Heal h policy 126 (2022) 680–687
A ailable online 25 Ap il 2022
0168-8510/© 2022 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-
nc-nd/4.0/).
Do indi iduals’ heal h p e e ences alida e he decen alisa ion o he
public heal h sys em in Spain?
Waleska Sigüenza
a
,
*
, Alai z A abe
b
a
Depa amen o Polí icas Públicas e His o ia Econ´
omica, Uni e sidad Del País Vasco (UPV/EHU), A enida Lehendaka i Agui e 83. 48015, Bilbao, SPAIN
b
Depa amen o Polí icas Públicas e His o ia Econ´
omica, Uni e sidad Del País Vasco (UPV/EHU), A enida Lehendaka i Agui e 83. 48015, Bilbao, SPAIN
ARTICLE INFO
Keywo ds:
Decen alised policy
Heal h se ices
Indi idual p e e ences
ABSTRACT
The objec i e o his pape is o es ima e indi iduals’ p e e ences abou public heal h se ices in wo Spanish
egions, he Basque Coun y (BC) and Cana y Islands (CI) and analyse whe he hey di e . This wo k was
mo i a ed by he ac ual economic si ua ion, whe e i is necessa y o ob ain equilib ium be ween he needed
heal h se ices and limi ed economic esou ces. Wi h his limi a ion in mind, poli icians ha e ied o design
heal h policies ha maximise indi iduals’ wel a e. Based on he heo y o decen alisa ion, he de olu ion o
public expendi u e decisions and managemen o egional go e nmen maximises indi iduals’ wel a e mo e
when indi idual p e e ences di e among egions. A disc e e choice expe imen was implemen ed wi h a su ey
designed o ob ain da a abou indi iduals’ choices. Using his da a and disc e e choice models, indi idual
p e e ences o heal h se ices we e es ima ed. Ou indings indica e ha hese p e e ences di e among egions,
so, o easons o e iciency, decen alising decisions and managemen o public heal h policies o egional
go e nmen s would be ecommended. Once heal h policies a e decen alised, ou esul s p o ide a ool o
iden i ying he heal h se ices mos alued by he indi iduals in each egion. This in o ma ion would be use ul
policymake s designing heal h policies.
1. In oduc ion
In any heal h sys em – public o p i a e – knowledge o po en ial
use s’ p e e ences is a key issue o policymaking. Many o he aspec s
such as epidemiological, echnical, o human a iables a e a leas as
impo an as hese p e e ences, bu he e is no doub abou he ele ance
o pa ien s’ (o po en ial pa ien s’) p e e ences in he design o heal h
sys ems. Such in o ma ion could be e y ele an o policymake s in
key issues, such as in which se ice o add (o de ac ) money, which is
he main a iable in he consume decision o pu chase p i a e heal h
insu ance o he accep abili y o co-paymen s in a public sys em. Fo
some na ional heal h sys ems (NHS) such as in G ea B i ain o Spain, in
which he pu chasing o p i a e insu ance implies ha ing duplica e
heal h co e age, he knowledge o hese p e e ences could shed ligh on
p i a e insu ance demand, especially i he p ecise p e e ence compo-
nen s a e well speci ied.
The case o Spain is o pa icula in e es because o he s uc u e o
i s NHS. The Spanish sys em is di ided in o au onomous communi ies
(ACs), which ha e a conside able amoun o eedom o decide upon
expendi u e and managemen . This c ea es a scena io whe e he e a e
di e en sys ems wi hin a sys em; he esul is a di e en ial homoge-
neous amewo k wi h di e en pe cei ed (and eal) quali ies in he
NHS. This scena io seems qui e app op ia e o ou objec i e (see he
Appendix o a mo e de ailed explana ion abou Spanish ins i u ional
s uc u e). The jus i ica ion o he decen alisa ion o expendi u e
policies lies in e iciency – ha is, he decen alisa ion o public
expendi u e o he egions maximises he wel a e o a g ea e numbe o
indi iduals [4,12,13]. To con i m ha he public heal h expendi u e
should be decen alised in Spain, i is necessa y o know indi iduals’
p e e ences abou heal h policies. I indi iduals’ e alua ions di e
among ACs and he objec i e is o maximise global wel a e, hen, i
would be easonable o decen alise heal h expendi u e and implemen
di e en heal h policies among egions.
This s udy is an ex ension o p e ious esea ch done by Sigüenza and
Ma iel [18]. They p o ide he easoning o assigning a mone a y
alua ion o he imp o emen in indi idual p e e ences caused by an
imp o emen in heal h se ices. The lack o e e ence p ices o public
heal h sys ems (PHS) has been iden i ied as a main p oblem, as his lack
o p ices makes he e icien alloca ion o public esou ces mo e di icul
among di e en heal h policies, while aking in o accoun hese policies’
* Co espondig au ho
E-mail add esses: [email p o ec ed] (W. Sigüenza), [email p o ec ed] (A. A abe).
Con en s lis s a ailable a ScienceDi ec
Heal h policy
jou nal homepage: www.else ie .com/loca e/heal hpol
h ps://doi.o g/10.1016/j.heal hpol.2022.04.010
Recei ed 23 Feb ua y 2020; Recei ed in e ised o m 12 July 2021; Accep ed 21 Ap il 2022
Heal h policy 126 (2022) 680–687
681
e ec on indi idual and social wel a e [3],[15]. We ag ee wi h Sigüenza
and Ma iel [18] ha ha ing knowledge o he mone a y alue o an
imp o emen in indi idual p e e ences caused by a speci ic imp o e-
men in heal h se ices may allow us o ob ain help ul ools o p io i ise
NHS policies and inc ease hei e iciency [5,9,14].
Sigüenza and Ma iel [18] measu ed he economic alua ion o in-
di idual p e e ences, conside ing he he e ogenei y o obse ed and
unobse ed indi idual p e e ences. Fo ha pu pose, indi idual cha -
ac e is ics, such as educa ional le el and income le el a e conside ed.
