Regional variation in the utilization of nursing home care in Germany

He , Annika; Lückemann, Maximilian; Sa ic-Babin, Amela
A icle — Published Ve sion
Regional a ia ion in he u iliza ion o nu sing home ca e
in Ge many
The Eu opean Jou nal o Heal h Economics
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: He , Annika; Lückemann, Maximilian; Sa ic-Babin, Amela (2024) : Regional
a ia ion in he u iliza ion o nu sing home ca e in Ge many, The Eu opean Jou nal o Heal h
Economics, ISSN 1618-7601, Sp inge , Be lin, Heidelbe g, Vol. 26, Iss. 5, pp. 757-776,
h ps://doi.o g/10.1007/s10198-024-01732-9
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/323293
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
The Eu opean Jou nal o Heal h Economics (2025) 26:757–776
h ps://doi.o g/10.1007/s10198-024-01732-9
ORIGINAL PAPER
Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
AnnikaHe 1,2,3 · MaximilianLückemann1,2 · AmelaSa ic‑Babin4
Recei ed: 11 Janua y 2024 / Accep ed: 16 Oc obe 2024 / Published online: 25 No embe 2024
© The Au ho (s) 2024
Abs ac
App oxima ely 32 pe cen o indi iduals aged o e 64yea s old, wi h ca e needs, a e esiding in nu sing homes in Ge many.
Howe e , his pe cen age exhibi s signi ican egional dispa i ies, anging om unde 15 pe cen in ce ain coun ies o o e
50 pe cen in o he s. The pu pose o his s udy is o elucida e he unde lying ac o s explaining his egional a ia ion in
nu sing home u iliza ion. We employed comp ehensi e adminis a i e da a encompassing he en i e elde ly ca e-dependen
popula ion and all nu sing homes. Ou analy ical app oach in ol es he use o linea eg ession models a he coun y le el,
accoun ing o an ex ensi e a ay o con ol a iables and ixed e ec s. Addi ionally, we analyzed egional dependencies by
applying spa ial lag models. In summa y, ou model success ully p edic s up o 73 pe cen o he obse ed egional a ia ion
in nu sing home u iliza ion. Key ac o s include ca e needs, he p esence o in o mal ca e suppo and he supply o p o es-
sional ca e. Spa ial dependencies can be de ec ed bu exhibi a mino in luence on hese a ia ions con olling o ca e needs.
No ewo hy, enabling ac o s, such as a egion’s weal h o u ali y, ha e a e y limi ed impac in a coun y wi h a gene ous
social insu ance sys em ha co e s ca e o hose wi h limi ed inancial esou ces.
Keywo ds Regional a ia ion· Long- e m ca e· Spa ial panel da a models· Nu sing home
JEL Classi ica ion I11· I18· C23
In oduc ion
When i comes o choosing he bes ype o ca e, in many
coun ies indi iduals wi h ca e needs ha e h ee op ions: (i)
in o mal ca e by amily membe s, o (combined wi h) (ii)
o mal ca e by home ca e p o ide s in hei homes, o (iii)
nu sing home ca e (NHC). In Ge many, home ca e is p io -
i ized o e NHC as pe he long- e m ca e (LTC) sys em’s
o ganiza ional p inciple, aiming o keep ca e ecipien s in
hei amilia en i onmen o as long as possible (SGB XI
§3). Despi e he na ional egula ion o he manda o y LTC
insu ance, he e is signi ican egional a ia ion in he u ili-
za ion o nu sing home ca e (NHC) in Ge many. The unad-
jus ed mean sha e a ies om below 15 pe cen o mo e
han 50 pe cen o all people o e 65 yea s old and in need
o ca e li ing in nu sing homes (NHs) ac oss he mo e han
400 Ge man coun ies, wi h an a e age o 32 pe cen .
Unde s anding he dispa i ies in NH u iliza ion is c u-
cial due o subs an ially highe public and p i a e expendi-
u es o NHC compa ed o home ca e o in o mal ca e. In
2021, he annual cos s o NH we e, on a e age, e21, 000
pe pe son compa ed o e9, 400 o home ca e [1]. Fu he -
mo e, since demand o LTC is inc easing, unde s anding
he egional di e ences is essen ial o shaping u u e LTC
policies.
Nume ous pape s ha e explo ed egional he e ogene-
i y in heal hca e u iliza ion. One p ominen example is
Finkels ein e al. [2], who explo e pa ien mig a ion and
ind ha a signi ican p opo ion o he a ia ion in heal h-
ca e u iliza ion can be a ibu ed o demand ac o s such
as p e e ences and heal h, besides he majo impac o he
place-speci ic supply ac o s ha ha e been p e iously
* Annika He
annika.he [email p o ec ed] e .de
Maximilian Lückemann
maximilian.luec[email p o ec ed] e .de
Amela Sa ic-Babin
[email p o ec ed]
1 Ins i u e o Heal h Economics (IHE), Leibniz Uni e si ä
Hanno e , CHERH Und CINCH, Königswo he Pla z 1,
30167Hanno e , Ge many
2 CINCH, Uni e si ä Duisbu g-Essen, Duisbu g, Ge many
3 Düsseldo Ins i u e o Compe i ion Economics (DICE), Hei
n ich-Heine-Uni e si ä , Düsseldo , Ge many
4 F ank u , Ge many
758 A.He e al.
iden i ied. Cu le e al. [3] a gue ha in he physician
ma ke , supply-side cha ac e is ics play a mo e subs an-
ial ole compa ed o demand in explaining he egional
he e ogenei y in expendi u es in he US. Godøy and Hui -
eld [4] highligh he in luence o socioeconomic ac o s
in explaining egional a ia ion in heal hca e u iliza ion
and mo ali y in No way. Be ge and Czypionka [5] show
ha demand-side ac o s explain mos o he a ia ions
in magne ic esonance imaging ac oss medical p ac ices.
Reich e al. [6] explo e dispa i ies o egional heal hca e
expendi u es in Swi ze land and p esen e idence o i s
co ela ion wi h supply-side densi ies and socioeconomic
ac o s.
Rega ding Ge many, Augu zky e al. [7] examine he
egional di e ences in he u iliza ion o hospi als, Kope sch
and Schmi z [8] analyze he a ia ion in he usage o ambu-
la o y physician se ices, and Göpp ah e al. [9] ocus on
o al heal hca e expendi u es. Ozegowski and Sundmache
[10] analyze he disc epancy be ween egional needs and
he u iliza ion o ou pa ien ca e. They iden i y supply ac-
o s as he p ima y con ibu o s o his gap. Some o he
s udies apply spa ial au o eg essi e models and show ha
co ela ions a e signi ican bu small ac oss coun ies. In
con as , Felde and Tauchmann [11] gene a e dis ic -le el
e iciency sco es in heal h p oduc ion and highligh he
impo ance o accoun ing o spa ial dependence o analyze
he s ong e ec o ede al s a e-speci ic egula ions on he
dis ic ’s e iciency. All s udies, as is common in his li -
e a u e, p esen co ela ions be ween he ou comes and he
egional explana o y ac o s a he han causal s a emen s.
One p ominen excep ion is Salm and Wu¨bke [12] who
u ilize indi idual exogenous pa ien mig a ion as an ins u-
men o in es iga e egional di e ences in ambula o y ca e
u iliza ion. They iden i y demog aphics and pa ien cha ac-
e is ics as being mos ele an , in con as o ins i u ional
di e ences.
In e ms o LTC p o ision, Pilny and S oka [13] exam-
ine how he egional a ailabili y o NHs a ec s elde ly ca e
decisions using a disc e e choice se ing whe e indi iduals
can choose be ween ou di e en ypes o o mal and in o -
mal ca e. Thei s udy employs esiden -le el adminis a i e
da a ob ained om a la ge Ge man heal h insu ance. They
ind ha he decision o choose NHC is signi ican ly d i en
by he egional supply o NH beds. Mennicken e al. [14]
ocus on he di e ences in emune a ion a es among NHs
in No h Rhine-Wes phalia. They ind ha app oxima ely 70
pe cen o he egional p ice di e ences can be explained,
he la ges pa by egional nego ia ion s yles be ween NHs
and he di e en LTC insu ance p o ide s. Las ly, Duell
e al. [15] examine he egional a ia ion in he eligibili y
o publicly inanced home ca e in he Ne he lands, which is
mainly d i en by he place o esidence a he han pa ien
expe iences.
