-
eases o he ci cula o y sys em (30), malignan umou s
(24), diseases o he espi a o y sys em (12), and endo-
c ine, nu i ional and me abolic diseases (5 %) [72]. In his
sense, despi e malignan umou s being he second mos
common cause o dea h, a e ci cula o y diseases, he e
a e s ikingly ew publica ions ocusing on his heal h
issue ( ou , o which wo a e ecological). This migh again
e lec he absence o a na ionally o ien ed esea ch policy,
in pa a ibu able o absence o poli ical a en ion o his
issue [3, 4, 8]. This is also he case o espi a o y diseases,
which a e also almos absen om ou analysis. In a ecen
epo o a conso ium published by he Eu opean
Commission on Heal h Inequali ies, Po ugal was de-
sc ibed as ha ing “[clea ] di icul ies in measu ing and
analysing heal h inequali ies”[73] (page 129). In e es -
ingly, he cu en Po uguese Na ional Heal h Plan iden i-
ies he educ ion o child obesi y as one o i s ou goals
o 2020, bu wi h no ocus on i s unequal dis ibu ion
among socioeconomic g oups [74]. This plan does men-
ion he impo ance o he social de e minan s o heal h,
bu ocuses almos exclusi ely on he access o heal h ca e
se ices as a emedy o heal h inequali ies [74].
The limi ed a en ion gi en o heal h inequali ies in
Po ugal can only be explained wi h an ex ensi e
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 7 o 10
explo a ion o mul iple ac o s, bu one o hese ac o s
is p obably he eng ained belie ha he Na ional Heal h
Se ice, as a uni e sal and ( ela i ely) inexpensi e se -
ice a poin o ca e, is enough o ace hese inequali ies.
Howe e , his is appa en ly no ue, as his e iew has
shown he e a e s ill impo an heal h inequali ies in
Po ugal. Tackling hese inequali ies will demand an im-
po an e o o build an o ganized esea ch and policy
s a egy ha will ha e o go beyond he Na ional Heal h
Se ice. I is impo an o no ice ha Po ugal is amongs
he mos unequal coun ies in Eu ope, so ha i could
bene i om a mo e p og essi e axa ion scheme and
highe social p o ec ion o he poo es , which a e majo
e idence-based and consensual measu es o igh inequal-
i ies in heal h [75].
Limi a ions
This e iew ied o b ing oge he analyses no always
compa able among hem. In ac , many o hese publica-
ions ocused on speci ic popula ions –mig an s, child en
o ce ain egions in Po ugal – ha migh ha e pa icula
pa e ns o heal h inequali ies. This migh ha e hidden
inequali ies ha a e no appa en when all g oups a e
pooled oge he . Ou sea ch s a egy migh ha e also
excluded impo an publica ions, namely in e na ional
compa isons ha included a Po uguese sample no
speci ically men ioned. Howe e , we ied o o e come
his by sea ching o publica ions by esea che s known
o ha ing published in his a ea. The quali y o he ana-
lyses in he e iewed publica ions was ound o be he e o-
geneous, wi h some p esen ing highly eliable analyses
and o he s elying on ‘con enience samples’,o onsmall
sample sizes. Following he PRISMA guidelines, we chose
no o sco e no selec he publica ions based on ‘quali y’,
bu o ca y ou a b ie assessmen o s eng hs and limi a-
ions on each ( able in Addi ional ile 2). Also, we ocused
ou e iew on pape s published in indexed pee - e iewed
jou nals acco ding o good p ac ices o scien i ic esea ch,
bu his may ha e excluded impo an publica ions, in
pa icula om he g ey li e a u e.
Finally, we es ic ed ou analysis o heal h ou comes,
and did no conside media ing ac o s such as li es yle
and heal hca e use. Also, we did no conside s udies on
in e en ions o dec ease inequali ies in heal h. We
adop ed his s a egy o a oid a oo la ge scope o he
e iew, which would ha e complica ed he iden i ica ion
o gene al ends and in e p e a ions. Fu he esea ch
should ocus on hese connec ed issues.
Along his pape , we e e ed o “inequali ies”in heal h
ins ead o o he possible e ms such as “inequi y”o
“di e ences”. In pa icula , inequi y e e s o di e ences
ha a e unjus , un ai and a oidable [76]. This op ion
was made because he concep o inequali y is mo e
neu al in e ms o in e p e a ions and alue judgemen s,
whe eas he e m “inequi y”implies s ong assump ions
abou he causes o di e ences, which none o he
e iewed pape s could con i m. Addi ionally, mos
e iewed pape s e e ed o inequali ies in heal h, so we
op ed o be ai h ul o au ho s’in e p e a ions.
Conclusions
We ha e shown ha he e is s ong e idence o socioeco-
nomic heal h inequali ies in Po ugal and compa a i e
analyses show ha hese a e possibly one o he highes
among Eu opean coun ies. We iden i ied educa ion and
gende as he main de e minan s o heal h inequali ies,
a ec ing mos ly he dis ibu ion o obesi y, sel - a ed
heal h and men al heal h symp oms. The publica ions we
iden i ied also e lec he absence o a na ionally o ien ed
esea ch s a egy on heal h inequali ies ocusing on he
mos p e alen diseases (such as malignan umou s and
espi a o y diseases), de e mining ac o s o inequali ies
(li ing con ex s, po e y o social capi al) and ulne able
popula ions (such as he elde ly). We hope his e iew will
help guide decision-making o ackle hese issues, as has
long been ecommended.
Addi ional iles
Addi ional ile 1: De ailed sea ch s a egy. (PDF 79 kb)
Addi ional ile 2: Table o ex ac ed da a om he 71 eligible
publica ions. (PDF 196 kb)
Addi ional ile 3: Desc ip ion o ules adop ed o build diag am in
Fig. 2 o he main ex . (DOCX 16 kb)
Addi ional ile 4: Comple e lis o he se en y one eligible publica ions
iden i ied by he sys ema ic e iew, by alphabe ical o de . (PDF 70 kb)
Compe ing in e es s
The au ho s ha e no compe ing in e es s o decla e.
Au ho s’con ibu ions
ICM con ibu ed o s udy concep ualisa ion and design, da a acquisi ion,
selec ion and analysis, and d a ed he manusc ip . GR con ibu ed o s udy
concep ualisa ion and design, da a selec ion and manusc ip e ision. JP
con ibu ed o s udy concep ualisa ion and design, da a selec ion and
ex ac ion and manusc ip e ision. All au ho s app o ed he inal manusc ip .
Acknowledgemen s
The au ho s wish o hank he con ibu ion o he esea che s om he No a
Heal hca e Ini ia i e, who p o ided commen s on a la e d a o his pape .
The inal pape is he esponsibili y o he au ho s.
This p ojec has been inanced by he Fundação pa a Ciência e Tecnologia
(G an VIH/SAL/0065/2011).
Au ho de ails
1
Ins i u o de Higiene e Medicina T opical, Uni e sidade NOVA de Lisboa,
Lisbon, Po ugal.
2
Cen o de In es igação em Saúde Pública, Lisbon, Po ugal.
3
Global Heal h and T opical Medicine, Ins i u o de Higiene e Medicina
T opical, NOVA Uni e si y o Lisbon, Lisbon, Po ugal.
4
Escola Nacional de
Saúde Pública, Uni e sidade NOVA de Lisboa, Lisbon, Po ugal.
Recei ed: 23 No embe 2015 Accep ed: 1 Feb ua y 2016
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 8 o 10
Re e ences
1. Ma mo M, F iel S, Bell R, Houweling TA, Taylo S. Closing he gap in a
gene a ion: heal h equi y h ough ac ion on he social de e minan s o
heal h. Lance . 2008;372(9650):1661–9.
2. Black D, Mo is J, Smi h C, Townsend P. Inequali ies in heal h: epo o a
Resea ch Wo king G oup. London: Depa men o Heal h and Social
Secu i y; 1980. p. 19.
3. Pe ei a J, Fu ado C. Equidade em saúde: do conhecimen o à acção. In: A
No a Saúde Pública: A saúde Pública na E a do Conhecimen o, edi o .
DGS/ENSP. Lisbon: G adi a; 2011.
4. Wo ld Heal h O ganiza ion. WHO e alua ion o he na ional heal h plan o
Po ugal (2004–2010). Copenhagen: WHO Regional O ice o Eu ope; 2010.
5. Wo ld Heal h O ganiza ion. Po ugal Heal h Sys em Pe o mance
Assessmen . WHO and Al o Comissa iado da Saúde [Po uguese High
Comissione o Heal h]. 2010.
6. Eu opean Commission. Social Si ua ion Moni o - Income inequali y in EU
coun ies. 2015. h p://ec.eu opa.eu/social/main.jsp?ca Id=1050&in PageId=
1870&langId=en. Accessed 22 No embe 2015.
7. Mackenbach JP, S i bu I, Roskam AJ, Schaap MM, Men ielle G, Leinsalu M,
e al. Socioeconomic inequali ies in heal h in 22 Eu opean coun ies.
N Engl J Med. 2008;358(23):2468–81.
8. Bago d’U a T. Equidade no sec o da saúde em Po ugal. In: Almedina E,
edi o . 30 Anos do Se iço Nacional de Saúde - um pe cu so comen ado.
Lisbon: Simões, Jo ge; 2010.
9. Mohe D, Libe a i A, Te zla J, Al man DG. P e e ed epo ing i ems o
sys ema ic e iews and me a-analyses: he PRISMA s a emen . BMJ.
2009;349:g7716.
10. O’Neill J, Tabish H, Welch V, Pe ic ew M, Po ie K, Cla ke M, e al. Applying
an equi y lens o in e en ions: using PROGRESS ensu es conside a ion o
socially s a i ying ac o s o illumina e inequi ies in heal h. J Clin Epidemiol.
2014;67(1):56–64.
11. Blax e M. A compa ison o measu es o inequali y in mo bidi y. Gowe :
Heal h inequali ies in Eu opean coun ies Alde sho ; 1989. p. 199–230.
12. Public Heal h Ins i u e o he Uni e si y o Po o [Ins i u o de Saúde Pública
da Uni e sidade do Po o]. Coho s: EPIPo o Po o, Po ugal: ISPUP.
h p://ispup.up.p / esea ch/ esea ch-s uc u es/coho s/?lang=en. Accessed
22 No embe 2015.
13. Alda de Cae ano Ca alho & José Pe ei a Miguel. Inqué i o Nacional de
Saúde 2005/2006. Ins i u o Nacional de Es a ís ica and Ins i u o Nacional de
Saúde D . Rica do Jo ge. 2009.