They calcula ed he willingness o pay (WTP) o public heal h se ices
o measu e he economic alua ion o indi idual p e e ences in he
Basque Coun y (BC). While hei main objec i e was o de e mine he
alue o heal h se ices in he BC, he main objec i e he e is o analyse
he di e ences in he alua ion be ween wo Spanish egions ha ha e
e y di e en heal h policies and expendi u e le els in public heal h and
pe cei ed quali y, such as he BC and he Cana y Islands (CI). O he
issues also a y be ween hese wo egions, such as clima ology, habi s
and cus oms. These di e ences may in luence indi iduals’ heal h se -
ice p e e ences. Conside ing ha he Spanish NHS is di ided in o e-
gions, di e ences may exis among hei economic alua ion o heal h
se ices. Di e ences in he WTP o BC and CI may be caused by obse ed
o unobse ed indi idual cha ac e is ics. Once we con ol o indi idual
cha ac e is ics and policy a ibu es, he di e ences be ween egional
WTPs indica e ha indi iduals in di e en ACs e alua e possible heal h
policies di e en ly. The main objec i e o his esea ch is o de e mine
whe he o no he economic alua ion o indi idual p e e ences abou
public heal h se ices a e equal o no be ween egions. I we can epo
a leas an example o wo egions whe e indi iduals ha e di e en
p e e ences abou heal h se ices, hen we can d aw he conclusion ha
he ini ial hypo hesis o equali y o p e e ences is no sa is ied. In his
case, he decen alisa ion o heal h policies in o egions would be
jus i ied o easons o e iciency and wel a e maximisa ion. Mo eo e ,
he iden i ica ion o indi idual WTP o imp o emen s in heal h se ices
may be help ul o designing heal h policies o inc ease indi iduals’
wel a e.
A disc e e choice expe imen (DCE) was implemen ed o assign a
mone a y alua ion o di e en heal h sys em se ices. A DCE can
conside he mul idimensional na u e o heal h se ices [10] and p o-
ide in o ma ion abou ma ginal changes in se ice. The DCE me hod is
hus a help ul ool o ob ain and analyse in o ma ion abou indi idual
p e e ences and o designing heal h policies [20]. The es o he pape
is o ganised as ollows. The nex sec ion desc ibes he s uc u e o he
su ey and he da a. The hi d sec ion explains he implemen ed
me hodology, ollowed by he esul s in sec ion ou . Finally, sec ions
i e and six p esen he discussion and conclusions.
2. Su ey s uc u e and da a
2.1. Su ey s uc u e
The su ey s uc u e used o e alua e he heal h se ices om he CI
is he same used by Sigüenza and Ma iel [18] o e alua e he same
se ices in he BC so he esul s should be compa able. The su ey is
di ided in o ou sec ions: basic in o ma ion, in o ma ion abou a i-
bu es, choice ca ds and sociodemog aphic a iables (see he Appendix
o mo e in o ma ion abou he su ey).The mos impo an pa o he
su ey a e he 12 choice ca ds ha ep esen hypo he ical choice sce-
na ios p esen ing h ee di e en al e na i es: A, B and C. These hypo-
he ical scena ios a e de ined by di e en a ibu es ep esen ing
di e en heal h co e age op ions abou which he esponden mus ake
a choice. The ele an heal h se ice a ibu es and p o ision le els we e
selec ed using he quali a i e echnique o ocus g oups. The a ibu es
we e Specialis Wai ing Lis (Specialis ), Su ge y Wai ing Lis (Su ge y),
Hospi al Com o (Com o ), Medical A en ion (A en ion), Den al Co e age
(Den al) and Cos .
Table 1 epo s he i e a ibu es o he selec ed se ice and he
co esponding cos o each possible combina ion o a ibu es – ha is,
he cos o each al e na i e – as se ou by he expe s in he ocus g oup.
Al e na i e A ep esen s he ac ual si ua ion, while al e na i es B and C
p esen di e en alues o each o he a ibu es as a esul o a speci ic
le el o change. These wo al e na i es ha e an associa ed cos ha he
indi idual would pay mon hly.
2.2. Da a
The su eyin e iews we e ca ied ou wi h 235 and 255 selec ed
indi iduals in he BC and CI, espec i ely, using simple andom sam-
pling in he yea s 2012 in he BC and 2016 in CI. Two dummy a iables
a e de ined o es ima e he model. Uni e si y is a dummy a iable ha
akes he alue 1 i he indi idual has uni e si y s udies. HighIncome
measu es he mon hly income le el and akes he alue 1 i he amily’s
mon hly ne income is mo e han
€
3,000.Two il e s a e applied o he
da a o ob ain highe eliabili y in he esul s [2]. The i s il e is a
a ionali y es , which p o ide an addi ional choice ca d o which e-
sponden s aced a highe ma ginal imp o emen in al e na i e C han in
al e na i e B, bu wi h a lowe cos , so we expec ed esponden s always
o choose al e na i e C because a choice o al e na i e B is conside ed
i a ional. O he o al 235 and 255 esponden s in BC and CI, espec-
i ely, 12 and 14 indi iduals chose he i a ional op ion and we e
excluded om he inal sample.
The second il e sough o iden i y p o es esponden s: we asked
indi iduals who chose al e na i e A o any o he 13 ca ds he eason o
hei choice. Those esponden s who answe ed ha his ini ia i e should
be inanced en i ely by he go e nmen and ha hey al eady pay oo
much ax a e iden i ied as p o es esponden s, because hey a e
conside ed o be unwilling o indica e hei WTP o imp o emen s in
heal h se ices in CI. We can conclude ha hese indi iduals a e agains
making any paymen s o imp o e he analysed se ice.
Once hese il e s we e applied, he es ima ions we e made exclu-
si ely wi h alid da a, which should yield mo e eliable esul s. The da a
analysis used, in o al, 150 and 146 comple ed ques ionnai es om BC
Table 1
A ibu es and le els conside ed in he DCE
Label Desc ip ion Le els
Specialis A e age wai ing ime o
specialis a en ion and
complemen a y medical
es s.
1
mon h
2
mon hs
3*
mon hs
—
Su ge y A e age wai ing ime o
no u gen su gical
in e en ion.
2
mon hs
3
mon hs
5
mon hs
8*
mon hs
Com o O e ed acili ies, such
as single oom equipped
wi h TV and in e ne ,
p esence o a companion
and ee access o he
amily.