Ou s udy is he i s o analyze spa ial a ia ions in NHC
u iliza ion. We con ibu e o he exis ing li e a u e by in e-
g a ing he heal hca e se ices u iliza ion model p oposed
by Ande sen and Newman [16]. The model comp ises h ee
ca ego ies: indi idual de e minan s (needs, p edisposing
ac o s, and enabling ac o s), heal h se ices supply, and
socie al de e minan s. We adjus he model o he Ge man
LTC se ing. We dis inguish be ween indi idual ca e needs,
he exis ence o in o mal ca e (p edisposing), weal h- ela ed
ac o s (enabling), supply measu es including p ices, and
socie al de e minan s.
We use comp ehensi e da a combining he Ge man Ca e
S a is ic wi h egional socioeconomic and demog aphic da a
a he coun y le el om 2007 o 2019 (biannually). The Ca e
S a is ic o he Ge man S a is ical O ices o he Lände
comp ises he en i e Ge man ca e-dependen popula ion and
all ca e acili ies. Ou me hodological app oach ollows he
small a ea a ia ion s udies by Cu le and Sheine [17] on
heal hca e expendi u es, success ully applied o o he Ge -
man heal hca e ma ke s by Augu zky e al. [7], Kope sch
1. ca egi ing needs
e.g., ca e le els, age
2. p edisposing ac o s
e.g., possibili ies o in . ca e
3. enabling ac o s
e.g., weal h, p ices
2.1 Indi idual de e minan s
e.g., pe sonnel, equipmen ,
ma e ials, geog aphic
dis ibu ion o esou ces
and p ocedu es
2.2 Heal h se ices supply
e.g., echnology a ailable
o physicians
and beha io al no ms
2.3 Socie al de e minan s
Fig. 1 De e minan s o he Ande sen-Newman model o heal hca e u iliza ion Illus a ion o he h ee indica o g oups de i ed om he
Ande sen-Newman model o heal hca e u iliza ion: (1) indi idual de e minan s, (2) heal h se ices supply, and (3) socie al de e minan s
759Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
and Schmi z [8], and Ozegowski and Sundmache [10]. We
es ima e o dina y leas squa es (OLS) models wi h egional-
ixed e ec s and wa e- ixed e ec s o unco e he a ia ion
o ou ou come a iable. To accoun o spa ial dependen-
cies in u iliza ion and egional shock spillo e s inspi ed by
Gup a e al. [18] and Ozegowski and Sundmache [10], we
addi ionally apply spa ial au o eg essi e (lag) models.
O e all, ou model achie es a good measu e o i
explaining 73 pe cen o he a ia ion in NH u iliza ion
ac oss he mo e han 400 Ge man coun ies, o which almos
64 pe cen can be a ibu ed o ca e needs (explaining 47
pe cen o he a ia ion). This is mo e han p e ious s ud-
ies ha e ound o o he ma ke s. Regional p edisposing
indica o s, such as in o mal ca e oppo uni ies measu ed
by he emale wo k o ce and demog aphics, accoun o an
addi ional 8–12 pe cen age poin s (depending on he o de
o inclusion). Regional enabling ac o s, including weal h
and u ali y, make a e y small con ibu ion, explaining an
addi ional 1–2 pe cen age poin s o he a ia ion. The supply
o heal hca e adds 5–10 pe cen age poin s when con olling
o ca e needs, which is much less han in p e ious s udies.
Regional and wa e- ixed e ec s inally explain 7 pe cen age
poin s o he a ia ion (added a las o he model). Fu he -
mo e, we iden i y a small ye signi ican p esence o spa ial
dependencies ha do no al e he main conclusions.
The indings ha e impo an policy implica ions since
hey demons a e ha besides he ca e needs, he deg ee o
in o mal ca e suppo co ela es highly wi h he demand o
NHC. Thus, policymake s could ei he s imula e in o mal
ca egi ing in some a eas by educing he double bu den o
wo k and ca ing o , in u n, inc ease he a ailabili y o NHC
o educe he need o in o mal ca e p o ision. I is also
impo an o no e ha weal h and income canno explain he
a ia ion e en when no con olling o he supply o o mal
ca e ha may be co ela ed.
Sec ion "A model o heal h se ices u iliza ion" discusses
he model o heal h se ices u iliza ion we apply o his se up,
ollowed by Sec ion "Da a and desc ip i e s a is ics”, which
p esen s he da a and desc ip i e s a is ics. We in oduce he
es ima ion s a egy in Sec ion "Es ima ion s a egy” and hen
p esen ou esul s in Sec ion "Resul s”. In Sec ion "Discus-
sion and conclusion”, we discuss ou indings and conclude.
A model o heal h se ices u iliza ion
We examine he egional dispa i ies in he u iliza ion o
NHC by employing a modi ied e sion o he Ande sen-
Newman model o heal hca e u iliza ion [16]. The model
dis inguishes be ween Indi idual de e minan s o u iliza ion
indi idual de e minan s, Heal h se ices supply supply o
LTC se ices, and Socie al de e minan s socie al de e mi-
nan s o u iliza ion. All pa ame e s a e agg ega ed a he
coun y le el. The indi idual cha ac e is ics om he ca e
s a is ics a e based on all elde ly indi iduals (aged 65 o
olde ) ecei ing any con ibu ion om he LTC insu ance
sys em (ac oss he h ee ca e ypes i no indica ed di e -
en ly). Fo an o e iew, all a iables a e g ouped and de ined
b ie ly, including sou ces, in Table5in he appendix.
Indi idual de e minan s o u iliza ion
As indi idual de e minan s o u ilizing heal hca e se ices
a e mul i- ace ed, we speci y h ee g oups o pe sonal cha -
ac e is ics: ca e needs, p edisposing ac o s, and enabling
ac o s.
Ca e needs co espond o indi idual disabili ies o heal h
s a us. Wi h ega d o ca e needs, we include he a e age
coun y-le el ca e le el i co esponds o. Ge man egula-
ions dis inguish be ween h ee le els o LTC se e i y in ou
da a.1 The need o ca egi ing a ises i indi iduals equi e
assis ance wi h he daily li ing ac i i ies due o ad anced
age, heal h- o men al- ela ed p oblems o a minimum o six
mon hs (SGB XI §14), which is assessed by ex e nal bod-
ies o he manda o y LTC insu ance, he so-called Medical
Re iew Boa ds (Medizinische Diens de K ankenkassen
[MDK] in Ge man). They also se ules o nego ia e p ices
be ween he p o ide s and he s a u o y LTC insu ances and
communi ies and moni o quali y on behal o he heal h
insu ance p o ide s (Fig.1). They ope a e a he ede al
s a e le el excep o No h Rhine-Wes phalia (sepa a ed
in o No h Rhine and Wes phalia-Lippe), Hambu g (com-
bined wi h Schleswig-Hols ein), and Be lin (combined wi h
B andenbu g).2 People who ha e been assigned a ca e le el
ecei e inancial suppo om hei LTC insu ance (see
espec i e allowances in Table6in he appendix). Le el 1
indica es mode a e needs while highe le els co espond o
mo e se e e heal h p oblems. Ca e ecipien s in ca e le el
3 a e disp opo iona ely highly ep esen ed in NHs [19].
Age co ela es wi h he eme gence o ail ies and he e-
o e se es as a p oxy o heal hca e needs and pa e ns o
medical ca e use [16]. We spli he elde ly ca e ecipien s
(ac oss all ca e ypes) in o h ee g oups: 65–74, 75–84 and
85 and o e , whe e he lowes age g oup is he e e ence
ca ego y. This ollows he idea ha olde ca e ecipien s a e
mo e likely o use NHC han hei younge pee s [19, 20].
P edisposing cha ac e is ics cap u e he su ound-
ing socio-demog aphic condi ions. We employ se e al
1 In 2017, wo le els we e added, which we map o he o me sys-
em he ollowing way: he new ca e g ades 1 and 2 co espond o
le el 1, ca e g ade 3 o le el 2, and ca e g ades 4 and 5 o le el 3.
2 The MRB is an independen non-p o i o ganiza ion p o iding
socio-medical specialis ad ice o he Ge man S a u o y Heal h Ca e
and LTC Insu ance. An o e iew o ede al s a e LTC egula ions is
a ailable a h p:// www. bi a. de/ gese ze/ laend e - heimg ese ze/, las
accessed in July 2024.