14. Public Heal h Ins i u e o he Uni e si y o Po o [Ins i u o de Saúde Pública
da Uni e sidade do Po o]. Coho s: Gene a ion XXI Po o, Po ugal: ISPUP.
h p://ispup.up.p / esea ch/ esea ch-s uc u es/coho s/ge acaoxxi/.
Accessed 22 No embe 2015.
15. Public Heal h Ins i u e o he Uni e si y o Po o [Ins i u o de Saúde Pública
da Uni e sidade do Po o]. Coho s: EPITeen Po o, Po ugal: ISPUP.
h p://ispup.up.p / esea ch/ esea ch-s uc u es/coho s/epi een/.
Accessed 22 No embe 2015.
16. Mi anda MJ, Cos a C, San ana P, Ba ozo LV. Associação espacial en e
a iá eis socioeconômicas e isco ela i o de nascimen os p é- e mo na
Região Me opoli ana de São Paulo (RMSP) e na Á ea Me opoli ana de
Lisboa (AML). Saúde Soc. 2014;23(4):1142–53.
17. Ha ding S, Bo ouje di M, San ana P, C uickshank J. Decline in, and lack o
di e ence be ween, a e age bi h weigh s among A ican and Po uguese
babies in Po ugal. In J Epidemiol. 2006;35(2):270–6.
18. Oli ei a C, Economou T, Bailey T, Mendonça D, Pina M. The in e ac ions
be ween municipal socioeconomic s a us and age on hip ac u e isk.
Os eopo os In . 2015;26(2):489–98.
19. on Humbold S, Leal I, Pimen a F. Li ing well in la e li e: he in luence o
sense o cohe ence, and socio-demog aphic, li es yle and heal h- ela ed
ac o s on olde adul s’sa is ac ion wi h li e. Applied Resea ch in Quali y o
Li e. 2014;9(3):631–42.
20. Bas os J, Pele ei o B, Ba os R, Al es L, Se e o M, Fá ima Pina M, e al.
Sociodemog aphic de e minan s o p e alence and incidence o Helicobac e
pylo i in ec ion in Po uguese adul s. Helicobac e . 2013;18(6):413–22.
21. Williamson LM, Rosa o M, Teyhan A, San ana P, Ha ding S. AIDS mo ali y in
A ican mig an s li ing in Po ugal: e idence o la ge social inequali ies.
Sex T ansm In ec . 2009;85(6):427–31.
22. Ha ding S, Teyhan A, Rosa o M, San ana P. All cause and ca dio ascula
mo ali y in A ican mig an s li ing in Po ugal: e idence o la ge social
inequali ies. Eu J Ca dio asc P e Rehabil. 2008;15(6):670–6.
23. Ne o F. P edic o s o men al heal h among adolescen s om immig an
amilies in Po ugal. J Fam Psychol. 2009;23(3):375.
24. Ne o F. Men al heal h among adolescen s om e u ned Po uguese
immig an amilies om No h Ame ica. N Am J Psychol. 2010;12(2):265.
25. San os DA, Sil a AM, San a-Cla a H, Ma ias CN, Fields DA, Sa dinha LB.
De e minan ac o s o ca dio espi a o y i ness in Po uguese adolescen s
o di e en e hnici ies. Re is a B asilei a de Cinean opome ia & Desempenho
Humano. 2011;13(4):243–9.
26. F aga S, Ma ques-Vidal P, Vollenweide P, Waebe G, Guessous I, Paccaud F,
e al. Associa ion o socioeconomic s a us wi h in lamma o y ma ke s: A wo
coho compa ison. P e en i e medicine. 2014.
27. Ha ding S, San ana P, C uickshank JK, Bo ouje di M. Bi h weigh s o black
A ican babies o mig an and nonmig an mo he s compa ed wi h hose o
babies o Eu opean mo he s in Po ugal. Ann Epidemiol. 2006;16(7):572–9.
28. Sil a PA. Indi idual and social de e minan s o sel - a ed heal h and well-
being in he elde ly popula ion o Po ugal. Cade nos de saude publica.
2014;30(11):2387–400.
29. Sousa-Ribei o M, S e ke M, Coimb a JL. Pe cei ed quali y o he psychosocial
en i onmen and well-being in employed and unemployed olde adul s: he
impo ance o la en bene i s and en i onmen al i amins. Econ Ind Democ .
2014;35(4):629–52.
30. Aze edo LF, Cos a-Pe ei a A, Mendonça L, Dias CC, Cas o-Lopes JM.
Epidemiology o ch onic pain: a popula ion-based na ionwide s udy on i s
p e alence, cha ac e is ics and associa ed disabili y in Po ugal. The jou nal
o pain. 2012;13(8):773–83.
31. Al es E, Co eia S, Ba os H, Aze edo A. P e alence o sel - epo ed
ca dio ascula isk ac o s in Po uguese women: a su ey a e deli e y.
In e na ional jou nal o public heal h. 2012;57(5):837–47.
32. Pe elman J, Fe nandes A, Ma eus C. Gende dispa i ies in heal h and
heal hca e: esul s om he Po uguese Na ional Heal h In e iew Su ey.
Cade nos de saude publica. 2012;28(12):2339–48.
33. S ewa -Knox B, Du y ME, Bun ing B, Pa H, de Almeida MDV, Gibney M.
Associa ions be ween obesi y (BMI and wais ci cum e ence) and socio-
demog aphic ac o s, physical ac i i y, die a y habi s, li e e en s, esilience,
mood, pe cei ed s ess and hopelessness in heal hy olde Eu opeans.
BMC Public Heal h. 2012;12(1):424.
34. Ca alho J, Nob e P. Gende issues and sexual desi e: he ole o emo ional
and ela ionship a iables. J Sexual Med. 2010;7(7):2469–78.
35. Cos a C, Ramos E, Se e o M, Ba os H, Lopes C. De e minan s o ea ing
diso de s symp oma ology in Po uguese adolescen s. A chi es o pedia ics
& adolescen medicine. 2008;162(12):1126–32.
36. Lawlo DA, Ha o M, Wedde kopp N, Ande sen LB, Sa dinha LB, Riddoch CJ,
e al. Associa ion o socioeconomic posi ion wi h insulin esis ance among
child en om Denma k, Es onia, and Po ugal: c oss sec ional s udy. BMJ.
2005;331(7510):183.
37. Co eia S, Ba os H. Small- o -ges a ional age Po uguese babies: he e ec
o childhood social en i onmen , g ow h and adul socioeconomic
condi ions. P e Med. 2015;70:102–7.
38. Mello T, An unes I, Waldman E, Ramos E, Rel as M, Ba os H. P e alence and
se e i y o den al ca ies in schoolchild en o Po o, Po ugal. Communi y
Den Heal h. 2008;25(2):119–25.
39. Schü e S, Chas ang J-F, Pa en -Thi ion A, Ve meylen G, Niedhamme I.
Social di e ences in sel - epo ed heal h among men and women in 31
coun ies in Eu ope. Scandina ian jou nal o public heal h. 2013;41(1):51–7.
40. Co eia S, Rod igues T, Ba os H. Socioeconomic a ia ions in emale e ili y
impai men : a s udy in a coho o Po uguese mo he s. BMJ open.
2014;4(1):e003985.
41. Pimen a F, Leal I, Ma oco J, Ramos C. Pe cei ed con ol, li es yle, heal h,
socio-demog aphic ac o s and menopause: Impac on ho lashes and
nigh swea s. Ma u i as. 2011;69(4):338–42.
42. Fe ei a-Valen e MA, Pais-Ribei o JL, Jensen MP. Associa ions be ween
psychosocial ac o s and pain in ensi y, physical unc ioning, and
psychological unc ioning in pa ien s wi h ch onic pain: a c oss-cul u al
compa ison. Clin J Pain. 2014;30(8):713–23.
43. Camões M, Lopes C, Oli ei a A, San os AC, Ba os H. O e all and cen al
obesi y incidence in an u ban Po uguese popula ion. P e Med.
2010;50(1):50–5.
44. San os A-C, Ba os H. P e alence and de e minan s o obesi y in an u ban
sample o Po uguese adul s. Public Heal h. 2003;117(6):430–7.
45. Noguei a H, Fe ão M, Gama A, Mou ão I, Rosado Ma ques V, Padez C.
Pe cep ions o neighbo hood en i onmen s and childhood obesi y:
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 9 o 10
e idence o ha m ul gende inequi ies among Po uguese child en.
Heal h & place. 2013;19:69–73.
46. Bingham DD, Va ela‐Sil a MI, Fe ão MM, Augus a G, Mou ão MI, Noguei a H,
e al. Socio‐demog aphic and beha io al isk ac o s associa ed wi h he high
p e alence o o e weigh and obesi y in po uguese child en. Am J Hum Biol.
2013;25(6):733–42.
47. Dias S, Gama A, Ma ins MO. Heal h s a us and p e en a i e beha io s o
immig an s by gende and o igin: a Po uguese c oss‐sec ional s udy.
Nu sing Heal h Sci. 2013;15(3):309–17.
48. Eikemo TA, Huisman M, Bamb a C, Kuns AE. Heal h inequali ies acco ding
o educa ional le el in di e en wel a e egimes: a compa ison o 23
Eu opean coun ies. Sociology o heal h & illness. 2008;30(4):565–82.
49. Von Dem Knesebeck O, Ve de PE, D agano N. Educa ion and heal h in 22
Eu opean coun ies. Soc Sci Med. 2006;63(5):1344–51.
50. Pe ei a M, Cana a o MC. Gende and age di e ences in quali y o li e and
he impac o psychopa hological symp oms among HIV-in ec ed pa ien s.
AIDS Beha . 2011;15(8):1857–69.
51. Bamb a C, Pope D, Swami V, S anis ee D, Roskam A, Kuns A, e al. Gende ,
heal h inequali ies and wel a e s a e egimes: a c oss-na ional s udy o 13
Eu opean coun ies. J Epidemiol Communi y Heal h. 2009;63(1):38–44.
52. Ribei o O, Teixei a L, A aújo L, A onso RM, Pachana N. P edic o s o anxie y
in cen ena ians: heal h, economic ac o s, and loneliness. In e na ional
Psychoge ia ics. 2014. p. 1–10.