Low* Medium High —
A en ion Pe sonalised, close and
p o essional medical
a en ion.
Low* Medium High —
Den al Low: oo h ex ac ion.
Medium: oo h
ex ac ion, den al
hygiene and den al
illings. High: oo h
ex ac ion, den al
hygiene, whi ening and
den al illings.
Low* Medium High —
Cos Mon hly p emiums paid
by he indi idual who
wan s o bene i om
co e age o e ed in each
choice op ion.
€
0*
€
30
€
60
€
90
Sou ce: compiled by au ho s
No es: *Cu en s a us (no ac ion le el)
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
682
and CI, espec i ely, yielding 1,800 and 1,752 obse a ions. As Sigüenza
and Ma iel [18] concluded, he BC sample is ep esen a i e. Fo he CI,
he sample is also ep esen a i e in e ms o gende , income and age.
(See he Appendix o mo e in o ma ion abou he da a and sampling).
2.3. Model
The andom u ili y heo y (RUT) assumes ha indi iduals’ u ili y is
de e mined by some obse able ac o s and an unobse able andom
pa ame e . To con e indi idual’s choice esponses in DCE in o es i-
ma ed pa ame e s, we employed he beha iou al amewo k o he RUT.
I is assumed ha he u ili y indi idual i ob ains when choosing al e -
na i e j in he h choice occasion, Ui ,j, is gi en by:
Ui ,j=Vi ,j+
ε
i ,j
whe e Vi ,j is he de e minis ic pa o he u ili y unc ion ha con ains
obse able ac o s and
ε
i ,j is an e o e m.
Mul inomial logi (MNL) is he basic model used o analyse he
choices in DCE o es ima e he e ec ha obse able ac o s ha e on he
u ili y. MNL assumes ha he de e minis ic pa o he u ili y, Vi ,j, is
shaped by obse able co a ia es and unknown pa ame e s ( o be es i-
ma ed) and he andom e o
ε
i ,j is Type I ex eme alue dis ibu ed
[11]. Thus, gi en a se o al e na i es, he u ili y indi idual i ob ains in
he h choice occasion om al e na i e j, Ui ,j, is gi en by
Ui ,j=Vi ,j+
ε
i ,j=Xi ,j
′β+
ε
i ,j(1)
whe e Xi ,j is a ec o o co a ia es o al e na i e j in choice occasion o
indi idual i and β is a ec o o pa ame e s. The ec o Xi ,j may include a
cons an e m o allow o al e na i e-speci ic cons an s (ASC). In ou
special case, we conside wo ASC.
Unde he s anda d assump ions, he p obabili y ha indi idual i
chooses al e na i e j in he h choice is as ollows:
P(j
Xi ,j)=eXi ,j
′β
∑jeeXi ,j
′β(2)
MNL assumes ha pa ame e s a e cons an among indi iduals. Thus,
i only al e na i e a ibu es a e in oduced in he de e minis ic pa , he
e ec ha hese a ibu es ha e on u ili y is homogenous among in-
di iduals. When he e ec s o indi idual-speci ic cha ac e is ics need o
be accoun ed o , hese a e in oduced in he model as in e ac ion e ms
wi h al e na i e a ibu es. This p ocedu e allows us o in oduce
obse ed p e e ence he e ogenei y caused by hese indi idual-speci ic
a iables – ha is, by di e en g oups o indi iduals. In ou case, he
e ec o educa ion and income is in oduced in he u ili y unc ion (1) as
ollows:
Because he samples a e ela i ely small, mo e obus model es i-
ma ion me hods such as MNL a e ecommended, while mo e complex
models such as andom pa ame e logi (RPL) should be a oided,
because hey may lead o ins abili y in he es ima ions. MNL wi h in-
e ac ions be ween a ibu es and sociodemog aphic cha ac e is ics o
p e e ence he e ogenei y was he e o e chosen as he es ima ion
me hod. The WTP o a uni o change in a gi en a ibu e o he p oduc
can be compu ed as he ma ginal a es o subs i u ion be ween he
quan i y exp essed by he a ibu es and he cos o he choice. In he
MNL, o example, he WTP o a ibu e Specialis is exp essed as
ollows:
WTPSpecialis = − β1+β12Uni e si y +β13HighIncome
β6+β62Uni e si y +β63HighIncome (4)
The WTPs o he emaining a ibu es a e de ined acco dingly.
The dis ibu ion o he WTP is ob ained by Boo s ap [7], d awing on
a la ge numbe o samples o size N (wi h eplacemen s) om he es i-
ma ion sample. In ou case, 600 samples o size 150 and 146 a e d awn
(wi h eplacemen s) om he es ima ion samples o BC and CI, espec-
i ely. Each o he samples is used o es ima e he model and each co -
esponding WTP, using equa ion 4. We hus ob ain he es ima ed
dis ibu ion o he WTP in BC and CI. The me hod p oposed by Poe e al.
[16,17] is used o es he di e ence be ween he dis ibu ions o WTP in
he wo egions.
3. Resul s
This pape analyses he di e ences in WTP o PHS se ices in he BC
and CI. Once we con ol o indi idual cha ac e is ics, we can analyse
he ac o s de e mining he di e ences in WTP be ween egions, as well
as iden i ying he ac o s ha in luence WTP and di e ences in WTP
among indi iduals o each egion.
As he main objec i e o his s udy is o compa e he WTP be ween
wo ACs, he same model speci ica ion is assumed o BC and CI. The
mos gene al speci ica ion is hus conside ed, whe e all possible in-
e ac ions be ween he a ibu es and dummy a iables a e in oduced.
Table 2 epo s he es ima ions o he MNL model wi h in e ac ions
be ween a ibu es and indi idual sociodemog aphic cha ac e is ics.
The i s se o pa ame e s (columns 2, 3, and 4) shows he es ima es o
he BC. The expec ed signs a e ob ained o he a ibu es pa ame e s.