760 A.He e al.
Table 1 Desc ip i e s a is ics
We epo desc ip i e s a is ics o all Ge man elde ly ca e-dependen indi iduals (aged 65+) en i led o
public and p i a e long- e m ca e insu ance allowances agg ega ed a he coun y le el o he yea s 2007,
2009, 2011, 2013, 2015, 2017
Sou ces: Resea ch Da a Cen e (RDC) o he Fede al S a is ical O ice and S a is ical O ices o he Lände ,
Ca e S a is ic, 2007 and 2017 (DOI: h ps:// doi. o g/ 10. 21242/ 22411. 2007. 00. 02.1. 1.0–h ps:// doi. o g/ 10.
21242/ 22411. 2017. 00. 02.1. 1.0) [32]; INKAR da abase o he Fede al O ice o Regional Planning [26];
IAB adminis a i e acancy in o ma ion om he Resea ch Ins i u e o he Fede al Employmen Agency
[35]; T anspa ency epo ca ds om he BKK compa ison engine o s a iona y LTC h ps://p lege inde .
bkk-dach e band.de/. Own calcula ions
*Labo o ce pa icipa ion is a sha e o males/ emales in he age g oup 15–65 in he ac i e wo k o ce
**Ru ali y is de ined as a sha e o coun y’s popula ion li ing in municipali ies wi h less han 150 esiden s
pe km
2
***OOP e e s o a p ice nego ia ed o each nu sing home ne o he allowance paid by he manda o y
LTC insu ance
****Vacancy a io is he a io o open posi ions o he o al numbe o budge ed posi ions
***** Nu sing ca e quali y is an indica o based on he quali y epo ca ds o he Medical Re iew Boa ds
(MRB), be ween ze o ( e y poo quali y) and one (excellen quali y) (see A1 o de ails in he Appendix).
2007-2017
Mean sd p1 p99
Nu sing home u iliza ion
 Elde ly ca e ecipien s in nu sing homes [%] 31.89 7.10 17.93 49.97
Ca egi ing needs
 Elde ly ca e ecipien s ac oss all ca e ypes
 Sha e in ca e le el 2 35.06 5.60 27.17 50.81
 Sha e in ca e le el 3 11.15 3.04 5.06 19.40
 Sha e aged 75-84 43.69 3.89 33.74 52.55
 Sha e aged 85+ 43.69 3.89 33.74 52.55
P edisposing
 Labo o ce pa icipa ion* [%]
 Female 76.69 4.90 67.00 85.25
 Male 84.59 4.26 72.30 92.60
 Inhabi an s by age g oup [%]
 Age g oup 35–49 27.17 2.19 22.60 32.40
 Age g oup 50–64 21.34 2.54 16.75 27.60
 Age g oup 65–74 11.08 1.42 8.20 14.80
 Age g oup 75–84 7.68 1.32 5.30 11.50
 Age g oup
≥
85 2.56 0.46 1.60 3.70
 A g. li e expec ancy a age 60[yea s] 83.41 0.678 81.97 84.98
Enabling
 GDP pe capi a [1000 EUR] 31.92 14.24 15.67 89.87
 Use o social assis ance o elde ly [%] 19.28 12.67 3.80 61.80
 A g. pension [EUR] 833.83 84.02 659.50 1036.75
 Mon hly a ailable household income [EUR] 1681.77 245.76 1228 2419
 Sha e u ali y** 0.29 0.29 0.00 1.00
 Communal deb s pe capi a [EUR] 1639.02 1365.52 0.00 6582.45
LTC supply
 Hospi al beds [pe 10,000 inh.] 6.44 3.84 0 19.29
 Home ca e acili ies [#] 34.49 41.26 7.00 180.00
 Beds pe esiden s (occupancy a e) in NH 0.89 0.06 0.74 0.99
 Single- oom sha e in NH [%] 61.12 11.06 36.02 85.18
 Nu se acancy a io**** 0.035 0.02 0.01 0.11
 Pe sonnel pe esiden , sha e 0.62 0.09 0.39 0.82
 OOP*** ca e le el 1 [EUR] 1436.82 316.40 719.29 2289.61
 OOP*** ca e le el 2 [EUR] 1572.76 334.09 809.01 2295.09
 OOP*** ca e le el 3 [EUR] 1788.19 368.65 1015.41 2535.70
Obse a ions 2,221

761Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
a iables as p oxies o he exis ence o in o mal ca e sup-
po : he sha es o men and women in he ac i e wo k o ce,
he egional demog aphics cap u ed by he sha e o inhab-
i an s in each age coho , each wi h a di e en p obabil-
i y o se ing as in o mal ca egi e s, and he a e age li e
expec ancy a age 60. Labo o ce pa icipa ion is de ined
as he sha e o men (o women) aged be ween 15 and 65
yea s in he ac i e wo k o ce. The e ec o employmen on
he choice o he ype o ca e is ambiguous. Employmen
educes he capaci y o p o ide in o mal ca e; ye lexible
wo k a angemen s (pa - ime, mini-jobs) can help econ-
cile wo k and ca e du ies [21]. The p opensi y o daugh e s
o p o ide in o mal ca e is gene ally highe han ha o
sons [22–25]. We expec ha a highe sha e o women
in he ac i e wo k o ce is associa ed wi h highe use o
NHC, while he e ec o men’s pa icipa ion is no clea a
p io i. Fo example, women could pa ly subs i u e wo k
wi h ca egi ing, while hei spouses compensa e o he
ea nings o egone by wo king mo e.
A ound 48 pe cen o women in he ac i e wo k o ce in
2019 we e pa - ime employed, while he co esponding
sha e o men was only 11.2 pe cen [26]. We include he
sha e o he popula ion aged below 34 as e e ence ca ego y
and add 35 o 49, 50 o 64, 65 o 74, 75 o 84, and 85 and
olde . The majo i y o ca egi ing ela i es (child en and
hei spouses) all in o he age span o 50–64. Augu zky
e al. [7] sugges ha mos in o mal ca egi e s a e be ween
55 and 69 yea s o age, which we app oxima e wi h he
a ailable da a (50–65 yea s old). We also include li e expec-
ancy a age 60 and pos ula e ha highe li e expec ancy is
associa ed wi h be e heal h in olde ages, esul ing in lowe
demand o NHC. Howe e , ha ing access o NHC migh
also inc ease li e expec ancy o he oldes ca e ecipien s.
Enabling cha ac e is ics p edominan ly co e he p o-
pensi y o he a o dabili y and accessibili y o NHC. He e,
we include GDP pe capi a, household income, a e age pen-
sion o senio s, u ali y, he u iliza ion o social aids o
he elde ly, and communal deb s. Ru ali y se es as a p oxy
o a el imes as dis ance in luences acili y choices [7,
27–29]. Ru ali y is measu ed as he sha e o he coun y’s
popula ion li ing in municipali ies wi h ewe han 150
inhabi an s pe squa e kilome e . We pos ula e ha NHC in
Ge many is u ilized mo e equen ly in u ban han in u al
a eas. NHC is conside ed o be a no mal good, meaning ha
demand is nega i ely a ec ed by highe p ices and posi i ely
by income and weal h. The e o e, we include a ious meas-
u es o income and impo e ishmen , such as he amoun o
social aid o he elde ly and communal deb s. Li ing in an
NH is ela i ely expensi e, wi h an a e age co-paymen o
e1, 620 pe mon h (Table1). In line wi h his, Bakx e al.
[30] epo ha being in he bo om income qua ile in Ge -
many dec eases he p obabili y o using any o mal LTC.
Simila ly, Augu zky e al. [4] and Eibich and Zieba h [31]
ind ha a highe income posi i ely impac s he u iliza ion
o hospi al and physician se ices.
Heal h se ices supply
Heal h se ices supply encompasses he esou ces and
o ganiza ion o heal hca e deli e y. This ca ego y includes
pe sonnel, equipmen , ma e ials, esou ces, and p ocedu es
employed once a pe son in need o ca e becomes pa o he
LTC sys em. To app oxima e he egional supply o heal h-
ca e se ices, we conside a ious p o ide cha ac e is ics.
Fi s ly, we include he numbe o hospi al beds pe 10,000
inhabi an s, which can po en ially se e as a empo a y al e -
na i e o NHC, especially ollowing su gical in e en ions
in senio s [32]. Secondly, we pos ula e ha he a ailabili y o
home ca e may also con ibu e o delaying he u iliza ion o
NHC. To assess he supply o NHC, we include he a e age
single- oom sha e, he pe sonnel acancy a io, he occupa-
ion a io, and he pe sonnel- o- esiden a io. We include
NH p ices by ca e le el 1 o 3 as he log o he ou -o -pocke
paymen , i.e., he sha e o he p ice ha is no co e ed by
he LTC insu ance (ei he inanced p i a ely ou -o -pocke
o , i no easible, co e ed by he social insu ance sys em).