53. Ama al MOP, de Figuei edo Pe ei a CM, Ma ins DIS, Sakella ides CT.
P e alence and isk ac o s o insomnia among Po uguese adolescen s.
Eu J Pedia . 2013;172(10):1305–11.
54. Ca alho M, de Ma os MG. Psychosocial de e minan s o men al heal h and
isk beha iou s in adolescen s. Global J Heal h Sci. 2014;6(4):22.
55. Bulhões C, Ramos E, Linde J, Dias S, Ba os H. Dep essi e symp oms and
i s associa ed ac o s in 13-yea -old u ban adolescen s. In e na ional jou nal
o en i onmen al esea ch and public heal h. 2013;10(10):5026–38.
56. San os NC, Cos a PS, Cunha P, Po ugal-Nunes C, Amo im L, Co e J, e al.
Clinical, physical and li es yle a iables and ela ionship wi h cogni ion and
mood in aging: a c oss-sec ional analysis o dis inc educa ional g oups.
F on ie s in aging neu oscience. 2014. p. 6.
57. Roskam A-JR, Kuns AE, Van Oyen H, Dema es S, Klumbiene J, Regido E,
e al. Compa a i e app aisal o educa ional inequali ies in o e weigh and
obesi y among adul s in 19 Eu opean coun ies. In J Epidemiol.
2010;39(2):392–404.
58. Gal DL, San os A-C, Ba os H. Leisu e- ime e sus ull-day ene gy
expendi u e: a c oss-sec ional s udy o seden a ism in a Po uguese u ban
popula ion. BMC Public Heal h. 2005;5(1):16.
59. Mo ei a PA, Pad ão PD. Educa ional and economic de e minan s o ood
in ake in Po uguese adul s: a c oss-sec ional su ey. BMC Public Heal h.
2004;4(1):58.
60. Oli ei a IT. Fecundidade e educação. Re is a de Es udos Demog á icos.
2007;40:5–19.
61. Da is EM, Zyzanski SJ, Olson CM, S ange KC, Ho wi z RI. Racial, e hnic, and
socioeconomic di e ences in he incidence o obesi y ela ed o childbi h.
Am J Public Heal h. 2009;99(2):294.
62. Kuns AE, Bos V, Lahelma E, Ba ley M, Lissau I, Regido E, e al. T ends in
socioeconomic inequali ies in sel -assessed heal h in 10 Eu opean coun ies.
In J Epidemiol. 2005;34(2):295–305.
63. Huisman M, Van Len he F, Mackenbach J. The p edic i e abili y o sel -
assessed heal h o mo ali y in di e en educa ional g oups. In J
Epidemiol. 2007;36(6):1207–13.
64. Oksuzyan A, Juel K, Vaupel JW, Ch is ensen K. Men: good heal h and high
mo ali y. Sex di e ences in heal h and aging. Aging Clin Exp Res.
2008;20(2):91–102.
65. McDonough P, Wal e s V. Gende and heal h: eassessing pa e ns and
explana ions. Soc Sci Med. 2001;52(4):547–59.
66. Macin y e S, Hun K, Swee ing H. Gende di e ences in heal h: a e hings
eally as simple as hey seem? Soc Sci Med. 1996;42(4):617–24.
67. Rieke PP, Bi d CE. Re hinking gende di e ences in heal h: why we need
o in eg a e social and biological pe spec i es. J Ge on ol Se B Psychol
Sci Soc Sci. 2005;60(Special Issue 2):S40–S7.
68. Eu os a . Mig an s in Eu ope. Belgium: A s a is ical po ai o he i s and
second gene a ion; 2011.
69. Eu opean Commission. People a isk o po e y o social exclusion.
h p://ec.eu opa.eu/eu os a /s a is ics-explained/index.php/People_a _ isk_
o _po e y_o _social_exclusion. Accessed 22 No embe 2015.
70. Kawachi I, Sub amanian S, Almeida-Filho N. A glossa y o heal h
inequali ies. J Epidemiol Communi y Heal h. 2002;56(9):647–52.
71. Eu os a . Old-age dependency a io pe 100 pe sons. h p://ec.eu opa.eu/
eu os a / gm/ able.
do? ab= able&ini =1&language=en&pcode= sdde510&plugin=1. Accessed 3
Feb 2015.
72. Di eção-Ge al da Saúde. A Saúde dos Po ugueses. Pe spe i a 2015 Lisbon:
Di eção-Ge al da Saúde. 2015.
73. Ma mo M. Heal h inequali ies in he EU. Final epo o a conso ium.
Heal h inequali ies in he EU Final epo o a conso ium. 2013.
74. Di ecção-Ge al da Saúde. Plano Nacional de Saúde: Re isão e Ex ensão a
2020. Lisbon: Di ecção-Ge al da Saúde; 2015.
75. Smi h KE, El anani MK. Wha kinds o policies o educe heal h inequali ies
in he UK do esea che s suppo ? Jou nal o Public Heal h. 2015;37(1):6–17.
76. B a eman P, G uskin S. De ining equi y in heal h. J Epidemiol Communi y
Heal h. 2003;57(4):254–8.
• We accep p e-submission inqui ies
• Ou selec o ool helps you o ind he mos ele an jou nal
• We p o ide ound he clock cus ome suppo
• Con enien online submission
• Tho ough pee e iew
• Inclusion in PubMed and all majo indexing se ices
• Maximum isibili y o you esea ch
Submi you manusc ip a
www.biomedcen al.com/submi
Submi you nex manusc ip o BioMed Cen al
and we will help you a e e y s ep:
Campos-Ma os e al. In e na ional Jou nal o Equi y in Heal h (2016) 15:26 Page 10 o 10

The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
95
2.4. Shi ing de e minan s o heal h inequali ies in uns able imes:
Po ugal as a case s udy
Re e ence:
Campos-Ma os I, Russo G, Gonçal es L. Shi ing de e minan s o heal h inequali ies
in uns able imes: Po ugal as a case s udy. Accep ed o publica ion in he Eu opean
Jou nal o Public Heal h in May 2017. DOI:
h ps://dx.doi.o g/10.1093/eu pub/ckx080
Online supplemen a y da a o his publica ion is in Appendix 4.
The ela ionship be ween con ex and heal h inequali ies Chap e 2. Resul s
96
Eu opean Jou nal o Public Heal h, 1–6
ßThe Au ho 2017. Published by Ox o d Uni e si y P ess on behal o he Eu opean Public Heal h Associa ion. All igh s ese ed.
doi:10.1093/eu pub/ckx080
.........................................................................................................
Shi ing de e minan s o heal h inequali ies in uns able
imes: Po ugal as a case s udy
Ine
ˆs Campos-Ma os
1,2
, Giuliano Russo
3
, Luzia Gonc¸al es
1,4
1 Ins i u o de Higiene e Medicina T opical, No a Uni e si y o Lisbon, Lisbon, Po ugal
2 Cen o de In es igac¸a
˜o em Sau
´de Pu
´blica, Lisbon, Po ugal
3 Depa men o P ima y Ca e and Public Heal h, Queen Ma y Uni e si y o London, London, UK
4 Cen o de Es a ı
´s ica e Aplicac¸o
˜es da Uni e sidade de Lisboa, Lisbon, Po ugal
Co espondence: Ine
ˆs Campos-Ma os, Depa amen o de Sau
´de In e nacional e Bioes a ı
´s ica, Ins i u o de Higiene e
Medicina T opical, Rua da Junquei a, n100, 1349-008 Lisboa, Po ugal, Tel: +44 750 282 3003,
e-mail: [email p o ec ed]
Backg ound: We explo e how heal h inequali ies (HI) changed in Po ugal o e he las decade, conside ing i is
one o he mos unequal Eu opean coun ies and has gone h ough majo economic changes. We desc ibe how
inequali ies in limi a ions changed conside ing di e en socioeconomic de e minan s, in o de o unde s and
wha d o e changes in HI. Me hods: We used c oss-sec ional wa es om he Eu opean Su ey on Income and
Li ing Condi ions da abase o de e mine how inequali ies in heal h limi a ions changed be ween 2004 and 2014 in
Po ugal in esiden s aged 16 yea s and o e . We calcula ed p e alence es ima es o limi a ions and di e ences
be ween income e ciles, he concen a ion index o each yea and i s decomposi ion and mul iple logis ic e-
g essions o es ima e he associa ion be ween socioeconomic de e minan s and limi a ions. Resul s: The
p e alence o heal h limi a ions inc eased in Po ugal since 2004, especially a e 2010, om 35 o 47%. Bu
he di e ence be ween op and bo om income e ciles dec eased om 23 o 10 pe cen age poin s, as iche
people expe ienced a s eepe inc ease. This was d i en by an inc ease in p e alence among economically ac i e
people, who, om 2011 onwa ds, had mo e limi a ions (OR and 95% CI we e 2.42 [2.13–2.75] in 2004 and 0.71
[0.65–0.78] in 2014). Conclusion: These esul s sugges wo sening heal h in Po ugal in he las decade, possibly
connec ed o pe iods o economic ins abili y. Howe e , absolu e HI dec eased conside ably in he same pe iod. We
discuss he possible ole o di e se adap a ion capaci y o socioeconomic g oups, and o high emig a ion a es o
young, heal hie people, e lec ing ano he side o he ‘mig an heal h e ec ’.
.........................................................................................................
In oduc ion
Socioeconomic heal h inequali ies (HI) a e ubiqui ous. They ha e
been obse ed wo ldwide as long as da a ha e been a ailable. I
appea s ha , ega dless o place and ime, heal h ends o ollow he
pa e ning o socioeconomic di e ences (1).
Va ious socioeconomic indica o s—educa ion, inancial
esou ces, employmen o occupa ion—de e mine HI, ope a ing
h ough di e en pa hways. Educa ion leads o be e in o ma ion,
cogni i e abili ies and de e mines p e e ences (2); inancial
esou ces, such as income o weal h, allow indi iduals o access
heal h-p oducing esou ces, such as heal hca e o housing.
