A anco, Na alia; Bauho , Sebas ian; Schwa z, Na alie; S ampini, Ma co
Wo king Pape
A e long hospi aliza ions subs i u ing p ima y and long-
e m ca e? E idence om B azil and Mexico
IDB Wo king Pape Se ies, No. IDB-WP-1632
P o ided in Coope a ion wi h:
In e -Ame ican De elopmen Bank (IDB), Washing on, DC
Sugges ed Ci a ion: A anco, Na alia; Bauho , Sebas ian; Schwa z, Na alie; S ampini, Ma co (2024) :
A e long hospi aliza ions subs i u ing p ima y and long- e m ca e? E idence om B azil and Mexico,
IDB Wo king Pape Se ies, No. IDB-WP-1632, In e -Ame ican De elopmen Bank (IDB), Washing on,
DC,
h ps://doi.o g/10.18235/0013126
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/302210
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 ps://c ea i ecommons.o g/licenses/by/3.0/igo/
A e Long Hospi aliza ions subs i u ing
P ima y and Long- e m Ca e? E idence om
B azil and Mexico
Na alia A anco
Sebas ian Bauho
Na alie Schwa z
Ma co S ampini
WORKING PAPER No IDB-WP-1632
In e -Ame ican De elopmen Bank
Social P o ec ion and Heal h Di ision
Augus 2024
A e Long Hospi aliza ions subs i u ing
P ima y and Long- e m Ca e? E idence om
B azil and Mexico
Na alia A anco
Sebas ian Bauho
Na alie Schwa z
Ma co S ampini
In e -Ame ican De elopmen Bank
Social P o ec ion and Heal h Di ision
Augus 2024
Ca aloging-in-Publica ion da a p o ided by he
In e -Ame ican De elopmen Bank
Felipe He e a Lib a y
A e long hospi aliza ions subs i u ing p ima y and long- e m ca e?: e idence
om B azil and Mexico / Na alia A anco, Sebas ian Bauho , Na alie Schwa z,
Ma co S ampini.
p. cm. — (IDB Wo king Pape Se ies ; 1632)
Includes bibliog aphical e e ences.
1. P ima y heal h ca e-B azil. 2. P ima y heal h ca e-Mexico. 3. Long- e m ca e
acili ies-B azil. 4. Long- e m ca e acili ies-Mexico. 5. Popula ion aging-
B azil. 6. Popula ion aging-Mexico. 7. Medical ca e-B azil. 8. Medical ca e-
Mexico. 9. Medical policy-B azil. 10. Medical policy-Mexico. I. A anco,
Na alia. II. Bauho , Sebas ian. III. Schwa z, Na alie. IV. S ampini, Ma co. V.
In e -Ame ican De elopmen Bank. Social P o ec ion and Heal h Di ision. VI.
Se ies.
IDB-WP-1632
h p://www.iadb.o g
Copy igh © 2024 In e -Ame ican De elopmen Bank ("IDB"). This wo k is subjec o a C ea i e
Commons license CC BY 3.0 IGO (h ps://c ea i ecommons.o g/licenses/by/3.0/igo/legalcode). The
e ms and condi ions indica ed in he URL link mus be me and he espec i e ecogni ion mus be
g an ed o he IDB.
Fu he o sec ion 8 o he abo e license, any media ion ela ing o dispu es a ising unde such license
shall be conduc ed in acco dance wi h he WIPO Media ion Rules. Any dispu e ela ed o he use o
he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he
Uni ed Na ions Commission on In e na ional T ade Law (UNCITRAL) ules. The use o he IDB's name
o any pu pose o he han o a ibu ion, and he use o IDB's logo shall be subjec o a sepa a e
w i en license ag eemen be ween he IDB and he use and is no au ho ized as pa o his license.
No e ha he URL link includes e ms and condi ions ha a e an in eg al pa o his license.
The opinions exp essed in his wo k a e hose o he au ho s and do no necessa ily e lec he iews o
he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun ies hey ep esen .
43
Abs ac
1
P olonged hospi al s ays, o hospi al s ays ha a e longe han medically necessa y, a e a majo
conce n o pa ien s, paye s, and p o ide s. We concep ualize and empi ically es ima e he
p e alence and cos o p olonged s ays among elde ly hospi al pa ien s (65 yea s and olde ) in
B azil and Mexico. We de elop a con inuum-o -ca e concep ual amewo k based on p io
li e a u e and insigh s ob ained h ough in e iews and ocus g oup discussions wi h expe s om
Mexico, A gen ina, and Colombia. In his amewo k, hospi als a e pa o a wide sys em. This
sys em in ol es bo h p e-admission and pos -discha ge medical and social ca e se ices. The e
a e h ee main sou ces o p olonged s ays: (i) lack o app op ia e p ima y heal hca e ha leads o
mo e complex admissions; (ii) hospi al ine iciency; and (iii) lack o ehabili a ion, social, and long-
e m ca e a discha ge. We es ima e he coun and sha e o inapp op ia e hospi al days due o
p olonged s ays o e all and o each sou ce. This es ima ion is based on adminis a i e eco ds
on discha ges om public sec o hospi als in 2019. Ou esul s show ha hospi al days due o
p olonged s ays accoun o app oxima ely hal o all hospi al days. Al hough mos o he
inapp op ia e days can be a ibu ed o hospi al ine iciency (36% in B azil and 49% in Mexico),
an impo an sha e is linked o he lack o ehabili a ion, social, and long- e m ca e. Lack o hese
se ices accoun s o 12% o o al hospi al days in B azil and 7% in Mexico. In a back-o - he-
en elope calcula ion, we es ima e ha p o iding six weeks o long- e m ca e se ices o add ess
he ca e needs b ough abou by only hi een causes o admission would gene a e annual ne
sa ings o app oxima ely US$174 million in B azil and US$45 million in Mexico.
Keywo ds: heal hca e cos s; p olonged hospi aliza ions; p ima y heal h ca e; long- e m ca e;
medical ca e; popula ion aging; olde pe sons; public policy; social ca e; ehabili a ion ca e; La in
Ame ica and he Ca ibbean; Mexico; B azil.
JEL classi ica ion: I10, J14, H5, J18
1
All au ho s a e wi h he Social P o ec ion and Heal h Di ision o he In e -Ame ican De elopmen Bank (IDB). Email:
[email p o ec ed]; sbauho @iadb.o g; [email p o ec ed]; ms[email p o ec ed]. This s udy was elabo a ed wi h
unding om IDB’s Economic and Sec o Wo k RG-E1871 “Can long- e m ca e se ices educe heal hca e cos s
h ough sho e hospi aliza ions?”. We hank Rica do Pé ez-Cue as, Ignacio As o ga, and Hugo Godoy o sugges ions
and guidance. We a e also g a e ul o Pablo Iba a án, Da id E ans, Agus in Filippo and an anonymous e iewe o
hei use ul commen s; Nadin Medellin and Diego Wachs o hei suppo in he da a p ocessing a he ea ly s ages o
his esea ch; Rocío Aguile a o he suppo in he quali a i e analysis. Finally, we hank ully acknowledge he
con ibu ions om expe s who kindly pa icipa ed in in e iews and ocus g oup discussions ha in o med he
cons uc ion o he concep ual amewo k. The documen was p o essionally edi ed by Guille mo Rubens. Remaining
e o s a e ou s only. The con en and indings o his pape e lec he opinions o he au ho s and no necessa ily hose
o he IDB, i s Boa d o Di ec o s, o he coun ies hey ep esen .
43
Table o Con en s
A e long hospi aliza ions subs i u ing p ima y and long- e m ca e? E idence om B azil and
Mexico ........................................................................................ E o ! Bookma k no de ined.
1. In oduc ion ......................................................................................................................... 3
2. Concep ual amewo k ........................................................................................................ 4
3. E idence om he exis ing li e a u e ................................................................................... 7
3.1. De e minan s o leng h o s ay ...................................................................................... 7
3.2. Cos s o long hospi aliza ions ....................................................................................... 8
3.3. De ini ions o p olonged hospi al s ays ......................................................................... 9
4. Da a and me hodology ......................................................................................................10
4.1. Da a sou ces ...............................................................................................................10
4.2. De ini ion o P olonged Hospi aliza ions and Decomposi ion o Leng h o S ay ...........12
5. E idence om B azil and Mexico .......................................................................................16
5.1. P olonged hospi aliza ions accoun o app oxima ely hal o all hospi al days ............16
5.2. Which condi ions accoun o mos excessi e days? ...................................................17
5.3. How much can be sa ed by p o iding ehabili a ion, social, and long- e m ca e se ices?
..........................................................................................................................................25
6. Discussion .........................................................................................................................26
7. Conclusions and policy ecommenda ions .........................................................................28
Re e ences ...........................................................................................................................30
Annex 1. Findings om in e iews and ocus g oup discussions ............................................35
Annex 2. Decomposi ion o hospi al days in B azilian and Mexican s a es .............................40
Annex 3. Sensi i i y analysis .................................................................................................42
43
1. In oduc ion
In he La in Ame ican and Ca ibbean egion, he con e gence o an aging popula ion and
echnological ad ancemen s is expec ed o signi ican ly aise heal hca e spending (Rao e al.,
2022). This inc ease is u he compounded by he ising p e alence o ch onic diseases and
dependence among olde adul s. O e 85% o hose aged 70 and abo e ha e a leas one ch onic
condi ion, and 14% o hose o e 65 equi e assis ance wi h ac i i ies o daily li ing (IHME, 2020;
A anco, Iba a án, and S ampini, 2022). Fu he mo e, he e is a no able lack o obus p ima y
heal h, social, pos -ope a i e, and long- e m ca e sys ems, as well as suppo o amily ca egi e s
(A anco e al., 2022).
Ra ionalizing hospi al use is a key s a egy o con olling ising heal h expendi u es. Hospi als
accoun o abou one- hi d o o al heal h spending in he egion and a e cen al o he adop ion
o cos ly medical echnology. They also bea he consequences o inadequa e p ima y ca e and
social sys ems, which can lead o hospi al s ays ex ending beyond wha is medically necessa y.