Las ly, we con ol o he NH’s quali y by including a quali y
indica o (see he appendix o he de ini ion).
Socie al de e minan s
Finally, socie al de e minan s o NH choice include he
echnology a ailable o physicians and beha io al no ms.
U iliza ion ela i e o Ge man a e age
1050-5-10
B andenbu g
Hesse
Mecklenbu g Wes e n Pome ania
Be lin
B emen
Thu ingia
Saa land
Rhineland-Pala ina e
No h Rhine-Wes phalia
Ge many
Saxony-Anhal
Lowe Saxony
Saxony
Baden-Wue embe g
Hambu g
Ba a ia
Schleswig-Hols ein
Pe cen age di e ence
Fig. 2 Nu sing home (NH) u iliza ion ac oss ede al s a es compa ed
o he Ge man a e age. Da a sou ce: S a is ical O ices o he Lände ,
Ca e S a is ic, 2007 o 2017 (DOI: h ps:// doi. o g/ 10. 21242/ 22411.
2007. 00. 02.1. 1.0–h ps:// doi. o g/ 10. 21242/ 22411. 2017. 00. 02.1. 1.0).
No e: This igu e shows a e age ede al-s a e de ia ion om ou main
dependen a iable mean o e ime om 2007 o 2017. The a e age
sha e o ca e ecipien s mo ing in o NHs is app . 31%. De ia ions a e
p o ided in pe cen age poin s
762 A.He e al.
Socie al de e minan s o u iliza ion a e gene ally no obse -
able [31]. We hypo hesize ha he egional ixed e ec s cap-
u e no only he geog aphic di e ences in echnology bu
also di e ences in beha io al no ms.
Da a anddesc ip i e s a is ics
We exploi h ee da a sou ces.
Fi s , we use he Ge man ca e s a is ics om he Fede al
S a is ical O ices and he S a is ical O ices o he Lände a
he Resea ch Da a Cen e (RDC) o Hanno e (“Ca e S a is-
ics”). The bi-annual s a is ic spans o e six wa es om 2007
o 2017, e lec ing he Ge man ca e si ua ion on Decembe
15 e e y wo yea s. The da a comp ise all ca e ecipien s
en i led o public o p i a e LTC insu ance allowances in
Ge many. This includes bo h in o mal and p o essional ca e
ecipien s, whe e he la e a e di e en ia ed be ween NHC
and home ca e. The da a also co e all NHC and home ca e
p o ide s, e.g., he numbe o a ailable places, he pe son-
nel, he ype o ca egi ing se ices o e ed, and he ees o
be paid o he ca e acili y o ca egi ing, accommoda ion,
and meals. To ocus on egional a ia ion a he han indi-
idual de e minan s o NH en y, we agg ega e he da a a
he coun y le el.
Second, we supplemen he ca e s a is ic wi h indica o s
o egional and u ban de elopmen , such as socio-econom-
ics, demog aphics, and heal h ca e supply, a he coun y le el
p o ided by he Fede al O ice o Building and Regional
Planning (INKAR) [7] (in Ge man: Indika o en und Ka en
zu Raum- und S ad en wicklung [INKAR], h ps:// www.
inka . de/.) Thi d, we add labo ma ke in o ma ion om he
Ge man Ins i u e o Employmen Resea ch, mo e speci i-
cally, he numbe o acancies o nu ses in ge ia ic ca e pe
coun y and yea [41].
In ou inal sample, we excluded all people unde 65
yea s o age in need o ca e and also all ca e acili ies ha
p o ide non-elde ly ca e (e.g., ca e o kids o psychia ic
ca e) o acili ies ha p o ide only sho - e m ca e o only
day o nigh ca e. In compliance wi h he da a secu i y ules
en o ced by he esea ch da a cen e , we agg ega ed neigh-
bo ing coun ies wi h ewe han h ee NHs pe owne ship
ype and coun y, which gi es us 374 (2007) and 367 (2017)
egions co e ing all 400 coun ies ac oss he 16 ede al s a es
(2221 obse a ions o e six wa es). The dependen a iable
is de ined as he coun y’s sha e o elde ly ca e ecipien s
aged 65 o olde li ing in NHs as opposed o li ing a home
( ecei ing only in o mal ca e o home ca e suppo ).
Figu e2 demons a es egional dispa i ies in he u iliza-
ion o NHC ac oss ede al s a es. NHC is mos equen ly
used in Schleswig-Hols ein, Ba a ia, and Hambu g. In
Schleswig-Hols ein, he p opo ion o ca e ecipien s in
NHs is consis en ly 9–10 pe cen age poin s highe han
he na ional a e age. Con e sely, B andenbu g, Hesse, and
Mecklenbu g Wes e n Pome ania all 5–6 pe cen age poin s
below he a e age.
The appendix p o ides maps ha isualize he egional
a ia ion in LTC u iliza ion ac oss ca e ypes and i s changes
o e ime. F om hese maps, we can disce n ce ain pa e ns.
Figu e4 illus a es he a ia ion in NHC ac oss coun ies in
2007 and 2017. NHC appea s o be mo e p e alen in he
no he n and sou he n egions o Ge many. Fo he wo o he
ca e ypes, he e is a highe u iliza ion o home ca e in he
no heas (Figu e5) and g ea e eliance on in o mal ca e in
he sou hwes (Figu e6).
Table1 summa izes all a iables included in ou model.
Fo de ailed a iable de ini ions compa e Table5 in he
appendix. The able shows ha he e is s ong egional a i-
a ion ac oss all a iables when looking a he lowes and
highes pe cen ile.
Es ima ion s a egy
Mul i a ia e linea eg essions oexplain a ia ion
inNHC
Ou es ima ion app oach ollows Cu le and Sheine [17],
Kope sch and Schmi z [8], and Be ge and Czypionka [5]
and allows us o ne ou he a ia ion ha is due o sys-
ema ic di e ences be ween he coun ies. We g adually add
g oups o explana o y a iables o he eg ession and in e
hei explana o y powe om he changes in goodness-o - i
measu es. Ou p e e ed o de o inclusion co esponds o
he models p esen ed in Sec ion "A model o heal h se ices
u iliza ion".
The simple linea eg ession equa ion is gi en as
whe e yc is he sha e o ca e ecipien s in NHC in coun y c
and wa e . X a ies ac oss models depending on he con-
ol a iables included in g oups. We s a wi h ca e needs
and subsequen ly include p edisposing, enabling, and supply
ac o s. Finally, in he ull model, we add wa e- ixed e ec s
τ and egional ixed e ec s R o accoun o unobse able
ac o s a he ele an adminis a ion le el , which a e he
Medical Re iew Boa ds (MRB o MDK in Ge man) since
hey se he ules o nego ia e p ices and moni o quali y on
behal o he heal h insu ance p o ide s as discussed abo e.
µc indica es he IID dis u bances e m. The s anda d e o s
a e clus e ed by egion and ime (MRB imes yea ).
Since he explana o y powe o each a iable block a -
ies based on he sequence o inclusion, we also speci y
(1)
yc
=
Xc
�
𝛽
+
R
+
𝜏
+
𝜇c
763Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
al e na i e sequences o obus ness analyses (compa e sec-
ion “Al e na i e O de ing”).
Spa ial au o‑co ela ion ac osscoun ies
As Ge man coun ies clus e in bigge egions, e.g., ede al
s a es, he e may be co ela ions in heal hca e use ac oss
coun ies. Tha is why, in a second s ep, we accoun o
con ounding spa ial dependencies by employing spa ial
eg ession models and es whe he he coe icien s o he
explana o y a iables change compa ed o he non-spa ial
linea eg essions.
In spa ial in e ac ion-based models, collec i e beha io s
and agg ega e pa e ns a e assumed o eme ge om he in e -
ac ion o agen s ac oss social, economic, and geog aphic
dimensions [35, 36]. In e ac ion can be (a) endogenous,
whe e he g oup causally in luences indi idual beha io , (b)
exogenous, whe e indi idual beha io a ies wi h exogenous
cha ac e is ics o he g oup, o (c) co ela ed, whe e simi-
la beha io is due o simila indi idual cha ac e is ics and
ins i u ional en i onmen s [37]. The unde lying idea is ha
ac ions chosen by one indi idual in luence he cons ain s,
expec a ions, and p e e ences in he e e ence g oup [38].