Employmen no only p o ides income, bu also a sense o con ol
o e one’s li e, lack o which is s ongly associa ed wi h impo an
s ess eac ions, which can de e io a e heal h (3). People wi h highe
occupa ional g ades also end o ha e a s onge sense o con ol
o e hei heal h, hei jobs and hei li es (4), bu occupa ion can
also e lec an indi idual’s place in socie y, showing he e ec o ank
and subjec i e eelings owa ds one’s posi ion in socie y (2). The
simul aneous analysis o a ious socioeconomic de e minan s o
HI can p o ide clues as o which p ocesses a e mo e impo an in
he c ea ion o HI (5). A be e unde s anding o which p ocesses
shape HI will help o build a base o design policies ha ackle hem
e ec i ely.
Po ugal is a pa icula ly in e es ing case s udy o HI. The
coun y has had low economic g ow h (6), and despi e subs an ial
in es men s in social p o ec ion, educa ion and heal hca e (7,8),
emains one o he mos unequal Eu opean Union coun ies in
income dis ibu ion (9). This is e lec ed in heal h dis ibu ion:
se e al analyses ound Po ugal o ha e some o he highes HI
among Eu opean coun ies (10–12). Addi ionally, Po ugal has
gone h ough a pe iod o economic c isis and implemen a ion o
aus e i y measu es in he las yea s, ha ha e led o a spike in
emig a ion (13) and a de e io a ion o public social se ices (14).
A ecen e iew o he impac o economic c ises ound ha hey
ended o agg a a e HI in a a ie y o coun ies (15). Howe e , he
e iew no ed ha esul s we e a iable, pe haps due o di e ing
wel a e policies, o he di e si y o heal h and socioeconomic
a iables. Poo unde s anding o how economic c ises shape HI
hinde s he in e p e a ion o hese esul s.
This wo k aims o suppo policy choices ha a emp o mi iga e
he e ec o economic c ises o o he con ex ual changes on HI. To
do his, we desc ibe how HI changed in Po ugal o e he las
decade, in ligh o he impo an social and mac oeconomic
changes ha he coun y has been h ough, and how he
socioeconomic de e minan s o hese inequali ies changed. We
used da a om he c oss-sec ional wa es o he Eu opean Su ey
on Income and Li ing Condi ions (EU-SILC), om 2004 o 2014.
Po ugal is used as a case s udy, bu his analysis is applicable o
o he coun ies as i desc ibes how de e minan s o HI can be shaped
by con ex ual ans o ma ions. This is pa icula ly use ul conside ing
ha many coun ies ha e ecen ly gone h ough simila mac oeco-
nomic changes as Po ugal.
Me hods
This analysis was pe o med using da a om he Po uguese c oss-
sec ional wa es o EU-SILC be ween 2004 and 2014 (p o ided by
Eu os a in Decembe 2015). EU-SILC is an annual su ey ca ied
ou in se e al Eu opean coun ies wi h a mixed longi udinal and
c oss-sec ional design. Despi e his mixed design, c oss-sec ional
samples a e ep esen a i e o he a ge popula ion when app op ia e
weigh s a e used (16). Po ugal pa icipa es since 2004 using a
s a i ied, mul i-s age, household-based sample. The su ey collec s
da a on li ing condi ions and includes h ee heal h ela ed ques ions:
limi a ions in daily ac i i ies due o heal h p oblems, sel - epo ed
heal h (SRH) and ch onic condi ions.
We used ‘limi a ions’ as ou heal h ou come. Indi iduals we e
asked i hey we e limi ed in ac i i ies hey usually did because o
heal h p oblems. Possible answe s included ‘Yes, s ongly limi ed’,
‘Yes, limi ed’ o ‘No’. The i s wo op ions we e collapsed, c ea ing a
bina y a iable (1 = ‘wi h limi a ions’, 0 = ‘wi hou limi a ions’).
This heal h ou come was chosen as i p o ides an objec i e
measu e han SRH and should cap u e heal h s a us mo e
accu a ely (17). The ini ial desc ip i e analysis was also done o
he o he wo heal h a iables: SRH and ch onic condi ions. SRH
is a widely used su ey measu e in which esponden s a e hei
o e all heal h; we used SRH as a bina y a iable in which ‘bad’
and ‘ e y ba ’ heal h we e he ou come. ‘Ch onic condi ions’ is a
sel -assessed ques ion in which esponden s a e asked whe he hey
ha e a ch onic condi ion; his was also used as a bina y a iable, in
which ha ing a ch onic condi ion was he ou come.
The ollowing a iables we e included in he analysis:
Age a in e iew (in yea s).
Sex (male o emale).
Income: yea ly household equi alised disposable income, in
eu os, de la ed using he ha monised index o consume p ices
(18).
Educa ion: de ined by highes In e na ional S anda d
Classi ica ion o Educa ion (ISCED) le el a ained (19),
ca ego ised in o ‘p ima y o less’ o ‘mo e han p ima y’.
Occupa ion: based on he In e na ional S anda d Classi ica ion
o Occupa ions (ISCO) used in EU-SILC, occupa ions we e
ca ego ised in whi e o blue colla , ollowing p e ious wo k
(ISCO codes 1–5 we e whi e colla , 6–9 blue colla and a med
o ces we e excluded) (20).
Ac i i y: based on he EU-SILC a iable ‘sel -de ined cu en
economic s a us’, people we e ca ego ised as ‘ac i e’ i hey
de ined hemsel es as being employed (pa o ull ime), in
aining o s udying, o ul illing domes ic asks; and ‘inac i e’
i hey we e unemployed, e i ed, un i o wo k o in he ‘o he
inac i e’ ca ego y.
Sa ings: EU-SILC u he asks households abou hei capaci y o
ace unexpec ed inancial expenses and o a o d one-week
annual holiday away om home. These wo a iables we e
me ged and ans o med in o a bina y a iable so ha he
alue ‘0’ was a ibu ed o households who could a o d bo h
and ‘1’ o he emaining households.
We used he comple e sample o esiden s aged 16 and o e . The
p opo ion o indi iduals who had limi a ions was calcula ed o
each yea in he o e all sample, wi hin each income e cile, and
s a i ied by age g oups. Income e ciles we e calcula ed acco ding
o he dis ibu ion o income o each yea .
The concen a ion index (CIx) o income- ela ed inequali ies in
limi a ions was calcula ed o each yea . The CIx is a measu e o
inequali ies based on he heal h concen a ion cu e. This cu e is
he esul o plo ing o he cumula i e pe cen age o indi iduals,
anked by income, wi h he cumula i e pe cen age o limi a ions. In
his plo , pe ec equali y is ep esen ed by a diagonal line, showing
an equal dis ibu ion o limi a ions among he popula ion,
ega dless o income. The CIx is calcula ed as wice he a ea
be ween he concen a ion cu e and he line o pe ec equali y.
When he e is pe ec equali y, he CIx is ze o. By con en ion, i
all limi a ions a e concen a ed in he iches (poo es ) pe son, he
CIx is 1 (-1). Howe e , wi h dicho omous ou come a iables, he
CIx is no wi hin he [-1,1] ange and be ween-yea compa abili y
may be limi ed; ollowing Wags a (21), o minimise his limi a ion,
we no malised he CIx by di iding i by 1 minus he p opo ion o
esponden s epo ing limi a ions in each yea .
Wags a e al. (22) showed ha he CIx can be decomposed in o
con ibu ions o indi idual ac o s o he income- ela ed HI. This
analysis allows o he quan i ica ion o how each ac o (i.e. each
socioeconomic a iable) con ibu es o he o e all dis ibu ion o
he heal h ou come among income anks. The con ibu ion o
each ac o is he p oduc o he elas ici y o ha ac o wi h
espec o he heal h a iable (i.e. he p opo ional change o a
speci ic ac o in ela ion o a p opo ional change in he heal h
a iable) and he CIx o ha ac o (i.e. he deg ee o income-
ela ed inequali y o ha ac o ).
Finally, we pe o med a mul iple logis ic eg ession o each yea ,
using he dicho omous heal h a iable (limi a ions) as an ou come.
We included all he demog aphic and socioeconomic a iables lis ed
abo e as explana o y a iables: age, sex, income, educa ion,
occupa ion, ac i i y and sa ings. These we e all added o he
model simul aneously.
Analyses we e weighed by a pe sonal c oss-sec ional weigh
p o ided by he EU-SILC da abase, which con ols o geog aphical,
household size, gende , and age g oup dis ibu ion, and non-
esponse wi hin each household. Analyses we e done on SPSS
S a is ics 21 and in ADePT So wa e 6.0 using a non-linea
model o he CIx.
Resul s
Table 1 summa ises he sample cha ac e is ics. Yea ly sample size
anged om 9947 indi iduals in 2007 o 14 650 in 2014. A e age age
inc eased om 46.3 o 49.0-yea s-old om 2004 o 2014. The
p opo ion o indi iduals wi h limi a ions also inc eased om 35.2
o 47.3%. Median income inc eased be ween 2004 and 2012, om
5869 o 8366 eu os pe yea , and d opped o 8265 eu os in 2014.
The e was also an inc ease in he p opo ion o people wi h
seconda y and e ia y educa ion and in whi e-colla occupa ions,
bo h ep esen ing app oxima ely hal he sample in 2014. The
p opo ion o ac i e people dec eased om 70% in 2004 o 58.5%
in 2014.
Figu e 1 shows he p opo ion o indi iduals wi h limi a ions by
yea . This p opo ion was s able a a ound 30% un il 2011, when i
inc eased o 43%, and hen inc eased again in 2014 o 47%. These
changes occu ed in all income e ciles, bu a ew di e ences we e
no iceable: (i) in almos e e y yea , he p opo ion o people wi h
limi a ions was highe in he i s e cile ( he lowes -income
popula ion g oup), ollowed by he second, and lowes in he
hi d; (ii) his di e ence was s able un il 2011, when he
p opo ion o limi a ions inc eased in all e ciles, mos ma kedly
in he second and hi d; (iii) his led o a dec ease in he absolu e
di e ence in limi a ions inequali ies be ween he i s and hi d
income e ciles. Figu e 1 poin s ou he absolu e di e ences
be ween he i s and hi d income e ciles in ou yea s (23% age
poin s in 2004 and 2010, 16 in 2011, and 10 in 2014). When
s a i ied by age g oups, he analysis p esen ed in igu e 1 shows
ha inequali ies in limi a ions we e highes in he olde age
g oups, he inc ease in limi a ions in 2011 occu ed in younge
age g oups, and he oldes age g oups showed a dec ease in limi a-
ions in 2012 (Supplemen a y igu e S1A).