A p olonged hospi aliza ion occu s when "a medically i pa ien is needlessly kep in hospi al due
o in e nal o ganiza ional/ope a ional ac o s o whe e a pa ien is lagged as in need o al e na e
le el o ca e and is delayed because o de e ed ansi ion o ca e and/o lack o ex e nal ans e -
o -ca e a angemen s" (Micalle e al., 2020, p. 105).
P olonged hospi aliza ions can be d i en by se e al ac o s. Fi s , p e en able como bidi ies o
pa ien ail y may ex end hospi al s ays. Second, ine iciencies wi hin he hospi al may lead o
longe s ays. Thi d, a lack o app op ia e discha ge des ina ions ha o e ehabili a i e ca e o
social suppo can also lead o p olonged hospi aliza ions. Addi ionally, hospi als ace admissions
and eadmissions ha a e en i ely a oidable wi h e ec i e p ima y and social ca e. Olde pe sons
a e pa icula ly a isk o p olonged hospi aliza ions due o hei mo e complex heal h condi ions,
ail y a admission, and he need o sa e discha ge a angemen s (Picone e al., 2003; Lenzi e
al., 2014).
P olonged hospi al s ays a e common, cos ly, and isky o pa ien s. A me a-analysis o 64 s udies
conduc ed in Eu ope and No h Ame ica ound ha such s ays accoun o an a e age o 22.8%
o all bed days. The igu es ange om 1.6% in England o 91.3% in Canada, depending on he
me hodologies, da a sou ces, and popula ions s udied (Landei o e al., 2019). The e is also
subs an ial wi hin-coun y a ia ion. These p olonged s ays con ibu e o inc eased heal hca e
cos s and can wo sen access and wai imes when hospi al capaci y is limi ed (Falcone e al.,
n.d.; Landei o e al., 2019). Addi ionally, p olonged hospi aliza ions can be po en ially unsa e o
pa ien s (Lingsma e al., 2018; Landei o e al., 2019; Rojas‐Ga cía e al., 2018). Exis ing esea ch
has iden i ied se e al d i ing ac o s o p olonged hospi aliza ions, including a lack o adequa e
ca e s uc u es ou side he hospi al, as well as hospi al and in e -hospi al p ocesses om
admission o discha ge, such as ea ly admission o ese e a bed o a scheduled p ocedu e o
adminis a i e delays (Landei o e al., 2019; Siddique e al., 2021; Micalle e al., 2020).
In his s udy, we concep ualize and empi ically measu e hospi al days due o p olonged s ays
among olde people in B azil and Mexico. Fi s , we p opose a con inuum-o -ca e concep ual
amewo k o p olonged s ays based on exis ing li e a u e and quali a i e insigh s om in e iews
and ocus g oup discussions wi h expe s om Mexico, A gen ina, and Colombia. The amewo k
ca ego izes d i e s o p olonged s ays in o h ee pa s: (i) lack o app op ia e p ima y heal hca e,
leading o mo e complex admissions; (ii) hospi al ine iciency; and (iii) lack o ehabili a ion, social,
and long- e m ca e a discha ge. Second, we es ima e he p e alence o p olonged s ays and he
con ibu ion o hese h ee ac o s using adminis a i e eco ds on discha ges o pa ien s aged
65 and olde om public sec o hospi als in B azil and Mexico om 2019. Finally, we es ima e he
o al cos o inapp op ia e days, by mul iplying hei numbe by he a e age cos o one day o
43
hospi aliza ion (including in as uc u e and equipmen amo iza ion, p ocedu es and human
esou ces).
Ou esul s sugges ha p olonged s ays a e highly p e alen and cos ly, p ima ily d i en by
hospi al ine iciencies and he lack o discha ge des ina ions ha p o ide pos -ope a i e,
ehabili a i e, and social suppo . Speci ically, we es ima e ha inapp op ia e hospi al days
accoun o 48.1% o hospi al days in B azil and 56.2% in Mexico. The sca ce supply o
ehabili a ion, social, and long- e m ca e se ices accoun s o 12.1% o all hospi al days in B azil
and 6.9% in Mexico.
To he bes o ou knowledge, his is he i s pape o p o ide such a decomposi ion o he causes
o p olonged hospi aliza ions, allowing o an ini ial es ima ion o he sa ings ha could be
achie ed h ough he p o ision o long- e m ca e. The emainde o he pape is o ganized as
ollows. Sec ion 2 ou lines he concep ual amewo k ha si ua es hospi als wi hin a b oade
heal h and social ca e sys em. We illus a e how p ima y heal hca e, hospi al ine iciency, and
ehabili a ion, social, and long- e m ca e can a ec he leng h o s ay. Sec ion 3 e iews he
li e a u e on he de e minan s, de ini ion, and cos s o long hospi al s ays. Sec ion 4 desc ibes he
da a and explains he me hodology used o de ine p olonged hospi aliza ions, and decompose
he leng h o s ay in o i s componen s, ollowing he logic o ou concep ual model. Sec ion 5
p esen s he esul s on he magni ude o inapp op ia e hospi al days and in o ma ion on he
condi ions ha con ibu e mos o hese excessi e days, as well as he po en ial sa ings om
imp o ed pos -discha ge se ices. In Sec ion 6, we u he discuss ou indings. Sec ion 7
concludes and p o ides policy ecommenda ions.
2. Concep ual amewo k
We de eloped a con inuum-o -ca e concep ual amewo k o p olonged hospi al s ays based on
exis ing li e a u e and quali a i e insigh s om in e iews and ocus g oup discussions wi h
expe s om Mexico, A gen ina, and Colombia. These discussions and in e iews, which included
6 pa icipan s om ocus g oups and 3 medical doc o s, we e conduc ed i ually be ween June
and Oc obe 2023. The da a was analyzed using hema ic analysis, an induc i e app oach ha
helped us iden i y key hemes and pa e ns. This analysis allowed us o concep ualize he
in o ma ion in o h ee s ages o ca e: p e-hospi al, in-hospi al, and pos -hospi al. Box 1 p esen s
selec ed quo es ha in o med ou model, and Annex 1 p o ides a mo e de ailed summa y o he
indings.
Ou amewo k iews hospi als as pa o a b oade heal h and social ca e sys em ha includes
p ima y ca e and se ices o ehabili a ion, social suppo , and long- e m ca e (Falcone e al.,
n.d.). P olonged hospi al s ays can esul om ine iciencies and bo lenecks a any s age o his
ca e con inuum, leading o a oidable admissions o longe hospi aliza ions han medically
necessa y. Fo ins ance, pa ien s may be admi ed oo ea ly o oo la e, expe ience delays in
becoming clinically i o discha ge once admi ed, o ace discha ge delays due o a lack o pos -
hospi al ca e suppo .
Figu e 1 illus a es ou concep ual amewo k and iden i ies h ee po en ial sou ces o
inapp op ia e hospi al days:
1. In ake Issues: Bo h inadequa e p ima y heal hca e (A) and insu icien ehabili a ion,
social, and long- e m ca e se ices (B) can lead o a oidable admissions o eadmissions
and inc ease pa ien ail y, con ibu ing o p olonged hospi aliza ions (Componen 1)
(F ei as e al., 2012; Lenzi e al., 2014; Bo e al., 2016; Toh e al., 2017). Fo example, an
olde pe son hospi alized o a emu ac u e may equi e a longe s ay i hey ha e poo ly
43
managed ch onic condi ions ha need s abiliza ion o addi ional ca e needs ha
complica e ea men . The same conside a ions apply o eadmissions. Lack o p ima y
heal hca e (A) and insu icien ehabili a ion, social, and long- e m ca e (B) inc ease he
p obabili y o ehospi aliza ion o people who ha e been p e iously discha ged and may
complica e he clinical pic u e, ex ending he du a ion o hese eadmissions (Misky e al.,
2010). Addi ionally, oo-ea ly discha ges due o in e nal hospi al issues may also inc ease
he likelihood o eadmission.
2. In-Hospi al Ine iciencies: P olonged s ays du ing hospi aliza ion (C) may a ise om
ine iciencies such as lack o esou ces, delays in p ocedu es, o poo planning and
managemen (Componen 2) (Holmås, Kam ul Islam, e al., 2013). Fo ins ance, hospi als
wi h a lowe physician- o-pa ien a io may expe ience longe s ays as pa ien s wai longe
o consul a ions o es esul s (Ma il-Ga za e al., 2018; Ca ey e al., 2005).
3. Discha ge Delays: Discha ges o clinically i pa ien s may be delayed due o a lack o
a ailable ehabili a ion, social, and long- e m ca e se ices (B) (Componen 3) (Toh e al.,
2017; Landei o e al., 2016; Moo e e al., 2015; Ca ey e al., 2005). Fo example, olde
pa ien s may emain hospi alized i hey need ca e o ehabili a ion ha canno be p o ided
a home due o a lack o amily suppo o public home ca e se ices. Al e na i ely, hey
may ace delays i sui able ins i u ional a angemen s (e.g., ehabili a ion cen e s o long-
e m ca e acili ies) a e una ailable. Such needs may a ise om he hospi aliza ion i sel
o p eexis bu become mo e p onounced pos -hospi aliza ion. Addi ionally, amilies migh
use he hospi al s ay as an oppo uni y o ob ain ca e om public se ices and may seek
o delay o p e en he pa ien 's discha ge.
Below, we ope a ionalize his amewo k o es ima e he con ibu ion o he h ee componen s –
in ake, in-hospi al, and discha ge – o he o e all coun and p e alence o inapp op ia e hospi al
days (p olonged s ays) (Figu e 1). In p ac ice, dis inguishing be ween hospi aliza ions and
eadmissions is challenging because each hospi aliza ion episode is eco ded sepa a ely and
canno be linked o p e ious s ays. Mo eo e , we canno de e mine whe he eadmissions esul
om issues wi h he ini ial s ay (e.g., p ema u e discha ge) o de iciencies in non-hospi al suppo
se ices. Thus, he e ec on eadmissions is also conside ed unde Componen 1. Addi ionally,
we canno de e mine whe he a s ay is p olonged due o amily e usal o discha ge he pa ien .