The posi i e sign o ASC1 and ASC2, which a e associa ed wi h al e -
na i e B and al e na i e C, espec i ely, indica e ha indi iduals p e e
al e na i e B and C o he s a us quo –, ha is, hey would p e e he
imp o emen s p oposed in al e na i e B and C (assuming hei cos s) o
he ac ual si ua ion. Fu he mo e, a nega i e e ec in Specialis indica es
ha indi idual u ili y dec eases wi h an inc ease in wai ing lis s. The
same e ec on indi idual u ili y is ob ained wi h an inc ease in Su ge y
wai ing lis s. On he o he hand, posi i e and signi ican es ima es o
pa ame e s o he a ibu es Com o , A en ion, and Den al e lec ha
inc eases in hospi al com o , medical a en ion and den al co e age
inc ease BC indi iduals’ u ili y. A nega i e and signi ican e ec o he
in e ac ion Com o ×HighIncome indica es ha an inc ease in Com o
inc eases indi idual u ili y mo e o indi iduals wi h a low income han
o indi iduals wi h a high income le el. The same e ec can be seen o
he in e ac ion A en ion ×HighIncome, which indica es ha he inc ease
in u ili y caused by an imp o emen in A en ion is highe o indi iduals
wi h low income han o indi iduals wi h a high income. An inc ease in
cos implies a highe u ili y dec eases among low-educa ed indi iduals.
The second se o pa ame e s (columns 5, 6 and 7) epo s he es i-
ma es o he MNL model o CI. As in he case o he BC, o CI, he
posi i e sign o ASC1 and ASC2 also indica es ha indi iduals p e e
Ui ,j=X′
i ,jβ+
ε
i ,j=ASCj+β1Specialis i ,j+β12 Specialis i ,j×Uni e si yi+β13 Specialis i ,j×HighIncomei+β2Su ge yixa ,j
+β22 Su ge yi ,j×Uni e si yi+β23 Su ge yi ,j×HighIncomei+β3Com o i ,j+β32 Com o i ,j×Uni e si yi
+β33 Com o i ,j×HigIncomei+β4A en ioni ,j+β42 A en ioni ,j×Uni e si yi+β43 A en ioni ,j×HighIncomei
+β5Den ali ,j+β52 Den ali ,j×Uni e si yi+β53 Den ali ,j×HighIncomei+β6Cos i ,j+β62 Cos i ,j
×Uni e si yi+β63 Cos i ,j×HighIncomei+
ε
i ,j
(3)
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
683
al e na i es B and C o he s a us quo – ha is, hey would p e e he
imp o emen s p oposed in al e na i es B and C (assuming hei cos s) o
he ac ual si ua ion. Mo eo e , an inc ease in Su ge y dec eases indi-
idual u ili y, while an imp o emen in A en ion inc eases u ili y. This
inc ease is highe o indi iduals wi h uni e si y s udies. Ne e heless,
an inc ease in Cos dec eases indi idual u ili y, and his dec ease is also
highe o indi iduals wi h Uni e si y s udies. This se o indi iduals,
hose wi h Uni e si y s udies, also show inc eased u ili y wi h imp o e-
men s in Com o .
The highe absolu e alue coe icien es ima es o he BC imply ha
he e ec o he unobse ed ac o s ep esen ed in he e o e m is
lowe han in CI, and hus ha BC indi iduals a e mo e de e minis ic –
ha is, he a ibu es ha e mo e impo ance ela i e o an unobse ed
ac o , in he BC han in he CI.
The WTP es ima es epo ed in Table 3, wi h each co esponding
95% con idence in e al, we e ob ained using he MNL es ima es om
Table 2. Fou di e en subg oups o indi iduals (based on sociodemo-
g aphic cha ac e is ics) we e conside ed o calcula e he WTP. The i s
subg oup (uppe le block o Table 3) comp ises indi iduals wi h no
educa ion, p ima y educa ion, and seconda y educa ion (Uni e si y=0),
along wi h low income (HighIncome=0). The second subg oup (uppe
igh block) consis s o indi iduals wi h uni e si y educa ion (Uni-
e si y=1) and low income (HighIncome=0). The indi iduals wi h high
income (HighIncome=1) and p ima y and seconda y educa ion (Uni-
e si y=0) a e in he hi d subg oup (lowe le block), while he in-
di iduals wi h high income (HighIncome=1) and uni e si y educa ion
(Uni e si y=1) a e in he ou h subg oup (lowe igh block). The i s
column o each es ima e subg oup con ains he a ibu es, and he sec-
ond and hi d columns epo he WTP o each a ibu e in he BC and
CI, espec i ely, wi h he co esponding 95% con idence in e al. The
WTP o each a ibu e in he BC is compa ed wi h each co esponding
WTP in he CI, using he Poe e al. [16,17] equali y es . The ou h
column o each es ima e subg oup epo s he p- alue o he null hy-
po hesis ha he wo coe icien s a e equal.
The WTP o he Specialis and Su ge y a ibu es a e nega i e due o
hei de ini ion, so hese alues mus be in e p e ed as a WTP o a oid a
ma ginal inc ease in he wai ing lis s. Howe e , he WTPs o Com o ,
A en ion and Den al a e in e p e ed as a WTP o a ma ginal inc ease in
he le els o hospi al Com o , medical A en ion and Den al co e age,
espec i ely.
When low income (HighIncome=0) indi iduals wi h p ima y and
seconda y educa ion (Uni e si y=0) a e conside ed (uppe le block), a
signi ican di e ence is obse ed be ween he WTP in he BC and in he
CI o he a ibu es Com o , A en ion and Den al, wi h BC indi iduals
ha ing a highe WTP o an imp o emen in Com o , A en ion and
Den al co e age. Howe e , he WTP o a dec ease in he wai ing lis s is
equi alen o bo h g oups. I he analysis is cen ed on he se o in-
di iduals wi h high income (HighIncome=1) and p ima y and seconda y
educa ion (Uni e si y=0; uppe igh block), he WTPs o he BC and CI
indi iduals a e equal o mos a ibu es, excep o Su ge y, whe e BC
indi iduals a e willing o pay a highe amoun o educe he wai ing
lis s.