The CIx was nega i e e e y yea , as he p e alence o limi a ions
was highe in poo e people ( igu e 2). The CIx anged be ween 0.15
and 0.18 (in absolu e alues) un il 2010 and d opped in 2011 o 0.09
and o 0.05 in 2014. Un il 2010, e e y socioeconomic a iable had a
nega i e con ibu ion o he CIx, meaning ha hey all con ibu ed
o p o-poo inequali y in he dis ibu ion o limi a ions. Howe e ,
a e 2010 he e we e a ew no iceable changes.
Fi s , ac i i y now ga e a posi i e con ibu ion o he CIx.
De ailed analysis o he con ibu ion o each a iable
(Supplemen a y able S1A) showed ha he elas ici y o limi a ions
wi h espec o ac i i y changed in 2011, om posi i e o a nega i e
con ibu ion; he CIx o ac i i y, on he o he hand, emained s able.
This means ha , in all yea s, inac i e people had lowe incomes
when compa ed wi h ac i e people. Howe e , while limi a ions
we e mo e p e alen in inac i e people un il 2010, hey we e mo e
p e alen in ac i e people a e his yea .
2o 6 Eu opean Jou nal o Public Heal h
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
105
We used da a om six ounds o he ESS, be ween 2002 and 2012, which included 36
coun ies and 237,535 indi iduals. Da a was o ganized in h ee le els – indi iduals,
yea s and coun ies – and analysed using mul ile el s a is ical echniques. Coun ies
we e g ouped acco ding o hei wel a e egime ype and analyses we e done
sepa a ely o each g oup. Each indi idual was a ibu ed one o h ee social mobili y
pa hs: upwa d, s able, o downwa d, acco ding o he di e ences be ween hei own
and hei mo he and a he ’s educa ional achie emen (analyses we e done sepa a ely
o mo he and a he ). Heal h was ope a ionalized as a bina y a iable: 1 o ‘bad’ o
‘ e y bad’ SAH and 0 o ‘ ai ’, ‘good’, and ‘ e y good’. In o de o ensu e ha he
social mo emen i sel was analysed and he e ec s o childhood ci cums ances we e
excluded, analyses we e con olled o pa en al educa ion. We calcula ed ela i e
(OR) and absolu e ( isk di e ence) measu es o he associa ion be ween social
mobili y and SAH o each wel a e egime ype.
Resul s showed ha upwa d mobili y (when compa ed o being socially ‘s able’) was
posi i ely associa ed wi h be e heal h in all wel a e egime ypes, measu ed bo h
om mo he and a he ’s achie emen , using absolu e and ela i e measu es. On a
ela i e scale, hese esul s we e s a is ically signi ican o p<0.05. On an absolu e
scale, o me USSR coun ies showed he bigges and only signi ican di e ence o
upwa d mo emen (4.1 and 3.8% di e ence, when social mobili y was measu ed
om he mo he ’s o he a he ’s achie emen , espec i ely). Scandina ian coun ies
showed he smalles and no signi ican di e ences: 0.8 and 0.9% di e ence, om
mo he o a he ’s, espec i ely.
O e all, his analysis showed ha social mobili y was associa ed wi h di e ences in
heal h in all wel a e egime ypes, bu Scandina ian coun ies showed he smalles
associa ion. Despi e ha ing high le els o social mobili y, hese coun ies seemed o
e icien ly sepa a e i om heal h, mo e so han coun ies om o he wel a e egime
ypes. These esul s sugges ha he ‘pa adox’ o high HI in no he n Eu opean
coun ies is unlikely o be due o social mobili y di e ences among wel a e egime
ypes. This analysis did no es o he hypo heses on why HI emain high in no he n
Eu opean coun ies, bu showed ha wel a e egimes play a ole in de e mining bo h
le els o popula ion heal h and heal h dis ibu ion in Eu opean coun ies.

The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
106
3.1.3. Social De e minan s o Heal h in Po ugal
As one o he mos unequal coun ies in Eu ope, bo h in e ms o income dis ibu ion
(13) and heal h dis ibu ion (14), Po ugal p esen s an in e es ing case s udy o
unde s anding HI. Despi e he high le els o inequali y, WHO iden i ied HI as an
‘impo an policy gap’ in he Po uguese Na ional Heal h Plan (15), which e lec s he
low poli ical in e es in he subjec and he absence o a na ional s a egy in place o
ackle hese inequali ies. I was hus impo an o sys ema ize cu en knowledge on
socioeconomic HI in Po ugal, o se a s epping-s one owa ds a possible s a egy o
ackle HI in he coun y. Fo his, we ca ied ou a SR o he li e a u e ha ga he ed
he exis ing e idence abou socioeconomic HI in Po ugal.
The PRISMA s a emen was used o guide and epo he e iew (16). The e iew
began by de ining wha measu es o SES would be included. Fo his, he
PROGRESS2 amewo k was ollowed, s anding o Place o esidence, Race /
e hnici y / cul u e / language, Occupa ion, Gende /sex, Religion, Educa ion,
Socioeconomic s a us and Social capi al (17). Bo h indi idual and con ex ual
de e minan s we e included. Heal hca e u iliza ion o access, and heal h ela ed
beha iou s we e excluded om he analysis, as we we e in e es ed solely on
inequali ies in heal h ou comes. We included e e y s udy ha quan i ied an
associa ion be ween he socioeconomic and heal h a iables, and con olled o , a
leas , gende and age. S udies ha used da a om 2000 onwa ds and om he
Po uguese esiden popula ion ( ega dless o na ionali y) we e included. We
excluded quali a i e s udies. A icles w i en in Po uguese and English we e
included. We sea ched Scopus, Web o Science and Pubmed o pape s ha me he
eligibili y c i e ia. Addi ionally, we scoped publica ions o esea che s in Po ugal
who egula ly publish esea ch in his a ea o publica ions ha me he eligibili y
c i e ia, o complemen ou online sea ch. Da a was ex ac ed om he selec ed
a icles and a b ie quali y e alua ion was pe o med. Resul s we e p esen ed using a
na a i e desc ip ion and a diag am ha summa ized he indings.
The inal selec ion included se en y-one pape s, all o which epo ed obse a ional
s udies, and mos o which used c oss-sec ional da a. Mos publica ions epo ed
2 The PROGRESS amewo k was c ea ed as an aide-memoi o help esea che s apply an equi y lens
o hei esea ch, and public heal h p o essionals o conside all po en ially inequi able ci cums ances in
public heal h in e en ions.
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
107
signi ican HI ha a ou ed indi iduals o g oups wi h be e social s anding. Some
associa ions we e pa icula ly common and showed s ong e idence o HI: lowe
educa ion wi h obesi y and wi h subjec i e ill heal h; and emale gende wi h men al
heal h symp oms and subjec i e ill heal h.
O he se en y-one selec ed pape s in he SR, se en een analysed, in one way o
ano he , he e ec o a con ex ual a iable. Se en o hese s udies had an ecological
design (18-24). Fou in es iga ed how an indi idual’s pe cep ions o hei
neighbou hood we e associa ed o hei heal h (25-28). Fou o he s udies explo ed
he di e ence be ween esiden s o u ban and u al se ings (29-32). The wo
emaining s udies looked a di e ences be ween dep i ed and a luen
neighbou hoods (33) and municipali ies (34). O e all, hese we e e y he e ogeneous
s udies, looking a di e en opics, di e en popula ions, and wi h di e ing deg ees
o quali y.
This e iew o e ed a sys ema iza ion o cu en e idence on HI in Po ugal, which
had no been done be o e. I s esul s show ha , while he e a e a conside able numbe
o publica ions ouching on he subjec , hey end o ocus on speci ic opics,
ep oducing simila esul s; hey do no necessa ily ocus on wha a e he mos
p e alen heal h and social issues in Po ugal; and hey end o use simila
me hodologies. This analysis was also able o iden i y wha a e he mos impo an HI
in Po ugal o which e idence is a ailable – educa ion and gende inequali ies in
obesi y, SAH and men al heal h – and he majo gaps in he esea ch li e a u e
ega ding his opic – analyses ocused on he mos p e alen heal h issues, he mos
impo an SES ac o s, and he mos ulne able popula ions. I is clea om hese
esul s ha , a he p esen momen , he e is no na ionally o ien ed esea ch s a egy
which would be c ucial o guide esea ch in a coun y whe e HI emain pa icula ly
high.
3.1.4. Shi ing De e minan s o Heal h Inequali ies in Po ugal
Many indi idual socioeconomic a iables can de e mine HI. Educa ion, one o he
mos commonly used indica o s, can de e mine be e cogni i e abili ies and be e
knowledge, ha allow indi iduals o p e en illness, ha e be e heal h, and be e
manage disease (35). Financial esou ces can allow indi iduals o acqui e heal h-
p oducing esou ces (36). Employmen p o ides income, bu also a sense o pu pose,
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
108
a s uc u e o he day, and social connec ions, which a e all associa ed wi h be e
heal h (37). Occupa ion, on he o he hand, is a e lec ion o social ank, and has also
been s ongly connec ed o mul iple heal h ou comes; amously, he Whi ehall s udies
showed ha hese di e ences we e a consequence o eelings o con ol o e one’s
li e, which a e s onge among people in highe occupa ional anks (38). This a ie y
o SES indica o s is a e lec ion o he mul i ude o pa hways ha lead o HI.
Obse ing which indica o s a e mo e impo an , and how hese change o e ime, can
gi e impo an clues o which p ocesses a e he mos ele an in c ea ing HI. Po ugal
has been h ough impo an changes o e he las decade, wi h pe iods o poli ical
ins abili y and implemen a ion o aus e i y measu es (39). These ans o ma ions
p o ide a unique oppo uni y o obse e how con ex ual changes can lead o changes
in o e all heal h, heal h dis ibu ion, and o he de e minan s o HI. Wi h he ou h
publica ion, we aimed o desc ibe how he de e minan s o HI changed in Po ugal
o e he las decade, in o de o unde s and wha p ocesses c ea ed HI and how hese
changed o e ime. This unde s anding can hope ully lead o in o med policies ha
can success ully ackle HI.
The ou h and las publica ion o his disse a ion used da a om he c oss-sec ional
wa es o he Po uguese sample o EU-SILC om 2004 o 2014 was used, wi h
limi a ions in daily ac i i ies as he ou come a iable. We calcula ed he p e alence o
limi a ions in each income e cile in each yea and absolu e di e ences be ween i s
and hi d e ciles o selec ed yea s. We hen calcula ed he CIx o each yea and i s
decomposi ion in a ious SES indica o s: occupa ion, employmen , educa ion,
income, sa ings, age, and sex. Finally, we an a mul iple logis ic eg ession analysis
o each yea , o de e mine he OR o each o hese indica o s.