43
Table 1. Main cha ac e is ics o he sample
B azil
Mexico
Numbe o hospi aliza ions
2,166,900
768,173
Numbe o days o hospi aliza ion
15.006.168
4,639,140
A e age leng h o s ay (days)
(S anda d de ia ion)
6.9
(10.6)
6.0
(30.4)
Age 65-69 (%)
26.8
27.9
Age 70-74 (%)
23.2
24.0
Age 75-79 (%)
19.6
19.7
Age 80-84 (%)
15.0
14.4
Age 85+ (%)
15.3
14.0
A e age age (yea s)
75.7
75.4
Females (%)
49.7
51.8
Males (%)
50.3
48.2
% wi h como bidi ies
22.1
-
Cha lson como bidi y index
0.27
Type o admission: Elec i e (%)
17.3
-
Type o admission: Eme gency (%)
82.7
-
P ocedu e complexi y: Medium (%)
89.6
-
P ocedu e complexi y: High (%)
10.4
-
In-hospi al mo ali y (%)
12.8
10.6
Sou ce: Au ho s’ elabo a ion based on DATASUS hospi aliza ion da abase, 2019 and Mexico's Heal h Sec o Hospi al
Discha ge Da abase, 2019.
4.2. De ini ion o P olonged Hospi aliza ions and Decomposi ion o Leng h o S ay
Following he concep ual amewo k, we decompose a hospi al leng h o s ay (LOS) in o ou
pa s: (i) he medically app op ia e s ay (T); (ii) excessi e days due o he lack o app op ia e
p ima y heal hca e ha leads o mo e complex admissions (ED1); (iii) excessi e days due o
hospi al ine iciency (ED2); (i ) excessi e days due he lack o ehabili a ion, social, and long- e m
ca e a discha ge (ED3). We include all p ima y diagnoses in he analysis, e en hose condi ions
ha should ha e been p e en ed a he p ima y ca e le el. Ex ensi e e idence sugges s ha a
obus p ima y ca e sys em can educe hospi aliza ions. Howe e , his pape does no aim o
quan i y he po en ial sa ings om such educ ions; ins ead, i ocuses on unde s anding he
ac o s con ibu ing o p olonged hospi al s ays once admission has occu ed. Box 2 illus a es
he decomposi ion using ou s e eo ypical examples.
Fo each condi ion, we de ine he medically app op ia e du a ion o s ay as he a e age leng h o
s ay in he mos e icien s a e ( he “benchma k s a e”) o , in o he wo ds, in he s a e wi h he
lowes a e age leng h o s ay o ha condi ion. To calcula e his a e age, we es ic he sample
o pa ien s wi hou como bidi ies (seconda y diagnoses) ha could ha e been p e en ed a he
p ima y le el. Tha is, only pa ien s wi hou Ambula o y-Ca e-Sensi i e Condi ions (ACSC)
como bidi ies a e conside ed o de e mine he benchma k s a e. By es ic ing he sample in his
way, we emo e he po ion o he s ay ha is a ibu ed o clinical complica ions linked o
como bidi ies ha could ha e been managed h ough app op ia e p ima y heal hca e.
3
Implici ly,
3
Fo hospi aliza ions due o ACSC (as p ima y condi ion), bo h LOS and ED1 could be a oided h ough app op ia e
p ima y heal hca e.
43
we assume ha in he benchma k s a e, ACSC como bidi ies a e p ope ly managed a he p ima y
le el.
We p oceed as ollows. Fo each condi ion, we calcula e he a e age du a ion o hospi aliza ions
o pa ien s wi hou ACSC como bidi ies by s a e (𝐿𝑂𝑆
𝑐,𝑠). We hen ake he lowes a e age alue
as he h eshold ha de ines he medically app op ia e s ay 𝑇𝑐. The calcula ions a e gi en by
equa ions [1] and [2].
𝐿𝑂𝑆
𝑐,𝑠 =∑𝐿𝑂𝑆𝑖,𝑐,𝑠
⬚
𝑖,𝑐,𝑠
∑𝐼𝑖,𝑐,𝑠
⬚
𝑖,𝑐,𝑠
o 𝑖 wi hou ACSC − como bidi ies [1]
𝑇𝑐=min𝑠( 𝐿𝑂𝑆
𝑐,𝑠) [2]
Whe e 𝐿𝑂𝑆𝑖,𝑐,𝑠 is he leng h o s ay o pa ien i, o condi ion c in s a e s. I is an indica o equal o
1 ha coun s he hospi aliza ions o he pu pose o calcula ing he a e age leng h o s ay.
Equa ion [2] iden i ies T and he co esponding benchma k s a e (BS), o each condi ion c.
Fo each hospi aliza ion, he numbe o excessi e days 𝐸𝐷 is he di e ence be ween he ac ual
leng h o s ay and he h eshold, as shown in equa ion [3]. 𝐸𝐷 is also equal o he sum o i s h ee
componen s, as shown in equa ion [4].
𝐸𝐷𝑖,𝑐 = 𝐿𝑂𝑆𝑖,𝑐 − 𝑇𝑐 𝑖𝑓 𝐿𝑂𝑆𝑖,𝑐 > 𝑇𝑐; 0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 [3]
𝐸𝐷 =𝐸𝐷1 + 𝐸𝐷2 + 𝐸𝐷3 [4]
To calcula e 𝐸𝐷1, i.e., he excessi e days ha could ha e been a oided h ough app op ia e
managemen o ACSC como bidi ies a he p ima y le el, we calcula e he a e age leng h o s ay
(𝑍𝑐) among pa ien s wi h ACSC como bidi ies hospi alized o condi ion c in he benchma k s a e
BS iden i ied in equa ion [2]. The o mula is shown in equa ion [5].
𝑍𝑐=∑𝐿𝑂𝑆𝑖,𝑐,𝐵𝑆
⬚
𝑖,𝑐,𝐵𝑆
∑𝐼𝑖,𝑐,𝐵𝑆
⬚
𝑖,𝑐,𝐵𝑆
o 𝑖 wi h ACSC − como bidi ies [5]
Only pa ien s epo ing ACSC como bidi ies a e conside ed in equa ion [5]. The assump ion is
ha poo managemen o ACSC como bidi ies would lead o longe s ays e en in he mos e icien
s a e, because i inc eases he complexi y o a pa ien ’s clinical pic u e a admission. This implies
ha Z is la ge han T.
4
Fo pa ien s wi h ACSC como bidi ies in all s a es, 𝐸𝐷1 is compu ed as he di e ence be ween 𝑍𝑐
and 𝑇𝑐, o he di e ence be ween he ac ual leng h o s ay and 𝑇𝑐 i he leng h o s ay is sho e
han 𝑍𝑐. Fo pa ien s wi hou ACSC como bidi ies, 𝐸𝐷1 is ze o by de ini ion. This is summa ized
in equa ion [6].
4
In ew cases whe e Z<T, we se Z=T. Also, in e y ew ou lying cases in which Z exceeds 2T, we se Z=2T.
43
𝐸𝐷1𝑖,𝑐 = 0 o 𝑖 wi hou ACSC − como bidi ies [6]
𝐸𝐷1𝑖,𝑐 = 𝑍𝑐− 𝑇𝑐 i 𝑖 has ACSC − como bidi ies and 𝐿𝑂𝑆𝑖,𝑐 > 𝑍
𝐸𝐷1𝑖,𝑐 = 𝐿𝑂𝑆𝑖,𝑐 − 𝑇𝑐 i 𝑖 has ACSC − como bidi ies and 𝑍𝑐>𝐿𝑂𝑆𝑖,,𝑐 > 𝑇𝑐
𝐸𝐷1𝑖,𝑐 = 0 i 𝑖 has ACSC − como bidi ies and 𝐿𝑂𝑆𝑖,𝑐 < 𝑇
This calcula ion o 𝐸𝐷1 can only be done o B azil, as he Mexican da a do no include in o ma ion
on como bidi ies. Consequen ly, o Mexico 𝐸𝐷1 is included pa ly in T and pa ly in 𝐸𝐷2. I should
also be no ed ha a s ong p ima y heal hca e would a oid admissions due o ACSC al oge he .
Tha is, he 𝑇𝑐 pa o admissions due o ACSC as p ima y condi ions can also be conside ed
excessi e days (mo e speci ically, ED1). Howe e , o his analysis, we aim o iden i y he
con ibu ion o ACSC como bidi ies o p olonged hospi aliza ions, e en in cases whe e a pe son
has been hospi alized due o a p ima y ACSC.
A e accoun ing o ED1, he decomposi ion o he emaining excessi e days depends on whe he
he condi ion ha caused admission gene a es new pos -discha ge ca e needs. Fo ins ance,
people wi h musculoskele al condi ions will equi e ehabili a ion ca e a e lea ing he hospi al.
Simila ly, neu ological condi ions and sys emic diseases, mainly espi a o y and ca diac ailu es,
a ec a pa ien 's mobili y and hus gene a e new pos -discha ge ehabili a ion needs. In con as ,
pa ien s wi h mo e gene ic condi ions, such as diabe es, can eco e wi hou ehabili a ion ca e.
We iden i y he condi ions ha gene a e new ca e needs (CN) h ough expe opinions who
assessed a se o 35 diagnoses ha mos con ibu e o excessi e hospi al days in ou analysis.
5
We assume ha condi ions no assessed belong o he no-new-ca e-needs (NCN) g oup.
I a condi ion does no c ea e new ca e needs (NCN), we assume ha all emaining excessi e
days a e due o hospi al ine iciency (ED2). I a condi ion gene a es pos -discha ge ca e needs
(CN), he emaining excessi e days a e u he disagg ega ed in o ED2 and lack o ehabili a ion,
social, and long- e m ca e (ED3).
Fo condi ions in he NCN g oup, he componen due o hospi al ine iciency is de ined by
equa ion [7].