The se o indi iduals wi h he bigges di e ences be ween he BC
and CI egions consis s o indi iduals wi h high income (HighIncome=1)
and uni e si y s udies (Uni e si y=1; lowe igh block). Indi iduals
wi h hese cha ac e is ics a e willing o pay mo e o an imp o emen in
Com o and A en ion in he CI han in he BC. Howe e , BC indi iduals
a e willing o pay mo e han indi iduals in he CI o dec ease Su ge y
wai ing lis s. These di e ences be ween he WTP o indi iduals in he BC
and CI may be due o he ac ha BC and CI indi iduals’ WTP a e
in luenced by di e en indi idual cha ac e is ics – ha is, he ac o s
ha in luence BC indi iduals’ alua ion o heal h se ices di e om
Table 2
MNL es ima es
Basque Coun y Cana y Islands
Coe . s d. de . Coe . s d. de .
A ibu es
ASC1 0.50 *** 0.17 1.07 *** 0.17
ASC2 0.28 * 0.07 0.81 *** 0.06
Specialis -0.16 * 0.08 -0.08 0.06
Su ge y -0.15 *** 0.03 -0.13 *** 0.02
Com o 0.31 *** 0.09 0.07 0.07
A en ion 0.65 *** 0.09 0.34 *** 0.07
Den al 0.41 *** 0.11 0.15 * 0.08
Cos -0.03 *** <0.01 -0.02 *** <0.01
In e ac ions
Specialis ×
Uni e si y 0.01 0.09 -0.08 0.09
HighIncome -0.01 0.09 0.05 0.12
Su ge y ×
Uni e si y -0.05 0.03 0.01 0.03
HighIncome -0.04 0.03 -0.02 0.04
Com o ×
Uni e si y -0.05 0.09 0.22 ** 0.09
HighIncome -0.18 ** 0.09 0.04 0.12
A en ion ×
Uni e si y 0.17 * 0.09 0.41 *** 0.09
HighIncome -0.36 *** 0.09 0.11 0.12
Den al ×
Uni e si y -0.05 0.12 0.09 0.12
HighIncome -0.16 0.11 -0.06 0.16
Cos ×
Uni e si y -0.01 * <0.01 -0.01 ** <0.01
HighIncome 7.37E-05 <0.01 -6.79E-04 <0.01
Numbe o obse a ions 1800 1752
Sample size 150 146
loglik. -1621.40 -1579.40
AIC 3282.83 3198.76
BIC 3392.74 3308.16
*,**, *** indica e 10%, 5%, 1% signi icance le el espec i ely
Coe : coe icien es ima e; S d.de .: s anda d de ia ion
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
684
hose ha in luence he alua ion o CI indi iduals.
Table 4 p esen s a compa ison o WTP es ima es, no be ween he
wo analysed egions, bu be ween subg oups o indi iduals in each
egion unde s udy. Table 4 is di ided in wo sub- ables, Table 4.1 and
Table 4.2, o he BC and CI, espec i ely. The i s se o esul s (uppe
le block o Table 4.1.) shows he WTP o BC indi iduals wi h low in-
come and di e en educa ional le els (columns 2 and 3) and he equali y
es s be ween he WTP o low-educa ed indi iduals and high-educa ed
indi iduals (column 4). BC indi iduals wi h no uni e si y educa ion
and wi h uni e si y educa ion appea o be willing o pay he same
amoun o PHS imp o emen s. Thus, educa ion le el is no a signi ican
ac o o ei he low- o high-income BC indi iduals.
In he lowe le block o Table 4.1, whe e he WTP equali y es is
epo ed be ween low- and high-income BC indi iduals, i appea s ha
low-income BC indi iduals a e willing o pay mo e han high-income
indi iduals o imp o emen s in Com o and A en ion. I can he e o e
be concluded ha gi en an educa ional le el, he e is an income-le el
in luence on BC indi iduals’ WTP.
The equali y es s epo ed in he le blocks o Table 4.2 conclude
ha gi en, CI indi iduals wi h low income, hose wi h uni e si y s udies
a e willing o pay mo e han hose wi h no uni e si y s udies o im-
p o emen s in Su ge y, Com o and A en ion (uppe le block o
Table 4.2). Educa ional le el is hus a edundan ac o in he WTP o CI
indi iduals wi h low income. Howe e , he WTP among high-income CI
indi iduals o imp o emen s is he same, ega dless o educa ional le el
(lowe le block o Table 4.2). Gi en he educa ional le el, he e a e no
di e ences be ween he WTP o high- and low-income CI indi iduals
( igh blocks o Table 4.2). Thus, income le el does no appea o be a
signi ican ac o o CI indi iduals when de e mining WTP o PHS
se ice imp o emen s.
4. Discussion
In his s udy, we conside ed possible ma ginal imp o emen s o
implemen a ion in he Spanish PHS, which is decen alised by egions.
The Spanish cen al go e nmen does no p o ide heal h se ices
di ec ly o he egions (excep o Ceu a and Melilla); i s wo k is pu ely
o coo dina e be ween he di e en egions. Thus, he esponsibili y o
he heal h sys em has de ol ed o he egions, and egional go e nmen s
decide he expendi u e le el, as well as he heal h policies o be
implemen ed. In he analysis o he bes assignmen o economic e-
sou ces, i is necessa y o conside he speci ic cha ac e is ics o each AC
[8], so a na ional-le el analysis would no be use ul. Howe e , ou s udy
conside s egion-speci ic cha ac e is ics o wo ACs, emphasising which
policies o implemen a a egional le el.
We ocus he analysis on wo egions ha ha e e y di e en heal h
se ice sys ems, BC and CI. The WTP among BC and CI indi iduals we e
calcula ed by con olling o income and educa ion le el indi idual
cha ac e is ics, and signi ican di e ences we e ound be ween analysed
egions. While he income le el signi ican ly in luenced BC indi iduals’
WTP, educa ional le el did no seem o in luence hei paymen de-
cisions. In he BC, indi iduals wi h a highe income we e willing o pay
signi ican ly less o ma ginal imp o emen s in Com o and A en ion.