The p e alence o limi a ions was ound o ha e inc eased in Po ugal in he las
decade, especially a e 2010. Howe e , his inc ease was s eepe in iche e ciles,
which led o a dec ease in bo h absolu e and ela i e inequali ies in limi a ions.
Analysis o he CIx decomposi ion and o he OR showed ha p o essional ac i i y
was he main de e minan o he dec ease in inequali y – ac i e people had ewe
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
109
limi a ions han inac i e people be o e 2011, bu his pa e n e e sed om his yea ,
as ac i e people had mo e limi a ions3.
This analysis was he i s ha looked a he change o HI in Po ugal o e ime.
Despi e ocusing on one coun y alone, se e al lessons can be applied o o he
con ex s o esea ch and policy pu poses. Fi s ly, HI changed conside ably in
Po ugal o e he las decade, possibly ela ed o changes in he coun y’s social and
economic ci cums ances. In 2011 Po ugal wen h ough a pe iod o conside able
unce ain y wi h he eques o a €78 billion bailou om he EU and he In e na ional
Mone a y Fund, esigna ion o he uling go e nmen and snap elec ions, and cons an
alks o aus e i y measu es ha would o ce he Po uguese o ‘ igh en hei bel ’ in
he nea u u e (40). This migh ha e in luenced o e all heal h and changes in i s
dis ibu ion, which highligh s he impo ance o con ex ual de e minan s on HI and
adds knowledge o he g owing body o e idence o he impac o economic c ises on
HI (41). Secondly, i shows ha HI, e en wi hin one coun y, a e nei he s a ic no
de e mined by he same ac o s o e ime. On he con a y, con ex ual socioeconomic
changes can ha e subs an ial impac s on HI and on wha d i es hem. Finally, his
unde s anding o he d i e s o HI can be ex emely use ul o ou line policies o ackle
hem. In ou analysis, he main d i e o dec easing HI was an in e sion in he
p e alence o limi a ions among ac i e and inac i e people. We hypo hesized his
may be a consequence o high emig a ion a es in he coun y (which led o an exi o
heal hie , iche people) o o di e en adap a ion capaci y among socioeconomic
g oups (as g oups o highe social s anding may be less capable o adap ing o wo se
socioeconomic ci cums ances).
3.1.5. Summa y
These ou publica ions pain a pic u e o how con ex ual de e minan s in e ac wi h
indi idual cha ac e is ics o in luence he dis ibu ion o heal h in Po ugal and in i s
wide Eu opean con ex . The analyses and hei esul s a e summa ized in able 2.
Con ex ual social capi al was ound o ha e no impac on popula ion heal h in
Eu opean coun ies, bu had an e ec o a pa icula g oup o people – hose wi h
low indi idual le els o in e pe sonal us . Wel a e egime ypes we e also associa ed
3 People we e conside ed ‘ac i e’ i hey we e employed (pa o ull ime), in aining o s udying, o
ul illing domes ic asks; ‘inac i e’ people we e unemployed, e i ed, un i o wo k o in he ‘o he
inac i e’ ca ego y.
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
110
wi h he magni ude o he heal h impac o social mobili y, wi h Scandina ian
coun ies showing he smalles absolu e di e ences. Finally, impo an con ex ual
changes in Po ugal o e he las decade seem o ha e in luenced heal h and i s
dis ibu ion in he coun y.
F om a di e en pe spec i e, he SR sough o summa ize he knowledge on HI in
Po ugal up o da e. This analysis showed ha he s udy o con ex ual de e minan s o
HI is s ill uncommon in Po ugal and ocused on a limi ed numbe o de e minan s.
Resul s o his pape mus be in e p e ed wi h cau ion as hey agg ega e all he da a on
HI in Po ugal o e he las decade; as he ou h publica ion showed, his was a ime
o in ense changes in he dis ibu ion o heal h in Po ugal, so agg ega ing he esul s
o e his pe iod o ime may ha e hidden impo an in o ma ion.
Table 2. Summa y o disse a ion publica ions, de e minan s es ed, geog aphic con ex , ime pe iod,
and main indings.
Publica ion
Social Capi al and
Heal h in Eu opean
Coun ies
Social Mobili y and
Heal h in Eu opean
Wel a e Regimes
Social
De e minan s o
Heal h in Po ugal
Shi ing
De e minan s o
Heal h Inequali ies
in Po ugal
Con ex ual
de e minan
Social capi al
Wel a e egime
n/a
Mac oeconomic
con ex
Indi idual
de e minan
Social capi al
Social mobili y
n/a
Socioeconomic
de e minan s (a)
Con ex
Eu ope
Eu ope
Po ugal
Po ugal
Time pe iod
2002-2012
2002-2012
2000-2014
2004-2014
Main
indings
Impo an c oss-
sec ional
in e ac ion – low
us indi iduals
ha e wo se heal h
in high us
con ex s
Scandina ian
coun ies had he
smalles associa ion
be ween social
mobili y and heal h
Few s udies on
con ex ual
de e minan s o
heal h.
HI dec eased in
Po ugal a e 2010,
mos ly due o
wo se heal h
among
p o essionally
ac i e people
(a) Fi e indi idual de e minan s we e es ed: occupa ion, ac i i y, income, educa ion, and sa ings.
HI: Heal h Inequali ies

The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
111
3.2. Limi a ions
The publica ions included in his analysis a e subjec o a numbe o limi a ions.
Limi a ions ha a e speci ic o each publica ion ha e been spelled ou in he esul s
sec ion. Some limi a ions a e common o wo o mo e publica ions.
The i s is ha he da a ha ha was used was c oss-sec ional. EU-SILC has a
longi udinal componen ha could ha e been used, bu panels only las ou yea s, and
his was no conside ed a su icien amoun o ime o assess changes in heal h
ou comes. As such, i was decided o use only c oss-sec ional da a. This decision
limi ed he s udy’s abili y o de e mine whe he heal h is an ou come o he
de e minan unde s udy, o he o he way a ound. Howe e , i was ied, when
app op ia e, o conside bo h di ec ions o he associa ion. Fo example, by
acknowledging bo h p ocesses o heal h selec ion and o he impac o social mobili y
in heal h in he second publica ion.
Ano he possible sou ce o limi a ions is om he used heal h a iables. The use o
SAH and o heal h limi a ions is open o c i icism, bu i mus also be conside ed ha
hese ha e impo an alue, and he e is a eason why mos popula ion su eys ask
hese ques ions speci ically. SAH has been shown o be a eliable measu e o o e all
heal h (42), and e en a good p edic o o mo ali y (43). A guably, SAH is a be e
measu e o heal h han an ‘objec i e’ one, such as a diagnosed illness, as i
inco po a es he indi idual’s pe cep ion o hei own heal h (42). Sel - epo ed
limi a ions in daily ac i i ies a e also a e lec ion o he indi idual’s pe cep ions, bu
p o ide a mo e objec i e measu e han SAH. This ou come has been used by o he
au ho s as an objec i e measu e o unc ional limi a ions (44, 45) and Eu os a uses i
as measu e o disabili y (46).
The wo i s publica ions mus also be in e p e ed wi h cau ion as hey consis o
c oss-coun y compa isons o subjec i e measu es, which may be in e p e ed
di e en ly in each coun y. This is pa icula ly impo an o he i s publica ion, as
in e pe sonal us showed conside able a ia ion be ween coun ies. We ied o
add ess his issue by using coun ies as ixed e ec s, hus emo ing om he model
ime-in a ian coun y cha ac e is ics.
As a whole, he ou publica ions may also occasionally seem o bea only a loose
connec ion, as hey ocus on di e en geog aphical a eas, on di e en de e minan s o
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
112
heal h, and a e some imes amed in ways ha do no seem o be connec ed. Indeed,
he SR o he li e a u e was sligh ly di e en in i s goals, bu was conside ed a
necessa y i s s ep, in o de o agg ega e all a ailable knowledge abou HI in
Po ugal be o e ca ying on u he esea ch. Excluding he SR, all he publica ions
look a con ex ual cha ac e is ics – be i wel a e egime, social capi al, o an economic
c isis – and how hese ha e an impac on he dis ibu ion o heal h wi hin pa icula
popula ion g oups.
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
113
3.3. The E ec o Con ex on Heal h and on Heal h Dis ibu ion
As was ou lined in he in oduc ion, he s udy o con ex and how i in luences heal h
and HI has been unde mined by a lack o a heo e ical basis on how con ex ual
de e minan s wo k. Bo h he CSDH and Dide ichsen ha e ecognised he ole o
con ex in c ea ing HI, bu only inso a as i is esponsible o social s a i ica ion and
o he p oduc ion o policies (47, 48). While help ul o policy o mula ion, his is an
incomple e iew o how con ex can in luence heal h and HI.
The amewo k p oposed he e seeks o ill his gap by ou lining he mechanisms by
which con ex in luences heal h and heal h dis ibu ion. I does his by desc ibing wo
mechanisms: (i) changes in o e all hea h and (ii) changes in heal h dis ibu ion. This
dis inc ion is impo an because, jus as is a gued o he CSDH’s amewo k (48),
policies aimed a imp o ing popula ion heal h do no always ha e a posi i e impac
on i s dis ibu ion. Hence, when seeking o in luence HI, i is no enough o
implemen a policy o imp o e o e all heal h. Ra he , i is impo an o conside how
ha policy can change heal h dis ibu ion also. I is hoped ha he amewo k
p oposed he e con ibu es o he absence o his heo e ical ounda ion, by ou lining
he mechanisms by which con ex in luences heal h and heal h dis ibu ion, p o iding
a basis o policy choices and empi ical analyses.
This amewo k was subs an ially inspi ed by Dide ichsen’s amewo k, d awing on
he indi idual pa hways ha he au ho ou lines o o m a basis o how con ex can
in luence hose pa hways. Unlike ha amewo k, howe e , his one ocuses on he
ole o con ex , and highligh s i s impac on popula ion heal h, on he one hand, and
on heal h dis ibu ion, on he o he . These impac s a e no mu ually exclusi e, as any
one change in con ex o con ex ual cha ac e is ic can in luence heal h and heal h
dis ibu ion h ough mul iple pa hways; howe e , i is hoped his dis inc ion suppo s
a e lec ion on how con ex ope a es and helps ill he ‘black box’ o con ex ual
e ec s on heal h (49).