𝐸𝐷2𝑖,𝑐 = 𝐿𝑂𝑆𝑖,𝑐 − 𝑇𝑐−𝐸𝐷1𝑖,𝑐 𝑖𝑓 𝑐 ∈ 𝑁𝐶𝑁 𝑎𝑛𝑑 𝐿𝑂𝑆𝑖,𝑐 >(Tc+𝐸𝐷1𝑖,𝑐); 0 o he wise [7]
Fo condi ions in he CN g oup, we assume ha hospi al ine iciency is equal o he a e age
ine iciency obse ed o he NCN g oup, 𝐸𝐷2
, which is de ined as:
5
The ollowing condi ions we e classi ied as gene a ing medium o high ca e needs: Angina pec o is; Bac e ial in ec ion
o unspeci ied si e; Bac e ial pneumonia, no elsewhe e classi ied; Ce eb al in a c ion; F ac u e o he emu ; F ac u e
o lowe leg, including ankle; Hea ailu e; O he ch onic obs uc i e pulmona y disease; O he degene a i e diseases
o ne ous sys em, no elsewhe e classi ied; O he sepsis; Pneumonia, o ganism unspeci ied; Sequelae o
ce eb o ascula disease; Shock, no elsewhe e classi ied; S oke, no speci ied as hemo hage o in a c ion. The
ollowing condi ions we e assessed as gene a ing no o low ca e needs: Acu e myoca dial in a c ion, Cholecys i is;
Choleli hiasis; Ch onic ischemic hea disease; Ch onic kidney disease; Epilepsy; Essen ial (p ima y) hype ension;
In ac anial inju y; Malignan neoplasm o colon; O he bac e ial diseases, no elsewhe e classi ied; O he
ce eb o ascula diseases; O he diseases o diges i e sys em; O he diso de s o luid, elec oly e and acid-base
balance; O he diso de s o skin and subcu aneous issue, no elsewhe e classi ied; O he diso de s o u ina y sys em;
O he pe iphe al ascula diseases; Pa aly ic ileus and in es inal obs uc ion wi hou he nia; Respi a o y ailu e, no
elsewhe e classi ied; Type 2 diabe es melli us; Unknown and unspeci ied causes o mo bidi y; Unspeci ied diabe es
melli us. See also Table 4.
43
𝐸𝐷2
=∑𝐸𝐷2𝑖,𝑐
⬚
𝑖,𝑐
∑𝐼𝑖,𝑐
⬚
𝑖,𝑐
, o 𝑐 ∈ 𝑁𝐶𝑁 [8]
𝐸𝐷2 is hen de ined by equa ion [9].
𝐸𝐷2𝑖,𝑐 =𝐸𝐷2
𝑖𝑓 𝑐 ∈ 𝐶𝑁 𝑎𝑛𝑑 𝐿𝑂𝑆𝑖,𝑐 >(Tc+𝐸𝐷1𝑖,𝑐 +𝐸𝐷2
) [9]
𝐸𝐷2𝑖,𝐶 = 𝐿𝑂𝑆𝑖,𝑐 − Tc−𝐸𝐷1𝑖,𝑐 i 𝑐 ∈ 𝐶𝑁 𝑎𝑛𝑑 (Tc+𝐸𝐷1𝑖,𝑐 +𝐸𝐷2
)> LOSi,c >(Tc+𝐸𝐷1𝑖,𝑐)
𝐸𝐷2𝑖,𝐶 = 0 o he wise
Finally, 𝐸𝐷3, he numbe o excessi e days ha a e due o he lack o ehabili a ion, social and
long- e m ca e se ices o condi ions ha gene a e new ca e needs is de ined by equa ion [10].
Fo condi ions ha gene a e no ca e needs, ED3 is ze o by de ini ion.
𝐸𝐷3𝑖,𝑐 = 0 o 𝑐 ∈ 𝑁𝐶𝑁 [10]
𝐸𝐷3𝑖,𝑐 = 𝐿𝑂𝑆𝑖,𝑐 − 𝑇𝑐−𝐸𝐷1𝑖,𝑐 −𝐸𝐷2𝑖,𝑐 i 𝑐 ∈ 𝐶𝑁 𝑎𝑛𝑑 𝐿𝑂𝑆𝑖,𝑐 >(Tc+𝐸𝐷1𝑖,𝑐 +𝐸𝐷2𝑖,𝑐),
𝐸𝐷3𝑖,𝐶 = 0 o he wise
Box 2 illus a es his decomposi ion analysis o di e en examples o condi ions and ypes o
pa ien s.
Box 2. Examples o decomposi ion calcula ions, B azilian da abase
Example 1: Pa ien wi h acu e myoca dial in a c ion, no ACSC como bidi ies.
LOS
Tc
Zc
ACSC
New ca e needs
ED
ED1
ED2
ED3
9
5.1
6.2
No
No
3.9
0
3.9
0
Example 2: Pa ien wi h acu e myoca dial in a c ion, wi h ACSC como bidi ies.
LOS
Tc
Zc
ACSC
New ca e needs
ED
ED1
ED2
ED3
10.2
5.1
6.2
Yes
No
5.1
1.1
4.0
0
Example 3: Pa ien wi h ac u e o he emu , no ACSC como bidi ies.
LOS
Tc
Zc
ACSC
New ca e needs
ED
ED1
ED2
ED3
12
6.7
7.0
No
Yes
5.3
0
3.1
2.2
Example 4: Pa ien wi h ac u e o he emu , wi h ACSC como bidi ies.
LOS
Tc
Zc
ACSC
New ca e needs
ED
ED1
ED2
ED3
12
6.7
7.0
Yes
Yes
5.3
0.2
3.1
2.0
43
5. E idence om B azil and Mexico
5.1. P olonged hospi aliza ions accoun o app oxima ely hal o all hospi al days
The decomposi ion analysis shows ha excessi e days ep esen 48.1% o o al hospi al days in
B azil and 56.2% in Mexico (Figu e 3). These excessi e days come om 1 million hospi aliza ions
in B azil and 440,000 hospi aliza ions in Mexico classi ied as p olonged, ep esen ing 46% and
57% o he o al numbe o hospi aliza ions, espec i ely.
In B azil, we es ima e ha 0.5% o o al hospi al days esul om he inc eased agili y and clinical
complexi y o pa ien s wi h seconda y condi ions ha could ha e been managed a he p ima y
ca e le el ( e e ed o as ACSC como bidi ies). Addi ionally, 35.5% o hospi al days a e
a ibu able o ine iciencies wi hin hospi als, such as lack o esou ces and managemen models.
Finally, 12.1% o hospi al days could be a oided by p o iding be e ehabili a ion, social, and
long- e m ca e se ices.
In Mexico, hospi al ine iciency is he p edominan ac o , accoun ing o 49.3% o o al hospi al
days. The pe cen age o hospi al days ha could be sa ed h ough he p o ision o ehabili a ion,
social and long- e m ca e se ices is lowe han in B azil, a a 6.9%. As discussed in he
me hodology sec ion, he da a om Mexico does no allow o he es ima ion o he sha e o
excessi e days due o mismanagemen o ACSC como bidi ies. The e o e, hese excessi e days
a e pa ly included in he es ima ed medically app op ia e s ay and pa ly in he sha e o excessi e
days due o hospi al ine iciency.
Ou analysis conside s all main causes o hospi aliza ions, including hose o condi ions ha
could ha e been a oided wi h be e p ima y ca e (i.e., hospi aliza ions wi h a p ima y diagnosis
o ACSC). F om he hospi al’s pe spec i e, hese a e alid admissions. Howe e , om he b oade
pe spec i e o a heal h sys em, hese admissions should no ha e occu ed and he e o e he
app op ia e leng h o s ay should be ze o days, making all days o hese condi ions excessi e.
Expec ed days o s ay om ACSC cons i u e 14.4% and 12.5% o all hospi al days in B azil and
Mexico, espec i ely. I we coun all days om admissions wi h p ima y diagnosis o ACSC as
medically inapp op ia e, excessi e days accoun o 64.7% and 68.7% o o al hospi al days in
B azil and Mexico, espec i ely.
43
Figu e 3. Decomposi ion o hospi al days in B azil and Mexico, 2019
Sou ce: Au ho s’ elabo a ion based on DATASUS hospi aliza ion da abase, 2019 and Mexico's Heal h Sec o Hospi al
Discha ge Da abase, 2019. No e: Componen ED1 canno be es ima ed o Mexico, due o lack o in o ma ion on ACSC
como bidi ies.
The sha e o excessi e days ha can be a ibu ed o he di e en componen s a ies by s a e
(Annex 2, Tables A2.1 and A2.2), especially o he pa ela ed o he lack o ehabili a ion, social,
and long- e m ca e se ices. In B azil, his componen accoun s o a sha e o o al hospi al days
anging be ween 3.6% in Ma o G osso do Sul and 19% in Amapá. In Mexico, i accoun s o a
sha e o o al hospi al days anging om 5% in Michoacán de Ocampo and Tabasco o 10.7% in
Mexico Ci y. Hospi al ine iciency accoun s o 24.8% in Pa aná and up o 46.0% in Rio de Janei o.
In Mexico, hese igu es ange om 42.4% in Colima o 52.9% in Baja Cali o nia.
5.2. Which condi ions accoun o mos excessi e days?
43
Table 2 shows ha in bo h B azil and Mexico, he expec ed s ay 𝑇𝑐 a ies subs an ially ac oss
condi ions. Conside ing he 20 condi ions esponsible o mos o he excessi e days, in B azil,
his pa ame e anges be ween jus o e 2 days o hype ension o nea ly 10 days o sequalae
o s oke. In Mexico, i a ies om 2.6 days o unknown causes o mo bidi y o sligh ly mo e han
6 days o a ac u e o he emu . Simila ly, he obse ed a e age leng h o s ay a ies g ea ly
ac oss condi ions. Fo he same 20 condi ions in bo h coun ies, he a e age leng h o s ay in
B azil anges om 4.3 days o choleli hiasis o 32.6 days o sequelae o ce eb o ascula
disease. In Mexico, i a ies be ween 4.7 days in he case o hype ension and 9.7 days o
ac u e o he emu .