To in e p e his esul , we conside ed wo ideas. Fi s , hose who
conside com o and a en ion a e mo e likely o pay o p i a e heal h
insu ance, as do people wi h a high income le el [1]. Second, p io
esea ch [6,19] has shown ha , in he BC, indi iduals a e mo e likely o
pay o p i a e heal h insu ance, so we would expec ha BC indi iduals
wi h a high income ha wan imp o emen s in com o and medical
a en ion would in es in p i a e insu ance. I is hus unde s andable
ha hese indi iduals would be less willing o pay o ma ginal
Table 3
WTP es ima es and egional compa ison
LOW EDUCATION
(Uni e si y=0)
UNIVERSITY EDUCATION
(Uni e si y =1)
LOW INCOME
(HighIncome=0)
SET 1 WTP(BC) WTP(CI) p-
alue
SET 2 WTP(BC) WTP(CI) p-
alue
Specialis -6.32 -3.51 0.18 Specialis -4.41 -4.76 0.46
(Con .In .) (-11.19,-
0.83)
(-6.89,0.22) (Con .
In .)
(-7.70,-1.13) (-9.83,0.12)
Su ge y -6.09 -5.75 0.45 Su ge y -6.32 -3.79 0.03
(Con .In .) (-9.70,-3.22) (-8.30,-3.61) (Con .
In .)
(-8.13,-4.62) (-5.97,-2.11)
Com o 12.59 3.17 0.03 Com o 8.54 9.11 0.45
(Con .In .) (5.09,21.87) (-2.61,9.52) (Con .
In .)
(3.00,14.86) (4.01,15.23)
A en ion 25.73 15.18 0.04 A en ion 25.39 24.06 0.40
(Con .In .) (17.13,36.63) (8.17,24.11) (Con .
In .)
(17.95,33.07) (16.96,33.32)
Den al 16.22 6.66 0.03 Den al 11.09 7.74 0.23
(Con .In .) (8.33,24.59) (0.08,13.48) (Con .
In .)
(4.21,17.23) (1.73,13.53)
HIGH INCOME
(HighIncome=1)
SET 3 WTP(BC) WTP(CI) p-
alue
SET4 WTP(BC) WTP(CI) p-
alue
Specialis -6.99 -1.06 0.14 Specialis -4.95 -3.23 0.31
(Con .In .) (-12.47,-
1.99)
(-12.04,8.85) (Con .
In .)
(-8.68,-1.15) (-8.93,1.86)
Su ge y -7.63 -6.49 0.33 Su ge y -7.57 -4.22 0.01
(Con .In .) (-11.44,-
4.49)
(-11.56,-2.81) (Con .
In .)
(-9.84,-5.53) (-6.14,-2.39)
Com o 4.90 5.58 0.47 Com o 2.56 10.71 0.04
(Con .In .) (-3.91,12.67) (-8.52,19.59) (Con .
In .)
(-2.27,7.36) (3.22,19.96)
A en ion 11.34 19.92 0.19 A en ion 14.13 26.65 0.02
(Con .In .) (3.69,19.73) (4.46,42.53) (Con .
In .)
(9.36,19.43) (16.61,38.79)
Den al 9.51 4.32 0.29 Den al 5.94 6.06 0.49
(Con .In .) (0.99,18.77) (-13.09,19.81) (Con .
In .)
(0.09,10.89) (-2.91,14.66)
Sou ce: compiled by au ho s
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
685
Table 4
E ec o indi iduals cha ac e is ics on WTP
Table 4.1: BASQUE
COUNTRY
EFFECT OF
EDUCATION
EFFECT OF
INCOME
Low
educa ion
Uni e si y
educa ion
p- alue Low
income
High
income
p-
alue
LOW INCOME Specialis -6.32 -4.41 0.26 LOW EDUCATION Specialis -6.32 -6.99 0.44
Su ge y -6.09 -6.32 0.43 Su ge y -6.09 -7.63 0.26
Com o 12.59 8.54 0.22 Com o 12.59 4.90 0.09
A en ion 25.73 25.39 0.49 A en ion 25.73 11.34 0.01
Den al 16.22 11.09 0.15 Den al 16.22 9.51 0.13
Low
educa ion
Uni e si y
educa ion
p- alue Low
income
High
income
p-
alue
HIGH INCOME Specialis -6.99 -4.95 0.28 UNIVERSITY
EDUCATION
Specialis -4.41 -4.95 0.42
Su ge y -7.63 -7.57 0.49 Su ge y -6.32 -7.57 0.19
Com o 4.90 2.56 0.31 Com o 8.54 2.56 0.07
A en ion 11.34 14.13 0.28 A en ion 25.39 14.13 0.01
Den al 9.51 5.94 0.28 Den al 11.09 5.94 0.12
Table 4.2: CANARY
ISLANDS
EFFECT OF
EDUCATION
EFFECT OF
INCOME
Low
educa ion
Uni e si y
educa ion
p- alue Low
income
High income p- alue
LOW INCOME Specialis -3.51 -4.76 0.34 LOW EDUCATION Specialis -3.51 -1.06 0.31
Su ge y -5.75 -3.79 0.10 Su ge y -5.75 -6.49 0.39
Com o 3.17 9.11 0.08 Com o 3.17 5.58 0.36
A en ion 15.18 24.06 0.06 A en ion 15.18 19.92 0.34
Den al 6.66 7.74 0.41 Den al 6.66 4.32 0.59
Low
educa ion
Uni e si y
educa ion
p- alue Low
income
High income p- alue
HIGH INCOME Specialis -1.06 -3.23 0.34 UNIVERSITY
EDUCATION
Specialis -4.76 -3.23 0.34
Su ge y -6.49 -4.22 0.17 Su ge y -3.79 -4.22 0.37
Com o 5.58 10.71 0.24 Com o 9.11 10.71 0.38
A en ion 19.92 26.65 0.25 A en ion 24.06 26.65 0.36
Den al 4.32 6.06 0.44 Den al 7.74 6.06 0.62
Sou ce: compiled by au ho s
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
686
imp o emen s in hese public heal h se ices. Abo e a ce ain income
le el, hose who can a o d p i a e heal h insu ance p e e o in es in
p i a e heal h han in he PHS. Sigüenza and Ma iel [18] simila ly ound
ha BC indi iduals wi h a high income ha e a lowe WTP o ma ginal
imp o emen s in he Specialis ca ego y; hey conjec u e ha his is
because indi iduals wi h a high income sol e hei heal h p oblems by
eso ing o p i a e heal h se ices.