3.3.1. Fi s Mechanism: Changes in O e all Heal h
The i s mechanism leads o changes in he o e all heal h o a popula ion. Using he
indi idual pa hway o Dide ichsen’s amewo k as a base, his mechanism can
ope a e by changing SES, exposu e o isk ac o s, o heal h s a us o indi iduals.
Figu e 6 shows hese h ee e ec s. The co e o he amewo k s a s om social
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
114
posi ion, which de e mines di e en ial exposu es; exposu es lead o di e en ial
ulne abili y; and inally disease o illness lead o di e en ial consequences, which
can ha e an impac back on social posi ion.
Figu e 6. Fi s mechanism: changes in o e all heal h.
The i s e ec e e s o how indi iduals’ social posi ion can be al e ed by con ex .
Fo example, when a coun y’s inances imp o e, po e y ends o decline, and
indi idual economic s a us o i s ci izens ends o imp o e, leading o o e all
imp o emen s in heal h. The second e ec e lec s changes in he exposu e o isk
ac o s; examples o his a e wa e luo ida ion o educ ions in ai pollu ion, which
can educe he exposu e o isk ac o s in an en i e popula ion, hus imp o ing i s
o e all heal h. The hi d e ec e e s o e en s ha change heal h s a us, such as when
an inno a i e ea men o a p e alen disease is disco e ed and made a ailable. All
hese e ec s ha e he po en ial o imp o e o e all popula ion heal h, ei he di ec ly,
by a ec ing indi iduals’ heal h o indi ec ly, by a ec ing social posi ion o exposu e
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
121
p og ams (con ex ual cha ac e is ic) ha e he po en ial o educe his inequali y, as
hey p o ide easie access o sc eening o people who o he wise migh no use i (69).
3.3.4 Summa y
The amewo k p oposed he e ou lines how con ex in luences heal h and heal h
dis ibu ion. I is p oposed ha con ex ope a es h ough wo mechanisms, one ha
changes o e all popula ion heal h ( hough e ec s on social posi ion, exposu e o isk
ac o s, and disease o illness) and ano he ha changes heal h dis ibu ion ( h ough
e ec s on social s a i ica ion, di e en ial exposu e, di e en ial ulne abili y, and
di e en ial consequences).
This amewo k seeks o ill a gap in he esea ch li e a u e, by which he pa hways
be ween he con ex and heal h and HI ha e no been ou lined be o e. I akes a s ep
o wa d om o he concep ual amewo ks, as i acknowledge he ole o con ex on
se e al di e en s eps be ween social posi ion and heal h ou comes. In his
amewo k, con ex does mo e han jus s a i y indi iduals o hei social posi ion.
The amewo k will hope ully be used as a basis o u u e policy and empi ical
analyses, helping cla i y he mechanisms by which con ex in luences heal h and i s
dis ibu ion.

The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
122
3.4. Applica ion o he Concep ual F amewo k
The amewo k desc ibed in he p e ious sec ion ( igu e 8) ou lines he mechanisms
by which con ex ual de e minan s can in luence heal h and heal h dis ibu ion. This
amewo k seeks o summa ize how his in luence ope a es, and can po en ially be
used as a basis o policy and empi ical analyses. In his chap e , he amewo k is
applied o h ee esea ch pape s om his disse a ion, in o de o illus a e how i can
be used in he in e p e a ion o e idence. I is no applied o one publica ion – he SR
– as his did no look a how one con ex ual de e minan in e ac s wi h indi idual
cha ac e is ics bu a he se ou he backg ound o he las piece o in es iga ion on
Po ugal (see able 2).
3.4.1. Social Capi al and Heal h in Eu opean Coun ies
This analysis showed a complex e ec o na ional-le el social capi al on indi idual
heal h, making i an in e es ing case s udy o he applica ion o he amewo k. The
amewo k ocuses on how he con ex ual de e minan (con ex ual social capi al) can
ha e an impac on popula ion heal h and heal h dis ibu ion, ope a ing h ough an
indi idual cha ac e is ic (indi idual social capi al). The key esul s we e:
• Con ex ual social capi al was no associa ed wi h indi idual heal h, and
• High con ex ual social capi al was associa ed wi h wo se heal h in low us
indi iduals and be e heal h in high us indi iduals.
These esul s show ha con ex ual social capi al had no impac on o e all popula ion
heal h. Wi hin he concep ual amewo k ( igu e 8), his means ha social capi al will
no ope a e h ough he ull a ows ( he i s mechanism – changes in o e all
popula ion heal h).
On he o he hand, con ex ual social capi al had a di e en ial e ec on di e en social
g oups, leading o changes in heal h dis ibu ion (dashed a ows, second mechanism –
changes in heal h dis ibu ion). This is p obably a e lec ion o how con ex ual social
capi al is no a esou ce enjoyed equally by all indi iduals – when high us
indi iduals a e he majo i y, he social capi al hey p oduce be ween hem is no
sha ed wi h he mino i y, low us indi iduals. Mo eo e , no only is his esou ce ou
o hei each, bu disc imina ion om he majo i y and dissemina ion o ‘bad social
capi al’ (such as ein o cing social no ms ha a e ha m ul o heal h) migh u he
damage a g oup ha is al eady ulne able. This di e en ial e ec can hus be a
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
123
consequence o p ocesses o ne wo k closu e, disc imina ion, and dissemina ion o
‘bad social capi al’. All hese p ocesses in essence e lec a di e en ial exposu e o
social capi al: high us indi iduals ha e access o con ex ual social capi al, while
low us indi iduals do no ; u he mo e, low us indi iduals may also be exposed
o ‘bad’ social capi al, u he damaging hei heal h.
Figu e 9. E ec o con ex ual social capi al on heal h dis ibu ion: esul s o he i s publica ion
(Campos-Ma os I, Sub amanian SV, Kawachi I. The ‘da k side’ o social capi al: us and sel - a ed
heal h in Eu opean coun ies. Eu opean Jou nal o Public Heal h. 2016;26(1):90-95).
I is impo an o no e ha applying he amewo k o only one analysis does no
exhaus all possible mechanisms by which one con ex ual de e minan ope a es.
Con ex ual social capi al can impac HI, as people a e no equally exposed o i s
e ec s, and i can also impac o e all popula ion heal h – as some au ho s a gue and
indeed ha e shown (4, 70). This amewo k is help ul o clea ly iden i y he
mechanisms unde s udy in an empi ical analysis, and u he sugges s o he
mechanisms ha may also exis and no ha e been de ec ed in ha pa icula analysis.
3.4.2. Social Mobili y and Heal h in Eu opean Wel a e Regimes
This analysis ocused on how he associa ion be ween social mobili y and heal h can
di e be ween Eu opean wel a e egimes. Impo an ly, i p o ided some e idence ha
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
124
heal h selec ion does no seem o be an impo an de e minan o HI in coun ies wi h
high le els o social mobili y. To apply he amewo k o his analysis, he ocus will
be on how he con ex ual de e minan (wel a e egime) can ha e an impac on heal h
and heal h dis ibu ion h ough he indi idual de e minan (social mobili y). The key
messages ega ding he impac o wel a e egime h ough social mobili y we e:
• The e we e signi ican di e ences in o e all heal h be ween di e en wel a e
egime ypes – Scandina ian coun ies showed he bes and o me USSR
coun ies he wo se esul s; and
• Indi iduals who we e upwa dly mobile showed be e heal h han hose who
we e socially s able, bu upwa d mobili y had a small impac in Scandina ian
coun ies, compa ed o a la ge impac in o me USSR coun ies.
Fi s , i is clea ha wel a e egime is associa ed wi h di e ences in o e all
popula ion heal h. The e a e many ways in which ce ain wel a e egimes can ha e an
impac on heal h – i can be ela ed o economic de elopmen , heal h se ices
p o ision, and cul u al aspec s, among o he s. One possible way is h ough social
mobili y. Poli ical choices, such as how o p o ide educa ion o how o edis ibu e
weal h, ha e he po en ial o b eak he in e -gene a ional ansmission o social
disad an age and imp o e he social posi ion o many, ega dless o hei pa en s’
social s anding. Thus, he ex en o social mobili y a ies subs an ially be ween
wel a e egimes (71) and can po en ially lead o be e socioeconomic ci cums ances
ac oss he popula ion. This is e iden by he obse a ion ha coun ies whe e social
mobili y is highes also end o ha e highe o e all educa ional le els (72). I is by
his e ec on social mobili y ha wel a e egimes, h ough imp o emen s in people’s
social posi ion, can lead o a be e (o wo se) le el o o e all popula ion heal h
( igu e 10).
Second, upwa d mobili y was associa ed wi h be e heal h in all coun ies, bu he
di e ence was conside ably bigge in o me USSR coun ies and smalles in
Scandina ian coun ies. This sugges s ha Scandina ian coun ies a e mo e e ec i e
a sepa a ing social mobili y om heal h. This means ha when indi iduals climb up
(down) he social ladde , hey a e mo e likely o ha e be e (wo se) heal h, especially
i hey li e in o me USSR coun ies. This can e lec a mechanism o social
s a i ica ion, as wel a e egimes de e mine how indi iduals a e placed in a socie y
and, consequen ly, hei heal h. F om a heal h selec ion pe spec i e, i may also mean
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
125
ha indi iduals who a e ill (heal hy) a e much mo e likely o all (climb) in he social
ladde in o me USSR coun ies, when compa ed o Scandina ian coun ies. This
second possibili y can e lec a mechanism o di e en ial consequences, as a pe son’s
heal h leads o changes in hei social posi ion.
Figu e 10. E ec s o wel a e egimes and social mobili y on heal h and heal h dis ibu ion: esul s o
he second publica ion (Campos-Ma os I, Kawachi I. Social mobili y and heal h in Eu opean coun ies:
does wel a e egime ype ma e ? Social Science and Medicine. 2015;142:241-248).
The esul s o his analysis e lec social mobili y’s e ec s on heal h and HI, since
mo e gene ous wel a e egimes no only show be e o e all heal h, bu also ewe
inequali ies be ween social mobili y g oups. Thus, wel a e egime, h ough i s e ec s
on social mobili y, can impac bo h he heal h and he heal h dis ibu ion o a
popula ion ( igu e 10). This analysis, o cou se, does no include o he e ec s ha
wel a e egimes can ha e bo h on heal h and HI, as i ocuses exclusi ely on he
e ec s ha a e media ed by social mobili y.