19
Table 2 also shows ha jus 20 condi ions accoun o 51% o excessi e days in B azil (Panel A)
and 41% in Mexico (Panel B). In B azil, wo in ec ion- ela ed diseases, pneumonia and sepsis,
a e a he op o he cha , join ly accoun ing o mo e han 11% o excessi e days (Table 2,
Panel A). I we add bac e ial pneumonia, bac e ial in ec ion, e ysipelas, and o he bac e ial
diseases, he sha e due o in ec ious diseases eaches 18% o excessi e days. Ca dio ascula
diseases also ank high, wi h 5.5% o excessi e days due o hea ailu e, 4.3% o s oke, 3.1%
o acu e myoca dial in a c ion, 2.2% o angina pec o is, and 1.3% o sequelae o ce eb o ascula
diseases. Taken oge he , hese ca dio ascula condi ions accoun o 16.6% o excessi e days.
In Mexico, he ac u e o he emu is he condi ion ha accoun s o mos excessi e days, 5.4%
o he o al (Table 2, Panel B). Ch onic kidney disease, pneumonia, and diabe es melli us ( ype
2) join ly accoun o an addi ional 11.4%.
Table 3 indica es key demog aphic and hospi aliza ion cha ac e is ics o he condi ions epo ed
in Table 2, compa ing he ull sample o hospi al s ays wi h he subsample o p olonged
hospi aliza ions by coun y. O e all, he e a e no clea pa e ns di e en ia ing p olonged
hospi aliza ions om all hospi al s ays. In bo h B azil (Table 3, Panel A) and Mexico (Table 3,
Panel B), demog aphic and heal h cha ac e is ics a e simila among subsamples, excep o
condi ions like choleli hiasis, unspeci ied diabe es melli us, and cholecys i is ha show sligh ly
lowe pe cen ages o emales among p olonged s ays in Mexico (Table 3, Panel B). Addi ionally,
in B azil, he ype o admission (elec i e e sus eme gency) and he complexi y o he p ocedu e
unde aken do no di e subs an ially be ween he wo samples. In addi ion, in bo h coun ies, in-
hospi al mo ali y a ies sligh ly ac oss subsamples, bu wi hou a clea pa e n, and he
magni ude o he di e ences is ela i ely small. In B azil (Table 3, Panel A), o hea ailu e,
s oke, and bac e ial diseases, excessi e days appea co ela ed wi h a highe mo ali y a e. In
con as , o condi ions like sepsis and sequelae o s oke excessi e days seem associa ed wi h
lowe le els o mo ali y. In Mexico (Table 3, Panel B), p olonged hospi aliza ions show lowe in-
hospi al mo ali y a es o acu e myoca dial in a c ion and sepsis, and highe a es o diseases
classi ied as unknown and unspeci ied causes o mo bidi y.
F om Table 3, i seems ha he quan i y o excessi e days is p ima ily a ibu ed o he cause o
hospi aliza ion a he han he pa ien 's cha ac e is ics. This aligns wi h ou app oach o using an
index ha a ies by condi ion o iden i y he componen o excessi e days due o lack o
ehabili a ion, social and long- e m ca e se ices.
20
Table 2. Top 20 condi ions esponsible o excessi e days o hospi aliza ion - Panel A. B azil
Th esholds
All s ays
Long s ays
T1
Z2
a .
days3
SD4
days om
condi ion5
% in o al
days6
cases om
condi ion7
% o long
hosp.8
a .
days9
SD10
# exc. days11
sha e in
excessi e
days12
Pneumonia, o ganism unspeci ied
5.46
7.61
7.17
7.36
1,024,904
6.8%
142,972
46%
11.99
8.64
424,799
5.9%
O he sepsis
5.78
5.78
11.23
11.37
695,576
4.6%
61,914
64%
15.98
11.82
402,662
5.6%
Hea ailu e
5.11
6.97
7.39
7.87
868,204
5.8%
117,441
44%
12.82
9.23
400,795
5.5%
S oke, no speci ied as hemo hage o in a c ion
5.52
9.35
7.62
9.00
684,707
4.6%
89,896
44%
13.41
11.01
312,943
4.3%
Acu e myoca dial in a c ion
5.07
6.17
8.22
9.30
431,797
2.9%
52,547
48%
14.08
10.66
225,682
3.1%
Bac e ial pneumonia, no elsewhe e classi ied
5.00
8.67
7.47
7.86
472,917
3.2%
63,320
47%
12.39
9.19
219,401
3.0%
O he diso de s o u ina y sys em
4.10
5.00
6.41
7.15
430,282
2.9%
67,101
50%
10.36
8.51
206,121
2.9%
O he ch onic obs uc i e pulmona y disease
4.60
6.24
6.84
9.39
384,751
2.6%
56,228
47%
11.45
12.04
182,348
2.5%
F ac u e o he emu
6.76
7.00
8.62
8.21
461,928
3.1%
53,559
49%
13.71
9.22
181,857
2.5%
Ch onic kidney disease
5.84
6.50
9.73
11.30
311,414
2.1%
32,000
54%
15.67
12.62
169,391
2.3%
Angina pec o is
3.16
6.31
5.52
6.70
291,174
1.9%
52,735
46%
9.74
7.94
160,969
2.2%
Bac e ial in ec ion o unspeci ied si e
7.60
9.38
9.48
9.81
320,050
2.1%
33,760
42%
16.93
11.28
133,193
1.8%
Respi a o y ailu e, no elsewhe e classi ied
5.10
5.10
9.37
11.14
172,712
1.2%
18,430
50%
16.05
12.57
100,645
1.4%
Sequelae o ce eb o ascula disease
9.78
11.00
32.63
66.18
126,058
0.8%
3,863
60%
52.07
80.16
97,169
1.3%
O he diseases o diges i e sys em
3.66
5.24
5.42
6.18
201,272
1.3%
37,167
49%
8.96
7.26
96,586
1.3%
Essen ial (p ima y) hype ension
2.32
2.54
5.85
19.20
115,465
0.8%
19,723
46%
10.82
27.45
77,307
1.1%
Choleli hiasis
3.07
3.07
4.34
6.17
134,577
0.9%
31,037
31%
10.23
8.52
68,385
0.9%
O he bac e ial diseases, no elsewhe e classi ied
6.14
12.27
10.94
11.07
123,663
0.8%
11,303
56%
16.64
11.97
66,472
0.9%
Unspeci ied diabe es melli us
3.98
3.98
6.12
6.94
138,245
0.9%
22,587
55%
9.21
8.08
65,481
0.9%
E ysipelas
5.02
10.04
7.41
7.24
146,680
1.0%
19,804
50%
11.45
8.36
64,149
0.9%
To al
7,536,377
50%
987,387
3,656,357
51%
21
Panel B. Mexico
Th eshold
All s ays
Long s ays
T1
a . days3
SD4
days om
condi ion5
% in o al
days6
cases om
condi ion7
% o long
hosp.8
a . days9
SD10
# exc. days11
sha e in
excessi e
days12
F ac u e o he emu
6.34
9.66
11.46
307,947
6.6%
31,885
62%
13.42
13.15
139,905
5.4%
Ch onic kidney disease
3.26
5.70
10.32
228,505
4.9%
40,092
49%
9.84
13.58
128,604
4.9%
Pneumonia, o ganism unspeci ied
5.66
8.13
93.93
193,211
4.2%
23,757
50%
13.23
132.29
90,487
3.5%
Type 2 diabe es melli us
3.71
5.63
8.44
158,308
3.4%
28,102
52%
9.09
10.50
79,132
3.0%
O he diso de s o u ina y sys em
3.60
6.25
7.43
106,890
2.3%
17,099
63%
8.79
8.39
55,630
2.1%
O he ch onic obs uc i e pulmona y disease
3.88
5.64
7.57
120,058
2.6%
21,273
61%
8.01
8.95
53,145
2.0%
Shock, no elsewhe e classi ied
5.28
8.43
11.10
93,702
2.0%
11,118
47%
15.21
13.08
52,219
2.0%
Hea ailu e
3.69
6.01
6.16
89,461
1.9%
14,893
61%
8.54
6.70
44,278
1.7%
O he ce eb o ascula diseases
4.15
6.09
9.12
92,546
2.0%
15,190
46%
10.36
12.01
43,690
1.7%
O he diseases o diges i e sys em
4.01
5.64
8.22
101,288
2.2%
17,966
47%
9.13
10.89
43,439
1.7%
Choleli hiasis
3.64
4.94
8.79
88,039
1.9%
17,833
41%
9.40
12.31
42,584
1.6%
Ch onic ischemic hea disease
3.50
6.65
8.03
72,018
1.6%
10,829
56%
10.44
9.04
42,137
1.6%
Acu e myoca dial in a c ion
3.60
6.63
21.07
71,377
1.5%
10,769
63%
9.47
26.22
39,573
1.5%
O he sepsis
5.62
9.57
36.13
66,528
1.4%
6,953
52%
16.10
49.28
37,761
1.4%
Unknown and unspeci ied causes o mo bidi y
2.65
7.37
8.30
54,404
1.2%
7,383
69%
9.98
8.78
37,495
1.4%
Essen ial (p ima y) hype ension
3.21
4.73
7.13
60,199
1.3%
12,721
45%
8.33
9.41
29,327
1.1%
Unspeci ied diabe es melli us
4.71
6.28
6.72
65,272
1.4%
10,400
49%
10.49
7.55
29,215
1.1%
F ac u e o lowe leg, including ankle
5.03
7.29
9.10
53,434
1.2%
7,327
49%
12.14
10.90
25,783
1.0%
Cholecys i is
3.48
4.85
19.25
52,541
1.1%
10,832
41%
9.26
29.58
25,542
1.0%
O he diso de s o luid, elec oly e and acid-base
balance
3.50
6.42
19.35
44,100
1.0%
6,868
58%
9.77
24.99
24,791
1.0%
To al
2,119,828
46%
323,290
1,064,735
40.8%
No es: (1) medically app op ia e s ay o pa ien s wi hou ACSC como bidi ies; (2) h eshold o pa ien s wi h ACSC como bidi ies; (3) a e age leng h o s ay;
(4) s anda d de ia ion o leng h o s ay; (5) o al days a ibu ed o he condi ion; (6) pe cen age o o al days a ibu ed o he condi ion; (7) numbe o hospi aliza ions
a ibu ed o he condi ion; (8) pe cen age o hospi aliza ion om condi ion ha con ain excessi e days; (9) a e age leng h o s ay, in sample o hospi aliza ions ha
include excessi e days; (10) s anda d de ia ion o a e age leng h o s ay, in sample o hospi aliza ions ha include excessi e days; (11) excessi e days a ibu ed
o he condi ion; (12) excessi e days a ibu ed o he condi ion as a pe cen age o o al excessi e days.