Endogenous in e ac ions in NHC may a ise om cul u al
ac o s. Fo example, he high use o NHC may inc ease
i s b oade socie al accep ance. In ou con ex , exogenous
in e ac ion can be a ibu ed o a wide e ec o local de el-
opmen s. Fo example, he closu e o a la ge NH in one
coun y is likely o boos he demand o NHC in neighbo ing
coun ies. Nega i e economic shocks will inc ease unemploy-
men , which could inc ease he deg ee o in o mal ca e sup-
po . Co ela ed in e ac ions esul om ac o s ha canno
be obse ed in he da a. Examples include a high p e alence
o condi ions ha o en p ecede mo ing o an NH, such as
men al diseases and s okes, o a good quali y o ca e in a
pa icula egion.
We ollow Kope sch and Schmi z [8] and Gup a e al.
[18], who analyze spa ial dependencies in o he heal hca e
ma ke s, and assume ha he use o NHC ollows a spa-
ial au o eg essi e p ocess. Equa ions (2) and (3) o mally
desc ibe he es ima ion p ocedu e o he spa ial au o eg es-
si e combined model (SAC), including a spa ially lagged
ou come and e o e m:
whe e W o igina es om he spa ial con igui y weigh
ma ix, and
𝜇c
and
𝜖c
a e ec o s o spa ially co ela ed
esiduals and IID dis u bances, whe eby
𝜖c
∼ N (0, σ2I).
We deno e
Wyj
and
Wuj
as spa ial lags o he dependen
(2)
yc
=
𝜆Wyj
+
Xc
�
𝛽
+
R
+
𝜏
+
𝜇c
(3)
uc
=
𝜌Wuj
+
𝜖c ,
(Mo an's I=0.2301 and P- alue=0.0010)
3
2
1
0
-1
-2
-2 024
NH sha e
WNHsha e
Fi ed alues
Spa ially laggedNH_sha e
Fig. 3 Mo an sca e plo . Da a sou ce: RDC o he Fede al S a is i-
cal O ice and S a is ical O ices o he Lände , Ca e S a is ic, 2007
and 2017 (DOI: h ps:// doi. o g/ 10. 21242/ 22411. 2007. 00. 02.1. 1.0–
h ps:// doi. o g/ 10. 21242/ 22411. 2017. 00. 02.1. 1.0), own calcula ions.
No es: The Mo an sca e plo displays how he selec ed a ibu e’s
alues a each loca ion ela e o he a e age alue o he same a ib-
u e a neighbo ing loca ions. The uppe - igh (lowe -le ) quad an
ep esen s cases whe e bo h he a ibu e alue and he local a e age
alue exceed (lie below) he o e all a e age alue, indica ing posi i e
spa ial au oco ela ions. The o he wo quad an s indica e nega i e
spa ial au oco ela ions. The dominan g oups wi hin hese quad an s
de e mine he o e all endency owa d posi i e, nega i e, o no spa-
ial au oco ela ions (see, e.g., lec u e no es om Penn S a e, P ojec
4: Calcula ing Global Mo an’s I and he Mo an Sca e plo , h ps://
www.e- educa ion. psu. edu/ geog5 86/ node/ 672, las accessed in July
2024)
764 A.He e al.
Table 2 Reg ession esul s on a ia ion in NHC, 2007–2017
Model 1 Model 2 Model 3 Model 4 Model 5
Elde ly ca e ecipien s in le el 2 [%] – 0.0092 – 0.0307 – 0.0504 – 0.0995 0.0003
(0.0590) (0.0477) (0.0511) (0.0658) (0.0687)
Elde ly ca e ecipien s in le el 3 [%] 0.5560*** 0.4139*** 0.3959*** 0.1967*** 0.2422***
(0.0969) (0.0513) (0.0802) (0.0729) (0.0555)
Elde ly ca e ecipien s aged 75 – 84 [%] – 1.627*** – 0.7844*** – 0.8873*** – 0.7100*** – 0.4878***
(0.1480) (0.1816) (0.1815) (0.1447) (0.1013)
Elde ly ca e ecipien s aged 85+ [%] – 0.211* 0.8343*** 0.6711*** 0.7135*** 0.7842***
(0.1201) (0.1279) (0.1297) (0.1101) (0.0926)
P edisposing
Female labo o ce pa icipa ion, sha e 0.2433*** 0.2585*** 0.1116* 0.2746***
(0.0580) (0.0593) (0.0582) (0.0591)
Male labo o ce pa icipa ion, sha e – 0.1135*** – 0.1671*** – 0.0949*** – 0.2011***
(0.0414) (0.0384) (0.0336) (0.0313)
Inhabi an s 35–49, sha e 0.2404 0.1567 0.4522*** 0.2393
(0.1736) (0.1770) (0.1401) (0.1554)
Inhabi an s aged 50–64, sha e – 0.9781*** – 0.9556*** – 0.8012*** – 0.3347***
(0.1210) (0.1260) (0.1376) (0.1032)
Inhabi an s 65–74, sha e 0.5286*** 0.3698** 0.2794 – 0.6543***
(0.1835) (0.1724) (0.1994) (0.1846)
Inhabi an s 75–84, sha e 1.919*** 1.626*** 1.395*** 1.544***
(0.3374) (0.4323) (0.3940) (0.2699)
Inhabi an s
≥85
, sha e – 4.083*** – 3.357*** – 3.256*** – 2.204***
(0.8868) (1.0100) (0.8692) (0.6460)
A g. li e expec ancy a 60, yea s – 0.0386*** – 0.0430*** – 0.0369*** – .0325***
(0.0041) (0.0041) (0.0035) (.003)
Enabling
Log(GDP pe capi a [1,000 EUR]) 0.0143** – 0.0045 0.0011
(0.0061) (0.0057) (0.0052)
Social aids o elde ly [%] – 0.0740*** – 0.0185 – 0.0059
(0.0225) (0.0239) (0.0157)
Log (A g. pension) 0.0166 0.0465 0.1589***
(0.0278) (0.0282) (0.0268)
Sha e u ali y** – 0.0165* – 0.0237*** – 0.0215***
(0.0083) (0.0072) (0.0061)
Log (A ailable household income) 0.0229 0.0429** 0.0189
(0.0177) (0.0200) (0.0174)
Log (Communal deb s p.c. [EUR]) 0.0007 0.0010 0.0002
(0.0007) (0.0010) (0.0006)
Supply
Hospi al beds [pe 10,000 inh.] 0.0022*** 0.0022***
(0.0004) (0.0003)
Single- oom sha e in NH [%] – 0.0388*** – 0.0328**
(0.0138) (0.0136)
Home ca e acili ies – 0.0002*** – 0.0002***
(0.00005) (0.00005)
Occupancy a e in NH – 0.1002*** – .1556***
(0.0317) (0.0280)
Nu se acancy a io**** – 0.4489*** – 0.3009***
(0.0663) (0.0590)
Pe sonnel pe esiden a io – 0.0848*** – 0.1263***
771Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
unadjus ed
+ ca egi ing needs
+ p edisposing
+ enabling
+ supply
+ ixed e ec s
+ spa ial co . dep. a .
+ spa ial lag and e o model
0 .5 1 1.5
Fig. 7 Obse ed- o-p edic ed a io o NHC ecipien s by qua ile o
obse ed u iliza ion. No e: We classi y he coun ies in o qua iles
based on he unadjus ed sha es o ca e ecipien s in NH. blue: low-
es 25%, ed: 25–49%, g een: 50–75%, yellow: highes qua ile. The
speci ica ions used a e as ollows: 1. Unadjus ed; 2. Adjus ed o
need; 3. Addi ionally adjus ed o p edisposing ac o s; 4. + enabling
ac o s; 5. + NH supply; 6. + MRB and wa e ixed e ec s; 7 + 8 + spa-
ial co ela ions. Da a sou ce: RDC o he Fede al S a is ical O ice
and S a is ical O ices o he Lände , Ca e S a is ic, 2007 and 2017
(DOI: h ps:// doi. o g/ 10. 21242/ 22411. 2007. 00. 02.1. 1.0–h ps:// doi.
o g/ 10. 21242/ 22411. 2017. 00. 02.1. 1.0). Own calcula ions
Fig. 8 Sha es o ca e ecipien s
in NHC and spa ial dependence.