3.4.3. Shi ing De e minan s o Heal h Inequali ies in Po ugal
The ocus o his analysis was on how HI ha e changed in Po ugal o e he las
decade. Al hough we did no es he e ec o any con ex ual de e minan , we
in e p e ed he esul s conside ing he subs an ial economic changes ha we e
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
126
happening in he coun y a he ime. Fo he applica ion o he amewo k, he
ollowing key poin s summa ize he mos ele an esul s o his analysis:
• O e all popula ion heal h de e io a ed in Po ugal be ween 2010 and 2011;
• This de e io a ion was s eepe in he iches e cile, hus dec easing
inequali ies be ween income e ciles; and
• The main d i e o his dec ease was he change be ween ac i e and inac i e
g oups: while limi a ions we e mo e p e alen in inac i e people un il 2010,
hey became mo e p e alen in ac i e people a e his yea .
The o e all de e io a ion in popula ion heal h in Po ugal be ween 2010 and 2011 is
likely o be ela ed o o e all economic changes happening a he ime. These e en s,
such as implemen a ion o highly publicized aus e i y measu es, can ope a e h ough a
s ess-de e mined pa hway, as a clima e o unce ain y can lead o s ess and,
consequen ly, poo e heal h, leading o wo se heal h ou comes ac oss he popula ion.
In he amewo k, his can be in e p e ed as an e ec on exposu e o a isk ac o :
economic changes a e he con ex ual de e minan ha lead o an inc ease in he
exposu e o a isk ac o – s ess – hus ha ing an e ec on he whole popula ion’s
heal h ( igu e 11). Howe e , no e e yone eac ed he same way o his unce ain y. In
ac , iche people ( he iches e cile) seemed o su e he g ea es hi . This was
g ea ly media ed by he ac ha ac i e people – who also ended o be iche – had a
highe p e alence o limi a ions a e 2010 han inac i e people. Despi e no ha ing
o mally es ed his, we hypo hesized ha wo mechanisms migh be behind his
change. Fi s , people om highe socioeconomic classes (who end o ha e highe
incomes) may no be as used o dealing wi h unce ain y as people om lowe
socioeconomic classes a e. To hem, he p ospec o unce ain imes ahead could ha e
led o mo e in ense s ess eac ions and o an inabili y o deal wi h p ac ical day- o-
day p oblems on a mo e es ic ed budge . This i s mechanism can be iden i ied in
he amewo k as di e en ial ulne abili y, as all g oups we e exposed o unce ain y
and s ess, bu – pe haps coun e in ui i ely – high SES people we e mo e ulne able,
a leas du ing a ce ain pe iod o he ime. Second, conside ing he high emig a ion
a es Po ugal was going h ough a he ime (73), he g oup o ac i e people who was
‘le behind’ migh ha e had disp opo iona ely high a es o limi a ions, as mig a ion
is known o be a selec i e p ocess by which heal hie people end o mig a e mo e
(74). This second mechanism is a consequence o changes in he composi ion o he

The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
127
popula ion. This can be seen as a change in social s a i ica ion, as changes in he
economic con ex led o changes in how weal h and powe we e dis ibu ed o
di e en social posi ions.
Figu e 11. E ec s o mac oeconomic changes on heal h and heal h dis ibu ion: esul s o he ou h
publica ion (Campos-Ma os I, Russo G, Gonçal es L. (accep ed o publica ion). Shi ing de e minan s
o heal h inequali ies in uns able imes: Po ugal as a case s udy. Accep ed o publica ion in he
Eu opean Jou nal o Public Heal h in May 2017. DOI: h ps://dx.doi.o g/10.1093/eu pub/ckx080).
O e all, his analysis sugges s ha economic and social changes in a coun y can lead
o changes in heal h and heal h dis ibu ion h ough h ee mechanisms: social
s a i ica ion, exposu e o isk ac o s, and di e en ial ulne abili y o hose isk
ac o s ( igu e 11). Once again, his does no p o ide an exhaus i e desc ip ion o
how economic c ises impac heal h and heal h dis ibu ion, bu sugges s some o he
mechanisms ha may ope a e and ames he indings in a la ge con ex .
3.4.4. Summa y
The amewo k p oposed he e aims o ou line he mechanisms ha connec con ex ual
cha ac e is ics o popula ion heal h and heal h dis ibu ion. I was applied o he
analysis o h ee con ex ual de e minan s: social capi al, wel a e egime, and
mac oeconomic changes. Each o hese de e minan s in luenced heal h ou comes and
hei dis ibu ion h ough a ious mechanisms: di e en ial exposu e (social capi al);
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
128
social s a i ica ion, social posi ion, and di e en ial consequences (wel a e egimes);
and social s a i ica ion, exposu e, and di e en ial ulne abili y (economic c isis).
The amewo k p o ed a use ul ool o ame hese publica ions, in which con ex ual
de e minan s we e explo ed. In pa icula , i acknowledged he impo ance o con ex
in se e al s eps in o de o in luence bo h heal h and HI. In u u e esea ch, i can
p o ide a s uc u e o acili a e a e lec ion abou he heo e ical basis ha unde sco es
analyses, hus s eng hening hem and he a gumen s hey p opose.
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
129
3.5. Con ibu ion o Policy and Resea ch
3.5.1. Rega ding Heal h Inequali ies
This esea ch has p esen ed a new amewo k on how con ex can shape heal h and
heal h dis ibu ion. This amewo k is buil on a amewo k p oposed by Dide ichsen
e al. ( igu e 3) (52), and inco po a es a ew impo an di e ences om he o iginal
amewo k. Fi s ly, i explici ly ocuses on con ex . Fo Dide ichsen, con ex was
esponsible o social s a i ica ion and policies. In he amewo k pu o wa d he e, i
is p oposed ha con ex is seen is a b oade pe spec i e, encompassing also physical
and o he social elemen s (besides policies). I is also p oposed ha con ex no only
con ibu es by c ea ing a sys em o social s a i ica ion and policies, bu also by
in luencing all he o he s eps o he pa hway be ween social posi ion and heal h
ou comes. Fu he mo e, his amewo k highligh s ha con ex can in luence heal h
and heal h dis ibu ion, and i does so by di e en mechanisms.
Ano he impo an dis inc ion om Dide ichsen’s wo k is ha in he o me
amewo k di e en ial ulne abili y is he accumula ion o ha m ul isk ac o s in a
pa icula popula ion g oup. In applying he new amewo k o he analyses, i was
ound ha di e en ial ulne abili y does no necessa ily mean ha he same
indi iduals a e always ulne able. Simila ly, di e en ial exposu es and consequences
do no necessa ily mean ha he poo es o leas educa ed will always be mo e
exposed o isk ac o s o su e g ea ly he consequences o ill-heal h. In ou wo k, i
is p oposed o ede ine he e m ‘di e en ial’ o mean only ha e ec s a e di e en in
di e en socioeconomic g oups, wi hou sugges ing a di ec ion o he e ec .
Hope ully, his amewo k can be used in u u e esea ch p ojec s o collec e idence
o he e ec o a con ex ual de e minan . Fo example, o pe o m a e iew o how
economic g ow h can impac on heal h and heal h dis ibu ion, he collec ed e idence
can be o ganized using his amewo k. This amewo k could also help esea che s
ou line a clea heo e ical basis o hei wo k, hus building a s onge e idence base
a ound he opic o hei esea ch. This is pa icula ly impo an conside ing he lack
o a clea heo y abou he mechanisms by which con ex ope a es has been one o he
ac o s hampe ing he s udy o con ex ual in luences on heal h (75).
Rega ding i s con ibu ion o policy making, as i builds on and u he he cu en
heo ies on con ex ual e ec s on heal h, he amewo k may also help policy make s
The ela ionship be ween con ex and heal h inequali ies Chap e 3. Discussion and Conclusions
130
unde s and he op ions a ailable o ackle speci ic aspec s o HI. Pa icula ly, i could
p o e a use ul ins umen o equali y impac assessmen s – he analysis o policies
ha ies o ensu e hey do no disc imina e agains any ulne able g oup (76).
Besides being a p ac ical ins umen o esea che s and policy make s, he
amewo k will hope ully con ibu e o he ising end o a di e en hinking abou
HI. Namely, i emphasizes how HI a e no a p oduc o indi idual cha ac e is ics
alone, bu a consequence o he in e ac ion be ween con ex and indi iduals. This can
ha e impo an implica ions on how HI a e ackled – om a ocus on indi idual
beha iou change, policy make s should also hink abou enabling and disabling
cha ac e is ics ha con ex s can p o ide o in luence hose beha iou changes.
Ano he con ibu ion o he p esen esea ch is he concu en analysis o se e al
indi idual SES cha ac e is ics as an oppo uni y o explo e he mechanisms behind
HI. This ype o analysis is no common, bu p o ed e y p oduc i e. In his analysis,
his allowed o he explo a ion o wha migh ha e been behind he dec ease in HI in
Po ugal a e 2010, hus laying g oundwo k o u he esea ch on he opic. This
me hod p o ided in e es ing le els o analysis, and could be ep oduced in u u e
analyses in o he con ex s.
3.5.2. Rega ding Eu ope
This esea ch also p o ided some con ibu ion o he unde s anding o HI in Eu opean
coun ies. The pe sis ence o HI in Eu ope has been called a ‘pa adox’ and social
mobili y has been pu o wa d as one o a ew possible explana ions (10). The p esen
indings on social mobili y in di e en wel a e egimes in Eu ope show ha his is
highly unlikely, o coun ies wi h high social mobili y would show a leas as high
inequali ies in heal h be ween mobili y g oups as coun ies wi h low mobili y, which
was no he case. This can help mo e o wa d he explo a ion o he ‘pa adox’, as
o he explana ions a e now mo e likely and should be u he in es iga ed. In e ms o
policy-making, hese indings sugges ha i is possible o mi iga e he associa ion
be ween social mobili y and heal h. Al hough all wel a e egimes showed some kind
o associa ion be ween social mobili y and SAH, some we e signi ican ly smalle ,
sugges ing ha i is possible o sepa a e he wo mo e e icien ly, and ha he answe
lies in he di e ences be ween he wel a e egimes.