Sou ce: Au ho s’ elabo a ion based on DATASUS hospi aliza ion da abase, 2019 and Mexico's Heal h Sec o Hospi al Discha ge Da abase, 2019.
28
a e gene a es wo da a eco ds ha canno be linked. Simila ly, pa ien s who ha e been
ans e ed o ano he es ablishmen o ha e unde gone a change in p ocedu e gene a e se e al
da a eco ds om which we canno calcula e he o al leng h o s ay. These da a ea u es
a i icially sho en he a e age leng h o s ay, hus unde es ima ing he numbe o excessi e days.
Second, ou analysis does no adjus o in-hospi al mo ali y, a ac o ha unca es he leng h o
s ay o some pa ien s. To add ess his issue in s udies o long hospi aliza ions, some au ho s d op
he obse a ions ha end wi h he dea h o he pa ien . Ou da a, howe e , show ha he
co ela ion be ween mo ali y and leng h o s ay is posi i e in some cases and nega i e in o he s.
Fo his eason, we make no co ec ions.
Thi d, ou da a do no allow us o p ecisely iden i y he clinically app op ia e leng h o s ay o each
hospi aliza ion. This would only be possible by analyzing pa ien s’ medical eco ds and assessing,
case by case, he op imal clinical leng h o s ay. Ou h eshold is adjus ed solely o he eason o
admission and assumes ha all hospi aliza ions due o a gi en eason should las he same. We
plan o collec complemen a y in o ma ion h ough de ailed analysis o medical eco ds o u u e
esea ch.
S udies u ilizing da a om medical eco ds o p o essional opinions gene ally ocus on smalle
samples d awn om a speci ic heal hca e ins i u ion. Thei esul s con i m ha , e en conside ing
he clinically jus i ied delays, long hospi aliza ions accoun o a la ge p opo ion o o al
hospi aliza ions. Fo example, in a s udy conduc ed in I aly, Bo e al. (2016) show ha 31.5% o
hospi aliza ions could be classi ied as long by clinical s anda ds, while Hendy e al. (2012)
es ima e his igu e o be a nea ly 50% o a London hospi al.
The h eshold we use o iden i y excessi e days is a bi a y and assumes ha c oss-s a e
a iabili y p o ides in o ma ion on he medically app op ia e s ay. This may be subjec o some
e o s. Fo example, a s a e migh epo a sho leng h o s ay o a ce ain condi ion because o
high in-hospi al mo ali y o because complex cases a e ans e ed o ano he hospi al. S a es
wi h a la ge p opo ion o uni e si y hospi als may epo longe s ays as hese ins i u ions ecei e
complex cases and pa o he s ay is dedica ed o aining (F ei as e al., 2012; Walke e al.,
2021).
In ou h eshold, he ela ionship be ween pa ien s’ ail y and leng h o s ay is only adjus ed o
he eason o hospi aliza ion and he exis ence o ACSC como bidi ies. Fo B azil, only
app oxima ely 5% o hospi aliza ions a e eco ded wi h ACSC como bidi ies. Fo Mexico, his
in o ma ion is no a ailable in ou da a. Due o his likely unde epo ing (o lack o da a on) ACSC
como bidi ies, we may be o e es ima ing he pe cen age o p olonged days o e all, and he pa
due o hospi al ine iciency.
In Annex 3, we p esen a sensi i i y analysis based on a di e en h eshold, o each condi ion,
se a he a e age leng h o s ay in he s a e wi h he median – ins ead o he minimum – o his
a e age. Wi h his al e na i e h eshold, 36% o hospi aliza ion days a e excessi e in bo h B azil
and Mexico. Simila o ou main analysis, almos 10% o hospi al days in B azil and 6% in Mexico
a e due o he lack o ehabili a ion, social and long- e m ca e se ices (Figu e A3.1, Annex 3).
Fou h, he componen ha measu es hospi al ine iciency may be a ec ed by ac o s we a e no
con olling o . A eaching hospi al, o example, may epo la ge a e age s ays e en wi h high
e iciency le els.
Finally, ou analysis is likely o unde es ima e he sha e o hospi al days ha can be sa ed h ough
ehabili a ion, social, and long- e m ca e se ices. Ou index o pos -discha ge ca e needs only
classi ies 36 o he mo e han 1,400 admission condi ions a ailable in B azil and Mexico and we
only conside pos -discha ge ca e needs ha a e di ec ly caused by he eason o admission. Due
29
o lack o da a, we a e unable o accoun o inc eased ail y esul ing om long hospi aliza ions
due o any condi ion, e en in pa ien s ha ha e no p e ious ca e needs. Fo al eady ail pa ien s,
a common si ua ion among olde pe sons, a ew days a he hospi al can c ea e signi ican loss
o au onomy. Gi en he p opo ion o olde pe sons wi h ca e needs in B azil (10.5%) and Mexico
(25.2%) (A anco, Iba a án and S ampini, 2022), his unde es ima ion may hus be la ge.
7. Conclusions and policy ecommenda ions
Ou esul s highligh he impo ance o de eloping pos -discha ge ca e se ices o educe
excessi e days o hospi aliza ions among olde people. A s ong ca e sys em ou side hospi als
allows pa ien s who a e clinically i o discha ge, bu s ill need ehabili a ion o suppo se ices,
o be eleased in a imely manne and wi hou comp omising hei wellbeing. As shown in his
pape , in es ing in long- e m ca e se ices can gene a e subs an ial sa ings in he heal hca e
sys em.
Inc easing he co e age and quali y o long- e m ca e se ices is s ill a challenge in he egion.
Cu en ly, he e is a limi ed supply o long- e m ca e sys ems and se ices in he egion, and
whe e a ailable, hey a e signi ican ly unde unded and ocused on he socioeconomically
ulne able popula ion (A anco e al., 2022). The wo coun ies ha a e he ocus o his s udy,
B azil and Mexico, a e discussing he c ea ion o ca e sys ems, which include he p o ision o
long- e m ca e (da Mo a Pe oni e al., 2023; López-O ega and A anco, 2019). The pa h owa ds
a long- e m ca e sys em will a y by coun y, bu he e a e common s eps ha coun ies need o
ake (Ca agna e al., 2019; Me dellín e al., 2018).
Fi s , eligibili y o se ices needs o be assessed h ough a scale ha e alua es ca e needs
(Oli ei a e al., 2022). Second, he assessmen needs o be ansla ed in o he de ini ion o a ca e
plan o e e y pe son. Thi d, coun ies need o decide how o inance he sys em. This may be
achie ed h ough gene al axa ion, social insu ance, co-paymen s, o a combina ion he eo . Each
inancing mechanism has s eng hs and weaknesses which need o be assessed by he coun ies
in o de o hem o selec a mechanism ha gua an ees he sys em’s inancial, social, and poli ical
iabili y (Fabiani e al., 2022).
Fou h, i is impo an o ensu e quali y o se ices. This equi es he es ablishmen and moni o ing
o quali y s anda ds, and aining and p o essionaliza ion o human esou ces who a e essen ial
o quali y se ice p o ision (A oyo e al., 2023; Fabiani, 2023; Villalobos Din ans e al., 2022).
Fi h, i is c i ical o c ea e s ong coo dina ion mechanisms be ween hospi als and long- e m ca e,
social, and ehabili a ion se ices. This is no easy, pa icula ly in coun ies like B azil, Mexico,
and many o he s in he La in Ame ican and Ca ibbean egion. In hese coun ies, heal hca e and
social se ices a e deli e ed by di e en ins i u ions, ha e sepa a e unding, di e en egula ions,
and di e en eligibili y ules.
In eg a ion equi es a undamen al pa adigm change, one ha places pe sons in he cen e o he
ca e deli e y sys em, and ha encompasses he adap a ion o bo h p ocesses and in as uc u es
(Lloyd-She lock e al. 2024; Albe son e al., 2022). The ole o a ca e manage o coo dina o , a
p o essional who wo ks closely wi h pa ien s (and hei amilies) o gua an ee he con inui y o
ca e ac oss all le els, has eme ged as a good p ac ice.
In La in Ame ica, he e a e some examples o sys ems ha ha e a emp ed o coo dina e social
and heal h ca e. In B azil, he p og ams Maio Cuidado in he ci y o Belo Ho izon e, and he
P og ama Acompanhan e de Idosos (PAI) in he municipali y o São Paulo, a e wo p omising
examples o imp o ed sociosani a y coo dina ion ha ha e he po en ial o acili a e olde people’s
ansi ion om he hospi al o hei pos -discha ge des ina ion (Lloyd-She lock e al., 2023; Lloyd-
30
She lock e al., 2024). Always in B azil, he p og am Melho em Casa o he Minis y o Heal h is
a la ge-scale, na ional e o o educe he numbe and he leng h o hospi aliza ions by p o iding
heal hca e a home (da Mo a Pe oni e al. 2023, Minis é io da Saúde do B asil 2024).
An e alua ion analysis o Belo Ho izon e’s p og am shows ha he leng h o hospi al s ays o
pa ien s ha belong o he Maio Cuidado p og am is 0.22 days sho e compa ed o pa ien s ha
do no belong o he p og am, gene a ing sa ings o app oxima ely US$100 pe admission (Lloyd-
She lock e al., 2024). The au ho s iden i y wo ea u es c ucial in explaining he p og am's
success: (i) he join de elopmen o he p og am by he Depa men o Heal h and he Depa men
o Social Assis ance, wi h bo h ins i u ions wo king in close collabo a ion; (ii) he c ea ion o a new
wo ke ca ego y – he amily ca e suppo wo ke s – who a e ully in eg a ed in o he local heal h
and social assis ance eams (Lloyd-She lock e al., 2024).