Howe e , in he CI, educa ion ( a he han income) le el in luences
indi iduals’ paymen decisions. The WTP o ma ginal imp o emen s in
Com o and A en ion was highe o hose wi h uni e si y educa ion.
Gonz´
alez [6] concluded ha highly educa ed indi iduals ha e mo e
in o ma ion and g ea e capaci y o manage he knowledge, so hey ha e
highe equi emen s when demanding goods and se ices. Highly
educa ed CI indi iduals hus wan a highe quali y heal h sys em and a e
willing o pay mo e o ma ginal imp o emen s in ce ain heal h se -
ices in hei egion. Unlike he BC indi iduals, income le el does no
in luence CI indi iduals’ WTP.
All o hese esul s lead o he conclusion ha , o some g oups o
indi iduals wi h ce ain cha ac e is ics, he WTP o ma ginal im-
p o emen s in PHS se ices di e s be ween egions. Fi s , ocusing on
he g oup o indi iduals wi h low income and educa ional le els, he
WTP among BC indi iduals is signi ican ly highe han among CI in-
di iduals due o he combina ion o he wo e ec s. Low-educa ed CI
indi iduals do no alue he PHS highly, while BC indi iduals wi h a low
income highly app ecia e he PHS. The opposi e e ec is ob ained in he
g oup o indi iduals wi h high income and educa ional le els. Highly
educa ed CI indi iduals highly alue hei PHS, so he g oup o in-
di iduals wi h hese cha ac e is ics a e, in gene al, willing o pay mo e
han compa able BC indi iduals o ma ginal imp o emen s in PHS
se ices.
We analysed he di e ences in he alua ion o PHS se ices (WTP)
among indi iduals in he BC and CI. No only do di e ences exis in he
WTP be ween egions, bu he ac o s ha in luence hese alua ions
also di e be ween egions. Conside ing ha he maximisa ion o in-
di iduals’ global wel a e is an impo an objec i e o heal h policy, i
seems use ul o ha e a ool, such as he p esen analysis, ha p o ides
de ails abou indi iduals’ p e e ences. The heo y o decen alisa ion
indica es ha , om he poin o iew o e iciency, i is ad isable o
decen alise decisions and managemen o public expendi u e, espe-
cially when indi iduals’ p e e ences abou possible policies di e by
egion [4,12,13]. Ou esul s e i y his s a emen , because i does
appea ha indi iduals’ u ili y di e s be ween egions. I he main aim
o policymake s is o p o ide heal h se ices ha maximise indi iduals’
u ili y – ha is, global wel a e – hen he decisions abou and manage-
men o hese policies should be ans e ed o egions. Howe e , i he e
is a sho age o esou ces o ca y ou imp o emen policies in any o he
egions, hen decen alisa ion may cause egional equi y p oblems. In
his case, he cen al go e nmen should implemen income edis ibu-
ion policies o gua an ee public heal h se ices in he poo es egions.
Al hough he decisions a e decen alised o egional go e nmen s, i
is also likely ha indi iduals in he same egion ha e he e ogeneous
p e e ences, so he p e e ences o all indi iduals canno be ully sa is-
ied. In his case, he egional go e nmen should implemen heal h
policies ha maximise he o e all p e e ences o he egion, hus
ob aining an e icien alloca ion. Decen alisa ion o heal h policies o
he local le el, and e en o he pe sonal le el, would no be cos -
e ec i e, because i would no be possible o ake ad an age o he
bene i s o economies o scale.
To sum up, when heal h policies a e decen alised o he egions
while he cen al go e nmen con ols and gua an ees he edis ibu ion
o easons o equi y, i would be possible o achie e a balance be ween
he e iciency bene i s o decen alisa ion and he ad an ages in e ms o
economies o scale ha can be ound in managemen , as well as a oiding
possible equi y p oblems. The esul s o his s udy could also be use ul
o egional go e nmen s o iden i y and e lec on he peculia i ies o
hei egions. We ha e obse ed ha , wi hin each egion, he e a e
p e e ences o di e en se ices, and he iden i ica ion o hese p e -
e ences could guide egional go e nmen s in he dis ibu ion o eco-
nomic esou ces ha would inc ease global wel a e.
5. Conclusions
One o he main p oblems ha public ins i u ions ace is how o gi e
an economic alua ion o indi iduals’ p e e ences abou di e en
possible policies and hus de e mine which policies maximise in-
di iduals’ wel a e. When a coun y is di ided in o egions (wi h a cen-
al go e nmen and egional go e nmen s o each egion), i he
p e e ences o possible policies di e be ween egions, i would (based
on he heo y o decen alisa ion) be mo e e icien o decen alise
expendi u e decisions o egional go e nmen s.
This s udy analysed whe he he e a e di e ences in heal h p e e -
ences among indi iduals om wo ACs wi h di e en cha ac e is ics –
BC and CI – because i he e a e, we would be able o conclude ha
heal h policies and Spanish PHS should be decen alised in o egions.
We he e o e sough o iden i y and quan i y indi iduals’ p e e ences in
di e en egions. DCEs allowed us o measu e he WTP o indi iduals o
ma ginal imp o emen s in heal h se ices, so we could gi e an economic
alua ion o di e en heal h policies. I he economic alua ion o he
possible heal h policies di e s among egions, egional p e e ences
ega ding possible heal h policies also di e .
We he e o e calcula ed he WTP o BC and CI indi iduals o
di e en possible ma ginal imp o emen s in PHS se ice, and we
iden i ied di e ences be ween indi iduals’ heal h p e e ence in he wo
egions. The income e ec has a highe e ec in he BC, while in-
di iduals’ decisions a e mo e likely o be de e mined by educa ional
le el in he CI. We hus obse e ha he highe he income o a BC
ci izen, he lowe his o he alua ion o he PHS. This esul shows he
p e e ence ha BC indi iduals wi h high income ha e o p i a e heal h
se ices. CI indi iduals’ WTP, on he o he hand, is highe when hey
ha e comple ed uni e si y s udies. We conclude ha a highe le el o
exigency among ained people could be a eason o his esul .