A e age sha es o ca e ecipi-
en s in NH o 2007 and 2017
ac oss coun ies abo e o below
he na ional a e age and he
sign o i s spa ial dependence.
( – , +) u iliza ion below mean,
posi i e spa ial dependence;
( – , – ) u iliza ion below mean,
nega i e spa ial dependence;
(+ , – ) u iliza ion abo e mean,
nega i e spa ial dependence;
(+ , +) u iliza ion abo e mean,
posi i e spa ial dependence.
Da a sou ce: RDC o he
Fede al S a is ical O ice and
S a is ical O ices o he Lände ,
Ca e S a is ic, 2007 and 2017
(DOI: h ps:// doi. o g/ 10. 21242/
22411. 2007. 00. 02.1. 1.0–h ps://
doi. o g/ 10. 21242/ 22411. 2017.
00. 02.1. 1.0). Own calcula ions

772 A.He e al.
Tables
See Tables5, 6, 7
Table 5 Desc ip ion o explana o y a iables and sou ces, agg ega ed a coun y le el
Da a sou ces: Ca e s a is ic Ca e s a is ic, (DOI: h ps:// doi. o g/ 10. 21242/ 22411. 2007. 00. 02.1. 1.0–h ps:// doi. o g/ 10. 21242/ 22411. 2017. 00. 02.1.
1.0); INKAR da abase o he Fede al O ice o Building and Regional Planning (BBR); IAB Ge man Ins i u e o Employmen Resea ch; T ans-
pa ency epo s T anspa ency epo ca ds om he BKK compa ison engine o s a iona y LTC: h ps://p lege inde .bkk-dach e band.de/ [las
accessed Sep . 2024]
Desc ip ion Da asou ce
NHCu iliza ion
Elde ly ca e ecipien s in NH Sha e o people in need o ca e abo e 65 using NHC Ca e s a is ic
Ca e needs
Elde ly ca e ecipien s in ca e le el 1–3 ac oss
all ca e ypes
Sha e o people in need o ca e in ca e le els 2 and 3, compa ed o
ca e le el 1
Ca e s a is ic
Elde ly ca e ecipien s aged 75–84 (85+), sha e Sha e by age g oup, e e ence ca ego y: 65–74 aged ca e ecipien s Ca e s a is ic
P edisposing
Labo o ce pa icipa ion [%] Sha e o employed indi iduals in he coun y pe 100 inhabi an s o
wo king age (15–65)
INKAR
Inhabi an s, by age g oups, [%] Sha e o o al popula ion by age,
<34
is he e e ence ca ego y INKAR
A g. li e expec ancy a 60 [yea s] A e age li e expec ancy o a 60-yea -old pe son INKAR
Enabling
GDP pe capi a [1,000 EUR] G oss Domes ic P oduc (GDP) pe inhabi an INKAR
 Use o social aid o he elde ly [%] Sha e o he popula ion ecei ing basic elde ly suppo INKAR
A g. pension [EUR] A e age mon hly pension paymen in eu os INKAR
Mon hly household income pe capi a [EUR] A e age disposable household income in eu os pe inhabi an INKAR
Ru ali y ** [%] Sha e o a coun y’s inhabi an s li ing in municipali ies wi h less han
<150
inhab. pe km
2
INKAR
Communal deb s pe capi a [EUR] Communal deb s pe inhabi an INKAR
LTC supply
Hospi al beds [pe 10,000 inh.] Hospi al beds pe 10,000 inhabi an s INKAR
Single oom sha e in NH [%] A e age sha e o single ooms in NH Ca e s a is ic
Occupancy a e Sha e o esiden s pe NH bed Ca e s a is ic
Pe sonnel o esiden a io Sha e o ull- ime equi alen pe sonnel o all esiden s Ca e s a is ic
Home ca e acili ies [#] Numbe o home ca e p o ide s (ambula o y ca e) Ca e s a is ic
Nu se acancy a io Sha e o job openings o ge ia ic ca e nu ses o all budge ed posi-
ions
IAB
A e age OOP by ca e le el [EUR]*** Ou -o -pocke con ibu ion by ca e le el (=NH p ice−LTC insu ance
paymen as o Table6)
Ca e s a is ic
Nu sing ca e quali y Index o nu sing ca e quali y (see Equa ion4 o u he in o ma ion) T anspa ency epo s
773Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
Table 6 Long- e m ca e
insu ance unds’ allowance
We p esen he maximum mon hly allowance (eu os) paid by he Ge man public long– e m ca e insu ance
o in o mal ca e, home heal h ca e, and nu sing home ca e. Depending on hei con ibu ions, ca e ecipi-
en s wi h p i a e insu ance a e en i led o highe allowances. Sou ces: Bundesgesundhei sminis e ium;
P leges ä kungsgese ze I–III, P lege-Neuaus ich ungs-Gese z. 2007-2011 (pp. 44-45): h ps:// www. bunde
sgesu ndhei smin is e ium. de/ ilea dmin/ Da ei en/5_ Publi ka io nen/ P lege/ Be ic h e/5. P leg ebe i ch . pd , 2011–
2015 h ps:// www. bunde sgesu ndhei smin is e ium. de/ ilea dmin/ Da ei en/5_ Publi ka io nen/ P lege/ Be ic h e/6.
P leg ebe i ch . pd (pp. 116–117) and h ps:// www. bunde sgesu ndhei smin is e ium. de/ minis e ium/ meldu
ngen/ 2016/ dezem be - 2016/ neu e gelun gen- 2017. h ml o 2017 [las accessed No . 2024]
2007 2009 2011 2013 2015 2017*
In o mal ca e
 Ca e le el I 205 215 225 235 244 316
 Ca e le el II 410 420 430 440 458 545
 Ca e le el III 665 675 685 700 728 728
 Ha dship case – – – – – 901
Home heal h ca e
 Ca e le el I 384 420 440 450 468 689
 Ca e le el II 921 980 1040 1100 1144 1298
 Ca e le el III 1432 1470 1510 1550 1612 1612
 Ha dship case 1918 1918 1918 1918 1995 1995
Nu sing home ca e
 Ca e le el I 1023 1023 1023 1023 1064 770
 Ca e le el II 1279 1279 1279 1279 1330 1262
 Ca e le el III 1432 1470 1510 1550 1612 1775
 Ha dship case 1688 1750 1825 1918 1995 2005
774 A.He e al.
Table 7 Spa ial eg ession
esul s based on OLS, model 5,
2007–2017, ull esul s
∗p<0.10
,
∗∗ p<0.05
,
∗∗∗ p<0.01
. S anda d e o s supp essed o ease o p esen a ion
Speci ica ions: (1) Spa ial Lag Model (SLM), Equa ion (2), (2) Spa ial E o Model (SEM) Equa ion (3),
(3) Spa ial Au o eg essi e Combined Model (SAC), Equa ions (2) and (3).