The bene i s o implemen ing a long- e m ca e sys em a e no limi ed o he po en ial sa ings ha
can be achie ed in he heal hca e sec o bu ex end o o he a eas (A anco e al., 2022),
Villalobos-Di ans, 2018). Such sys ems can ee up ime om amily ca egi e s, mos ly women,
p omo ing hei labo ma ke pa icipa ion o allowing hem o pu sue o he ac i i ies. These
sys ems can also con ibu e o he p o essionaliza ion o ca egi ing wo k, gene a ing mo e and
be e jobs in a sec o ha is cu en ly cha ac e ized by high eminiza ion, low le els o educa ion,
and subop imal wo king condi ions (Fabiani 2023, Villalobos-Di ans e al. 2022).
31
Re e ences
Albe son, E. M., Chuang, E., O'Mas a, B., Miake-Lye, I., Haley, L. A., & Pou a , N. (2022).
Sys ema ic e iew o ca e coo dina ion in e en ions linking heal h and social se ices o
high-u ilizing pa ien popula ions. Popula ion heal h managemen , 25(1), 73-85.
A anco, N., Iba a án, P., and S ampini, M. (2022). P e alence o ca e dependence among olde
pe sons in 26 La in Ame ican and he Ca ibbean coun ies. Technical No e IDB-TN-2470.
In e -Ame ican De elopmen Bank. h p://dx.doi.o g/10.18235/0004250
A anco, N., Bosch, M., S ampini, M., Azua a, O., Goyeneche, L., Iba a án, P., Oli ei a, D.,
Re ana To e, M.R., Sa edo , B. and To es, E. 2022. Aging in La in Ame ica and he
Ca ibbean: Social P o ec ion and Quali y o Li e o Olde Pe sons. IDB Monog aph 1009.
Washing on, DC: In e -Ame ican De elopmen Bank. h p://dx.doi.o g/10.18235/0004287
A oyo, E. A., P ie o, E. B., Co de o, L. F., Ma ín, M. L., F anco, L. C. L., Gómez, A. L., Benede i,
F., and Díaz-Veiga, P. (2023). Towa d he P o essionaliza ion o Ca egi e s: T aining and
Skills Needed o Long-Te m. In e -Ame ican De elopmen Bank, Technical No e IDB-TN
02717 h ps://doi.o g/10.18235/0005055
Be akis, K. D., Aza i, R., Helms, L. J., Callahan, E. J., & Robbins, J. A. (2000). Gende di e ences
in he u iliza ion o heal h ca e se ices. Jou nal o amily p ac ice, 49(2).
Bo, M., Fon e, G., Pi a o, F., Bone o, M., Comi, C., Gio gis, V., Ma chese, L., Isaia, G., Maggiani,
G., Fu no, E., Falcone, Y., and Isaia, G. C. (2016). P e alence o and ac o s associa ed wi h
p olonged leng h o s ay in olde hospi alized medical pa ien s. Ge ia ics & Ge on ology
In e na ional, 16(3), 314–321. h ps://doi.o g/10.1111/ggi.12471
Ca agna, G., A anco, N., Iba a án, P., Oli e i, M. L., Medellín, N., and S ampini, M. (2019). Age
wi h Ca e: Long- e m Ca e in La in Ame ica and he Ca ibbean. In e -Ame ican De elopmen
Bank, IDB-MG-74. h ps://doi.o g/10.18235/0001972
Came on, K. A., Song, J., Manheim, L. M., & Dunlop, D. D. (2010). Gende dispa i ies in heal h
and heal hca e use among olde adul s. Jou nal o women's heal h, 19(9), 1643-1650.
Ca ey, K. (2015). Measu ing he Hospi al Leng h o S ay/Readmission Cos T ade-O Unde a
Bundled Paymen Mechanism. Heal h Economics, 24(7), 790–802.
h ps://doi.o g/10.1002/hec.3061
Ca ey, M. R., She h, H., and B ai hwai e, R. S. (2005). A p ospec i e s udy o easons o
p olonged hospi aliza ions on a gene al medicine eaching se ice. Jou nal o Gene al
In e nal Medicine, 20(2), 108–115. h ps://doi.o g/10.1111/j.1525-1497.2005.40269.x
Cho, Y.-M., and Kwon, S. (2022). E ec s o public long- e m ca e insu ance on he medical
se ice use by olde people in Sou h Ko ea. Heal h Economics, Policy and Law, 1–18.
h ps://doi.o g/10.1017/S174413312200024X
Cos a-Fon , J., Jimenez-Ma in, S., and Vilaplana, C. (2018). Does long- e m ca e subsidiza ion
educe hospi al admissions and u iliza ion? Jou nal o Heal h Economics, 58, 43–66.
h ps://doi.o g/10.1016/j.jhealeco.2018.01.002
Cu y, N. (2006). P e en i e Social Ca e: Is i cos e ec i e? Kings Fund.
h ps://www.kings und.o g.uk/si es/de aul / iles/p e en i e-social-ca e-wanless-backg ound-
pape -na asha-cu y2006.pd
32
Da Mo a Pe oni, F., G ucho ski Ve íssimo, L., Goes Shiba a, L., A anco, N. (2023).
En ejecimien o y a ención a la dependencia en B asil. In e -Ame ican De elopmen Bank.
No a Técnica IDB-TN-02677. h p://dx.doi.o g/10.18235/0004792
Din ans, P. V., Oli ei a, D., and S ampini, M. (2022). Es imación de las necesidades de ecu sos
humanos pa a la a ención a las pe sonas mayo es con dependencia de cuidados en Amé ica
La ina y el Ca ibe. In e -Ame ican De elopmen Bank. No a Técnica IDB-TN-02556.
h ps://doi.o g/10.18235/0004487
Fabiani, B. (2023). Ca ing o Ca egi e s: The Landscape o Paid Ca e Wo k in La in Ame ica
and he Ca ibbean. In e -Ame ican De elopmen Bank, Technical No e IDB-TN 02783
h ps://doi.o g/10.18235/0005147
Fabiani, B., Cos a-Fon , J., A anco, N., S ampini, M., and Iba a án, P. (2022). Funding Op ions
o Long-Te m Ca e Se ices in La in Ame ica and he Ca ibbean. In e -Ame ican
De elopmen Bank, Technical No e IDB-TN 02473 h ps://doi.o g/10.18235/0004306
Falcone, D., Bolda, E., and Leak, S. C. (n.d.). Wai ing o Placemen : An Explo a o y Analysis o
De e minan s o Delayed Discha ges o Elde ly Hospi al Pa ien s.
F ei as, A., Sil a-Cos a, T., Lopes, F., Ga cia-Lema, I., Teixei a-Pin o, A., B azdil, P., and Cos a-
Pe ei a, A. (2012). Fac o s in luencing hospi al high leng h o s ay ou lie s. BMC Heal h
Se ices Resea ch, 12(1), 265. h ps://doi.o g/10.1186/1472-6963-12-265
Gaughan, J., G a elle, H., and Siciliani, L. (2015). Tes ing he Bed-Blocking Hypo hesis: Does
Nu sing and Ca e Home Supply Reduce Delayed Hospi al Discha ges? Heal h Economics,
24(S1), 32–44. h ps://doi.o g/10.1002/hec.3150
Hassan, M., Tuckman, H. P., Pa ick, R. H., Koun z, D. S., and Kohn, J. L. (2010). Hospi al leng h
o s ay and p obabili y o acqui ing in ec ion. In e na ional Jou nal o Pha maceu ical and
Heal hca e Ma ke ing, 4(4), 324–338. h ps://doi.o g/10.1108/17506121011095182
Hauck, K., and Zhao, X. (2011). How dange ous is a day in hospi al? A model o ad e se e en s
and leng h o s ay o medical inpa ien s. Medical Ca e, 49(12), 1068–1075.
h ps://doi.o g/10.1097/MLR.0b013e31822e b09
Ins i u e o Heal h Me ics and E alua ion (IHME), 2020. Global Bu den o Disease Collabo a i e
Ne wo k. Global Bu den o Disease S udy 2019 (GBD 2019). Sea le, Uni ed S a es.
Hendy, P., Pa el, J., Ko dbacheh, T., Laska , N., and Ha bo d, M. (2012). In-dep h analysis o
delays o pa ien discha ge: A me opoli an eaching hospi al expe ience. Clinical Medicine,
12(4), 320–323. h ps://doi.o g/10.7861/clinmedicine.12-4-320
Holmås, T. H., Islam, M. K., and Kje s ad, E. (2013). In e dependency be ween social ca e and
hospi al ca e: The case o hospi al leng h o s ay. Eu opean Jou nal o Public Heal h, 23(6),
927–933. h ps://doi.o g/10.1093/eu pub/cks171
Holmås, T. H., Kam ul Islam, M., and Kje s ad, E. (2013). Be ween wo beds: Inapp op ia ely
delayed discha ges om hospi als. In e na ional Jou nal o Heal h Ca e Finance and
Economics, 13(3–4), 201–217. h ps://doi.o g/10.1007/s10754-013-9135-4
Hyun, K.-R., Kang, S., and Lee, S. (2014). Does long- e m ca e insu ance a ec he leng h o s ay
in hospi als o he elde ly in Ko ea?: A di e ence-in-di e ence me hod. BMC Heal h Se ices
Resea ch, 14(1), 630. h ps://doi.o g/10.1186/s12913-014-0630-1
Kozma, C. M., Dickson, M., Rau , M. K., Mody, S., Fishe , A. C., Schein, J. R., and Mackowiak,
J. I. (2010). Economic bene i o a 1-day educ ion in hospi al s ay o communi y-acqui ed
33
pneumonia (CAP). Jou nal o Medical Economics, 13(4), 719–727.