Taking in o accoun ha in each egion he e a e di e en p e e -
ences o heal h policies, heal h policies should be decen alised o
egional go e nmen s while he cen al go e nmen con ols o equi y
easons, i maximising he wel a e o indi iduals is a p ima y aim. The e
a e also di e ences be ween he p e e ences o indi iduals in he same
egion, howe e , so once heal h policies a e decen alised, we ha e
p o ided a ool ha makes i possible o iden i y he heal h policies ha
would maximise he wel a e o he indi iduals in each egion. Ou
esea ch could hus allow egional go e nmen s o unde s and hei
ci izens’ heal h se ice p e e ences, which could be help ul in alloca ing
economic esou ces o maximise he global wel a e o he egion.
Decla a ion o Compe ing In e es
The au ho s decla e ha hey ha e no known compe ing inancial
in e es s o pe sonal ela ionships ha could ha e appea ed o in luence
he wo k epo ed in his pape .
Acknowledgemen s
We acknowledge inancial suppo om FEDER/Minis y o Science,
Inno a ion and Uni e si ies h ough g an PID2020-113650RB-I00. We
he eby wan o hank suppo om Mai ena Poelemans and Ge aldine
Bachoue o he Cen e de Documen a ion e de Reche ches Eu op´
eennes
(Uni e si ´
e de Pau e des Pays de l’Adou ), he Uni e si y o he Basque
Coun y (UPV/EHU), he Minis y o Uni e si ies and he Eu opean
Union h ough g an RECUALI21/18 o he e-quali ica ion o he
Spanish uni e si y sys em o 2021-2023, unded by he Eu opean
Union-Nex Gene a ion EU. We would also like o acknowledge inancial
suppo om he UPV/EHU h ough g an GIU 18/147 and om he
Basque Go e nmen h ough g an IT1336-19.
W. Sigüenza and A. A abe
Heal h policy 126 (2022) 680–687
687
Supplemen a y ma e ials
Supplemen a y ma e ial associa ed wi h his a icle can be ound, in
he online e sion, a doi:10.1016/j.heal hpol.2022.04.010.
Re e ences
[1] A abe A, Sigüenza W. The e ec s o he economic ecession on spending on p i a e
heal h insu ance in Spain. In e na ional Jou nal o Heal h Economics and
Managemen 2019;19(2):155–91.
[2] Ba eman IJ, Ca son RT, Day BH, Hanemann WM, Hanley N, He T, e al. Economic
alua ion wi h s a ed p e e ences echniques: A manual. Chel enham: Edwa d
Elga ; 2002.
[3] Cos a-Fon J. Pa icipaci´
on colec i a y e elaci´
on de p e e encias sob e p og amas
sani a ios: un en oque de sis ema sani a io. Gace a Sani a ia 2005;19(3):242–52.
[4] C eme J, Es ache A, Seab igh P. Decen alizing public se ices: Wha can we
lea n om he heo y o he i m? Re ue d´Economie Poli ique 1996;106:37–60.
[5] Ga cía-Al ´
es A, Na as E, So iano MJ. E aluaci´
on econ´
omica de in e enciones de
salud pública. Gace a Sani a ia 2011;25(Supl 1):25–31.
[6] Gonz´
alez Y. La demanda de segu os sani a ios. Re is a de Economía Aplicada
1995;8:111–42. III.
[7] Ho owi z JL. The boo s ap in econome ics. E. Leame , J.J. Heckman (eds..
Handbook o econome ics, 5. Ams e dam: No h Holland; 2001. p. 3159–228.
[8] L´
opez-Casasno as G, Cos a-Fon J, Planas I. Di e si y and egional inequali ies in
he Spanish ‘sys em o heal h ca e se ices. Heal h Economics 2005;14(S1):
S221–35.
[9] L´
opez-Casasno as G, Soley M. The economic c isis and i s e ec s on he social
de e minan s o heal h, 201. Hacienda Pública Espa˜
nola /Re is a de Economía
Pública; 2012. p. 113–32.
[10] Lou ie e J, Henshe D, Swai J. S a ed choice me hods: Analysis and applica ions.
Camb idge: Camb idge Uni e si y P ess; 2000.
[11] McFadden D. Condi ional logi analysis o quali a i e choice beha io . P.
Za embka (ed.. F on ie s in econome ics. New Yo k: Academic P ess; 1974.
p. 105–42.
[12] Oa es WE. Fiscal ede alism. New Yo k: Ha cou B ace Jo ano ich; 1972.
[13] Oa es WE. An essay on iscal ede alism. Jou nal o Economic Li e a u e 1999;37
(3):1120–49.
[14] Pei ´
o S, A ells JJ, Meneub R. Iden i icaci´
on y p io izaci´
on de ac uaciones de
mejo a de la e iciencia en el Sis ema Nacional de Salud. Gace a Sani a ia 2011;25
(2):95–105.
[15] Pin o JL. Calidad de ida y asignaci´
on de ecu sos sani a ios. Gace a Sani a ia
2000;14(2):168–74.
[16] Poe GL, Gi aud KL, Loomis JB. Compu a ional me hods o measu ing he
di e ence o empi ical dis ibu ions. Ame ican Jou nal o Ag icul u al Economics
2005;87:353–65.
[17] Poe GL, Welsh MP, Champ PA. Measu ing he di e ence in mean willingness o pay
when dicho omous choice con ingen alua ion esponses a e no independen .
Land Economics 1997;73:255–67.
[18] Sigüenza W, Ma iel P. Valo aci´
on econ´
omica de los se icios sani a ios en la
Comunidad Au ´
onoma del País Vasco. Hacienda Pública Espa˜
nola/Re iew o
Public Economics 2013;207:71–99.
[19] Sigüenza W, Mai el P. De e minan es del segu o m´
edico p i ado y del uso de
u gencias. Re is a de Economía Aplicada 2015;23(68):111–42.
[20] Viney R, Lancsa E, Lou ie e J. Disc e e choice expe imen s o measu e consume
p e e ences o heal h and heal hca e. Expe Re iew o Pha macoeconomics &
Ou comes Resea ch 2002;2(4):319–26.
W. Sigüenza and A. A abe