Sou ces: Resea ch Da a Cen e (RDC) o he Fede al S a is ical O ice and S a is ical O ices o he Lände ,
Ca e S a is ics, su ey yea s 2007–2017 (DOI: h ps:// doi. o g/ 10. 21242/ 22411. 2007. 00. 02.1. 1.0–h ps://
doi. o g/ 10. 21242/ 22411. 2017. 00. 02.1. 1.0); INKAR da abase o he Fede al O ice o Building and
Regional Planning (BBR). Own calcula ions
SLM SEM SAC
Ca egi ing needs
Ca e ecipien s in ca e le el 2, sha e 0.0153 0.0114 0.0154
Ca e ecipien s in ca e le el 3, sha e 0.2468*** 0.3443*** 0.3476***
Sha e LTC ecipien s aged 75–84 – 0.4886*** – 0.4742*** – 0.4712***
Sha e LTC ecipien s aged 85+ 0.7536*** 0.7402*** 0.7225***
P edisposing
Female labo o ce pa icipa ion, sha e 0.2801*** 0.2707*** 0.2773***
Male labo o ce pa icipa ion, sha e – 0.2143*** – 0.1975*** – 0.2091***
Sha e inhabi an s
35 −49
0.1545 0.1678 0.1040
Sha e inhabi an s aged
50 −64
– 0.3499*** – 0.3886*** – 0.4043***
Sha e inhabi an s
65 −74
– 0.6831*** – 0.3948** – 0.4215**
Sha e inhabi an s
75 −84
1.4752*** 1.2300*** 1.2138***
Sha e inhabi an s
≥85
– 1.9391*** – 1.9298*** – 1.7317***
A g. li e expec ancy a 60 – 0.0316*** – 0.0317*** – 0.0315***
Enabling
Log (GDP pe capi a [1,000 EUR]) 0.0031 0.0022 0.0030
Use o social aids o elde ly – 0.0017 0.0164 0.0218
Log (A g. pension) 0.1355*** 0.1612*** 0.1507***
Sha e u ali y** – 0.0234*** – 0.0182*** – 0.0192***
Log (A ailable household income) 0.0220* 0.0205 0.0220*
Log (Communal deb s [EUR]) – 0.0001 0.0001 – 0.0001
Supply
Hospi al beds [pe 10,000 inh.] 0.0023*** 0.0019*** 0.0021***
Single oom sha e in NH – 0.0300*** – 0.0339*** – 0.0320***
Ra io ambula o y acili ies pe ca e ecipien – 0.0002*** – 0.0002*** – 0.0002***
Occupancy a e in NH – 0.1576*** – 0.1507*** – 0.1510***
Vacancy a io**** – 0.3091*** – 0.3332*** – 0.3334***
Nu ses pe esiden – 0.1187*** – 0.1249*** – 0.1195***
Log(A e age OOP*** [EUR], Ca e le el 1) 0.0967** 0.1277*** 0.1237***
Log(A e age OOP*** [EUR], Ca e le el 2) – 0.1306** – 0.1056 – 0.1053*
(0.056) (0.064) (0.064)
Log(A e age OOP*** [EUR], Ca e le el 3) – 0.0030 – 0.0553 – 0.0523
Nu sing ca e quali y***** 0.0197** 0.0168** 0.0177**
Cons an 2.3142*** 2.1161*** 2.1749***
Lambda (
𝜆
) 0.0060*** 0.0065***
(0.001) (0.002)
Rho (
𝜌
) 0.0823*** 0.0770***
(0.006) (0.006)
MRB FE, Time FE Yes Yes Yes
Obse a ions 2,221 2,221 2,221
775Regional a ia ion in heu iliza ion o nu sing home ca e inGe many
Acknowledgemen s We hank Ma in Ka lsson and And eas Schmid
as well as pa icipan s o he wo kshop DIBOGS 2017 in Munich, he
dggö con e ence 2022, and he CINCH semina a Essen o aluable
commen s. We hank U ban Janisch and Jonas Löbel om he Resea ch
Da a Cen e o he S a is ical O ices o Saxony-Anhal o p o iding us
wi h he ca e s a is ics. We would also like o exp ess ou g a i ude o
Flo ian Köhle om he Resea ch Da a Cen e a he S a is ical O ice
o Lowe Saxony o his suppo du ing ou on-si e wo k and da a p o-
cessing. Financial suppo om he Ge man Minis y o Educa ion and
Resea ch (BMBF; g an numbe s 01EH1102A and 01EH1602B) and
he Ge man Resea ch Founda ion (DFG; g an numbe HE 6825/3-1)
is g a e ully acknowledged.
Funding Open Access unding enabled and o ganized by P ojek
DEAL. Deu sche Fo schungsgemeinscha , HE 6825/3-1, Annika He ,
Bundesminis e ium ü Bildung und Fo schung, 01EH1602B, Annika
He , 01EH1102A, Annika He .
Da a a ailabili y The Ca e S a is ics a e publicly a ailable and can be
accessed ia he S a is ical O ice o he Lände unde he con ac ed
es ic ions (e.g., on-si e use and ees). DOI: h ps:// doi. o g/ 10. 21242/
22411. 2007. 00. 02.1. 1.0–h ps:// doi. o g/ 10. 21242/ 22411. 2017. 00. 02.1.
1.0.
Open Access This a icle is licensed unde a C ea i e Commons A i-
bu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a-
ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce,
p o ide a link o he C ea i e Commons licence, and indica e i changes
we e made. The images o o he hi d pa y ma e ial in his a icle a e
included in he a icle’s C ea i e Commons licence, unless indica ed
o he wise in a c edi line o he ma e ial. I ma e ial is no included in
he a icle’s C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will
need o ob ain pe mission di ec ly om he copy igh holde . To iew a
copy o his licence, isi h p://c ea i ecommons.o g/licenses/by/4.0/.
Re e ences
1. GKV-Spi zen e band: Kennzahlen de sozialen P lege e si-
che ung 2023, www. gk - spi z en e band. de/ se i ce/ zahlen_ und_
g a i ken/ sp _ kennz ahlen/ sp _ kennz ahlen. jsp, (2022)
2. Finkels ein, A., Gen zkow, M., Williams, H.: Sou ces o geo-
g aphic a ia ion in heal h ca e: e idence om Pa ien Mig a ion.
Q. J. Econ. 131, 1681–1726 (2016). h ps:// doi. o g/ 10. 1093/ qje/
qjw023
3. Cu le , D., Skinne , J.S., S e n, A.D., Wennbe g, D.: Physician
belie s and pa ien p e e ences: a new look a egional a ia ion in
heal h ca e spending. Am. Econ. J. Econ. Pol. 11, 192–221 (2019).
h ps:// doi. o g/ 10. 1257/ pol. 20150 421
4. Godøy, A., Hui eld , I.: Regional a ia ion in heal h ca e u iliza-
ion and mo ali y. J. Heal h Econ. 71, 102254 (2020). h ps:// doi.
o g/ 10. 1016/j. jheal eco. 2019. 102254
5. Be ge , M., Czypionka, T.: Regional medical p ac ice a ia ion in
high-cos heal hca e se ices. Eu . J. Heal h Econ. 22, 917–929
(2021). h ps:// doi. o g/ 10. 1007/ s10198- 021- 01298-w
6. Reich, O., Weins, C., Schus e schi z, C., Thöni, M.: Explo ing
he dispa i ies o egional heal h ca e expendi u es in Swi ze -
land: some empi ical e idence. Eu . J. Heal h Econ. 13, 193–202
(2012). h ps:// doi. o g/ 10. 1007/ s10198- 011- 0299-x
7. Augu zky, B., Kope sch, T., Schmi z, H.: Wha accoun s o he
egional di e ences in he u ilisa ion o hospi als in Ge many?
Eu . J. Heal h Econ. 14, 615–627 (2012). h ps:// doi. o g/ 10. 1007/
s10198- 012- 0407-6
8. Kope sch, T., Schmi z, H.: Regional a ia ion in he u ilisa ion
o ambula o y se ices in Ge many. Heal h Econ. 23, 1481–1492
(2014). h ps:// doi. o g/ 10. 1002/ hec. 3001
9. Göp a h, D., Kope sch, T., Schmi z, H.: De e minan s o egional
a ia ion in heal h expendi u es in Ge many. Heal h Econ. 25,
801–815 (2016). h ps:// doi. o g/ 10. 1002/ hec. 3183
10. Ozegowski, S., Sundmache , L.: Unde s anding he gap be ween
need and u iliza ion in ou pa ien ca e— he e ec o supply-side
de e minan s on egional inequi ies. Heal h Policy 114, 54–63
(2014). h ps:// doi. o g/ 10. 1016/j. heal hpol. 2013. 08. 005
11. Felde , S., Tauchmann, H.: Fede al s a e di e en ials in he e i-
ciency o heal h p oduc ion in Ge many: an a i ac o spa ial
dependence? Eu . J. Heal h Econ. 14, 21–39 (2013). h ps:// doi.
o g/ 10. 1007/ s10198- 011- 0345-8
12. Salm, M., Wübke , A.: Sou ces o egional a ia ion in heal hca e
u iliza ion in Ge many. J. Heal h Econ. 69, 102271 (2020). h ps://
doi. o g/ 10. 1016/j. jheal eco. 2019. 102271
13. Pilny, A., S oka, M.A.: De e minan s o ecei ed long- e m ca e –
Indi idual esponses o egional nu sing home p o isions. Heal h
Ca e Manag. Sci. Manag. Sci. 19, 326–337 (2016). h ps:// doi. o g/
10. 1007/ s10729- 015- 9333-3
14. Mennicken, R., Augu zky, B., Ro hgang, H., Wasem, J.:
Explaining di e ences in emune a ion a es o nu sing homes
in Ge many. Eu . J. Heal h Econ. 15, 401–410 (2014). h ps://
doi. o g/ 10. 1007/ s10198- 013- 0483-2
15. Duell, D., an Oo , R., Lindeboom, M., Koolman, X., Po ai ,
F.: Regional a ia ion in public long- e m home ca e : egional
di e ences in he Ne he lands and he ole o pa ien expe i-
ences. No dic J. Heal h Econ. 8, 87–113 (2020). h ps:// doi. o g/
10. 5617/ njhe. 6670
16. Ande sen, R., Newman, J.F.: Socie al and indi idual de e mi-
nan s o medical ca e u iliza ion in he Uni ed S a es. Milbank
Q. 83, 1–28 (2005). h ps:// doi. o g/ 10. 2307/ 33496 13
17. Cu le , D.M., Sheine , L.: The geog aphy o medica e. Am.
Econ. Re . 89, 228–233 (1999). h ps:// doi. o g/ 10. 1257/ ae .
89.2. 228
18. Gup a, D., Biswas, D., Kabi aj, P.: COVID-19 ou b eak and U ban
dynamics: egional a ia ions in India. GeoJou nal 87, 2719–2737
(2022). h ps:// doi. o g/ 10. 1007/ s10708- 021- 10394-6
19. Schulz, E.: De e minan s o ins i u ional long- e m ca e in Ge -
many. Enep i Resea ch Repo . 115, (2012)
20. Picka d, L.: Subs i u ion be ween o mal and in o mal ca e: a
`na u al expe imen ’ in social policy in B i ain be ween 1985 and
2000. Ageing Soc. 32, 1147–1175 (2011). h ps:// doi. o g/ 10. 1017/
s0144 686x1 10008 33
21. Alde s, P., Cos a-Fon , J., de Kle k, M., F ank, R.: Wha is he
impac o policy di e ences on nu sing home u iliza ion? The
cases o Ge many and he Ne he lands. Heal h Policy 119, 814–
820 (2015). h ps:// doi. o g/ 10. 1016/j. heal hpol. 2015. 02. 006
22. Cha les, K.K., Se ak, P.: Can amily ca egi ing subs i u e o
nu sing home ca e? J. Heal h Econ. 24, 1174–1190 (2005). h ps://
doi. o g/ 10. 1016/j. jheal eco. 2005. 05. 001
23. F eedman, V.A.: Family s uc u e and he isk o nu sing home
admission. J. Ge on ol.: Se . B 51B, S61–S69 (1996). h ps:// doi.
o g/ 10. 1093/ ge onb/ 51B.2. S61
24. an Hou en, C.H., No on, E.C.: In o mal ca e and heal h ca e
use o olde adul s. J. Heal h Econ. 23, 1159–1180 (2004). h ps://
doi. o g/ 10. 1016/j. jheal eco. 2004. 04. 008
25. Wol , D.A., F eedman, V., Soldo, B.J.: The di ision o amily
labo : ca e o elde ly pa en s. J. Ge on ol. B Psychol. Sci. Soc.
Sci.Ge on ol. B Psychol. Sci. Soc. Sci. 52B, 102–109 (1997).
h ps:// doi. o g/ 10. 1093/ ge onb/ 52b. speci al_ issue. 102
26. Bundesam ü Bauwesen und Raumo dnung: INKAR indika o en
und ka en zu aum- und siedlungsen wicklung, www. inka . de,
(2023)
776 A.He e al.
27. Kenney, G.M., Dubay, L.C.: Explaining a ea a ia ion in he use
o medica e home heal h se ices. Med. Ca e 30, 43–57 (1992).
h ps:// doi. o g/ 10. 1097/ 00005 650- 19920 1000- 00004
28. McAuley, W.J., Spec o , W., an Nos and, J.: Fo mal home ca e
u iliza ion pa e ns by u al–u ban communi y esidence. J. Ge -
on ol. Se . B. 64B, 258–268 (2009). h ps:// doi. o g/ 10. 1093/ ge -
onb/ gbn003
29. Schmi z, H., S oka, M.A.: Heal h and he double bu den o ull-
ime wo k and in o mal ca e p o ision — e idence om admin-
is a i e da a. Labou Econ. 24, 305–322 (2013). h ps:// doi. o g/
10. 1016/j. labeco. 2013. 09. 006
30. Bakx, P., de Meije , C., Schu , F., an Doo slae , E.: Going o -
mal o in o mal, who ca es? The in luence o public long- e m
ca e insu ance. Heal h Econ. 24, 631–643 (2015). h ps:// doi.
o g/ 10. 1002/ hec. 3050
31. Eibich, P., Zieba h, N.R.: Examining he s uc u e o spa ial
heal h e ec s in Ge many using hie a chical bayes models. Reg.
Sci. U ban Econ. 49, 305–320 (2014). h ps:// doi. o g/ 10. 1016/j.
egsc iu be co. 2014. 06. 005
32. Kümpel, C.: Do inancial incen i es in luence he hospi aliza-
ion a e o nu sing home esiden s? E idence om Ge many.
Heal h Econ. 28, 1235–1247 (2019). h ps:// doi. o g/ 10. 1002/
hec. 3930
33. S a is isches Bundesam : S a is ische be ich - p lege o aus-
be echnung - deu schland und bundeslände - 2022 bis 2070,
www. des a is. de/ DE/ Themen/ Gesel lscha - U mwel / Gesun dhei /
P leg e/_ inhal . h ml#_ 5o9 c unzo, (2020)
34. S a is ik de Bundesagen u ü A bei : Bes and an gemelde-
en a bei ss ellen nach ausgewähl en be u en und wi scha sz-
weigen, p ojec -no.: [292547], s a is ik.a bei sagen u .de/DE/,
(2023)
35. Anselin, L.: Unde he hood Issues in he speci ica ion and in e -
p e a ion o spa ial eg ession models. Ag ic. Econ. 27, 247–267
(2002). h ps:// doi. o g/ 10. 1111/j. 1574- 0862. 2002. b001 20.x
36. LeSage, J.: Spa ial econome ics. In: Handbook o esea ch me h-
ods and applica ions in economic geog aphy. Edwa d Elga Pub-
lishing (2015)
37. Manski, C.F.: Economic analysis o social in e ac ions. J. Econ.
Pe spec . 14, 115–136 (2000). h ps:// doi. o g/ 10. 1257/ jep. 14.3.
115
38. Re elli, F.: Pe o mance a ing and ya ds ick compe i ion in social
se ice p o ision. J. Public Econ. 90, 459–475 (2006). h ps:// doi.
o g/ 10. 1016/j. jpube co. 2005. 07. 006
39. Moscone, F., Knapp, M., Tose i, E.: Men al heal h expendi u e
in England: a spa ial panel app oach. J. Heal h Econ. 26, 842–864
(2007). h ps:// doi. o g/ 10. 1016/j. jheal eco. 2006. 12. 008
40. D ukke , D.M., Peng, H., P ucha, I.R., Racibo ski, R.: Maximum
likelihood and gene alized spa ial wo-s age leas -squa es es ima-
o s o a spa ial-au o eg essi e model wi h spa ial-au o eg essi e
dis u bances. S and. Genomic Sci. (2013). h ps:// doi. o g/ 10. 1177/
15368 67X13 01300 201
41. Elho s , J.P.: Spa ial econome ics. Sp inge , Be lin, Heidelbe g
(2014)
42. Woold idge, J.M.: Two-way ixed e ec s, he wo-way mundlak
eg ession, and di e ence-in-di e ences es ima o s. SSRN Elec-
on. J. (2021). h ps:// doi. o g/ 10. 2139/ ss n. 39063 45
43. Dale, M.R.T., Fo in, M.-J.: Spa ial au oco ela ion and s a is ical
es s: Some solu ions. J. Ag ic. Biol. En i on. S a . 14, 188–206
(2009)
44. Sünde kamp, S., Weiß, C., Ro hgang, H.: Analyse de ambulan en
und s a ionä en P legeno en hinsich lich de Nü zlichkei ü den
Ve b auche . P lege 27, 325–336 (2014). h ps:// doi. o g/ 10. 1024/
1012- 5302/ a0003 79
45. He , A., Ho en o , H.: Highe p ices, highe quali y? E idence
om Ge man nu sing homes. Heal h Policy 120, 179–189 (2016).
h ps:// doi. o g/ 10. 1016/j. heal hpol. 2016. 01. 008
46. He , A., Nguyen, T.-V., Schmi z, H.: Public epo ing and he
quali y o ca e o Ge man nu sing homes. Heal h Policy 120,
1162–1170 (2016). h ps:// doi. o g/ 10. 1016/j. heal hpol. 2016. 09.
004
Publishe 's No e Sp inge Na u e emains neu al wi h ega d o
ju isdic ional claims in published maps and ins i u ional a ilia ions.