h ps://doi.o g/10.3111/13696998.2010.536350
K ell, R. W., Gi o i, M. E., and Dimick, J. B. (2014). Ex ended Leng h o S ay A e Su ge y:
Complica ions, Ine icien P ac ice, o Sick Pa ien s? JAMA Su ge y, 149(8), 815–820.
h ps://doi.o g/10.1001/jamasu g.2014.629
Landei o, F., Leal, J., and G ay, A. M. (2016). The impac o social isola ion on delayed hospi al
discha ges o olde hip ac u e pa ien s and associa ed cos s. Os eopo osis In e na ional: A
Jou nal Es ablished as Resul o Coope a ion be ween he Eu opean Founda ion o
Os eopo osis and he Na ional Os eopo osis Founda ion o he USA, 27(2), 737–745.
h ps://doi.o g/10.1007/s00198-015-3293-9
Landei o, F., Robe s, K., G ay, A. M., and Leal, J. (2019). Delayed Hospi al Discha ges o Olde
Pa ien s: A Sys ema ic Re iew on P e alence and Cos s. The Ge on ologis , 59(2), e86–e97.
h ps://doi.o g/10.1093/ge on /gnx028
Lenzi, J., Monga di, M., Rucci, P., Ruscio, E. D., Vizioli, M., Randazzo, C., Toschi, E., Ca ado i,
T., and Fan ini, M. P. (2014). Sociodemog aphic, clinical and o ganisa ional ac o s
associa ed wi h delayed hospi al discha ges: A c oss-sec ional s udy. BMC Heal h Se ices
Resea ch, 14, 128. h ps://doi.o g/10.1186/1472-6963-14-128
Lewis, R., and Glasby, J. (2006). Delayed discha ge om men al heal h hospi als: Resul s o an
English pos al su ey. Heal h and Social Ca e in he Communi y, 14(3), 225–230.
h ps://doi.o g/10.1111/j.1365-2524.2006.00614.x
Lingsma, H. F., Bo le, A., Middle on, S., Kie i , J., S eye be g, E. W., and Ma ang- an de Mheen,
P. J. (2018). E alua ion o hospi al ou comes: The ela ion be ween leng h-o -s ay,
eadmission, and mo ali y in a la ge in e na ional adminis a i e da abase. BMC Heal h
Se ices Resea ch, 18(1), 116. h ps://doi.o g/10.1186/s12913-018-2916-1
Lloyd-She lock, P., Fialho de Ca alho, P., Giacomin, K., and Sempé, L. (2023). Add essing
p essu es on heal h se ices in Belo Ho izon e, B azil h ough communi y-based ca e o
poo olde people: A quali a i e s udy. The Lance Regional Heal h - Ame icas, 27, 100619.
h ps://doi.o g/10.1016/j.lana.2023.100619
Lloyd-She lock, P., Giacomin, K., Fialho de Ca alho, P., Sempé, L. (2024). P og ama Maio
Cuidado: An In eg a ed Communi y-Based In e en ion on Ca e o Olde People. Technical
no e 2889. In e -Ame ican De elopmen Bank. h p://dx.doi.o g/10.18235/0005535
López-O ega, M., A anco, N. (2019). En ejecimien o y a ención a la dependencia en México.
In e -Ame ican De elopmen Bank. No a Técnica IDB-TN-1614
h p://dx.doi.o g/10.18235/0001826
Makowsky, M. D., and Klein, E. Y. (2018). Iden i ying he ela ionship be ween leng h o hospi al
s ay and he p obabili y o eadmission. Applied Economics Le e s, 25(6), 375–380.
h ps://doi.o g/10.1080/13504851.2017.1324605
Ma il-Ga za, B. A., Belaunza án-Zamudio, P. F., Gulias-He e o, A., Zuñiga, A. C., Ca o-Vega,
Y., Ke shenobich-S alnikowi z, D., and Si uen es-Oso nio, J. (2018). Risk ac o s associa ed
wi h p olonged hospi al leng h-o -s ay: 18-yea e ospec i e s udy o hospi aliza ions in a
e ia y heal hca e cen e in Mexico. PLOS ONE, 13(11), e0207203.
h ps://doi.o g/10.1371/jou nal.pone.0207203
Ma us-López, M. (2023). “Análisis del impac o y aho o en salud de un sis ema de cuidados a la
dependencia en Chile”. Unpublished documen .
34
Medellín, N., Iba a án, P., and S ampini, M. (2018). Cua o elemen os pa a diseña un sis ema
de cuidados. Technical no e 1438. In e -Ame ican De elopmen Bank.
h ps://doi.o g/10.18235/0001129
Micalle , A., Bu igieg, S. C., Tomaselli, G., and Ga g, L. (2020). De ining Delayed Discha ges o
Inpa ien s and Thei Impac in Acu e Hospi al Ca e: A Scoping Re iew. In e na ional Jou nal
o Heal h Policy and Managemen , 11(2), 103–111. h ps://doi.o g/10.34172/ijhpm.2020.94
Minis é io da Saúde B asil (n.d). Depa amen o de In o má ica do Sis ema Único de Saúde,
DATASUS. A ailable a : h ps://da asus.saude.go .b /sob e-o-da asus/
Minis é io da Saúde B asil (2024). Melho em Casa inclui equipes de eabili ação e em no as
di e izes pa a ges o es. A ailable a : h ps://www.go .b /saude/p -
b /assun os/no icias/2024/ma co/melho -em-casa-inclui-equipes-de- eabili acao-e- em-
no as-di e izes-pa a-ges o es
Misky, G., Wald, H., and Coleman, E. (2010). Pos -Hospi aliza ion T ansi ions: Examining he
E ec s o Timing o P ima y Ca e P o ide Follow-Up. Jou nal o Hospi al Medicine : An
O icial Publica ion o he Socie y o Hospi al Medicine, 5, 392–397.
h ps://doi.o g/10.1002/jhm.666
Moo e, L., Cisse, B., Ba omen Kuimi, B. L., S el ox, H. T., Tu geon, A. F., Lauzie , F., Clémen ,
J., and Bou geois, G. (2015). Impac o socio-economic s a us on hospi al leng h o s ay
ollowing inju y: A mul icen e coho s udy. BMC Heal h Se ices Resea ch, 15(1), 285.
h ps://doi.o g/10.1186/s12913-015-0949-2
Na ional Heal h Se ice (NHS) England (n.d). “Reducing leng h o s ay”. A ailable a :
h ps://www.england.nhs.uk/u gen -eme gency-ca e/ educing-leng h-o -s ay/
Oli a es-Ti ado, D. P. (n.d.). Hospi alización Social en Adul os Mayo es en el Sis ema de Salud
de Chile.
Oli ei a, D., León-Moncada, S., and Te a, F. (2022). El uso de Ba emos de Valo ación de la
Funcionalidad y de la Dependencia de cuidados en pe sonas mayo es: P ác icas, a ances
y di ecciones u u as. h ps://doi.o g/10.18235/0004280
Pe elman, J., and Closon, M.-C. (2011). Impac o socioeconomic ac o s on in-pa ien leng h o
s ay and hei consequences in pe case hospi al paymen sys ems. Jou nal o Heal h
Se ices Resea ch and Policy, 16(4), 197–202. h ps://doi.o g/10.1258/jhs p.2011.010047
Picone, G., Ma k Wilson, R., and Chou, S.-Y. (2003). Analysis o hospi al leng h o s ay and
discha ge des ina ion using haza d unc ions wi h unmeasu ed he e ogenei y. Heal h
Economics, 12(12), 1021–1034. h ps://doi.o g/10.1002/hec.800
Rao, K. D., Vecino O iz, A. I., Robe on, T., Lopez He nandez, A., & Noonan, C. (2022). Fu u e
Heal h Spending and T ea men Pa e ns in La in Ame ica and he Ca ibbean: Heal h
Expendi u e P ojec ions & Scena io Analysis . In e -Ame ican De elopmen Bank.
Rauh, S. S., Wadswo h, E., and Weeks, W. B. (2010). The ixed-cos dilemma: Wha coun s
when coun ing cos - educ ion e o s? A hospi al’s ixed cos s a e a eali y ha can make he
idea o achie ing sa ings by educing leng h o s ay illuso y. Heal hca e Financial
Managemen , 64(3), 60–64.
Rod igues, L. P., de Oli ei a Rezende, A. T., Delpino, F. M., Mendonça, C. R., Noll, M., Nunes,
B. P., de Oli ie a, C., and Sil ei a, E. A. (2022). Associa ion be ween mul imo bidi y and
hospi aliza ion in olde adul s: Sys ema ic e iew and me a-analysis. Age and Ageing, 51(7),
a ac155. h ps://doi.o g/10.1093/ageing/a ac155
35
Rojas‐Ga cía, A., Tu ne , S., Pizzo, E., Hudson, E., Thomas, J., and Raine, R. (2018). Impac
and expe iences o delayed discha ge: A mixed‐s udies sys ema ic e iew. Heal h
Expec a ions : An In e na ional Jou nal o Public Pa icipa ion in Heal h Ca e and Heal h
Policy, 21(1), 41–56. h ps://doi.o g/10.1111/hex.12619
Siddique, S. M., Tip on, K., Leas, B., G eysen, S. R., Mull, N. K., Lane-Fall, M., McShea, K., and
Tsou, A. Y. (2021). In e en ions o Reduce Hospi al Leng h o S ay in High- isk Popula ions:
A Sys ema ic Re iew. JAMA Ne wo k Open, 4(9), e2125846.
h ps://doi.o g/10.1001/jamane wo kopen.2021.25846
S one, K., Zwiggelaa , R., Jones, P., and Mac Pa haláin, N. (2022). A sys ema ic e iew o he
p edic ion o hospi al leng h o s ay: Towa ds a uni ied amewo k. PLOS Digi al Heal h, 1(4),
e0000017. h ps://doi.o g/10.1371/jou nal.pdig.0000017
Tahe i, P. A., Bu z, D. A., and G een ield, L. J. (2000). Leng h o s ay has minimal impac on he
cos o hospi al admission11No compe ing in e es s decla ed. Jou nal o he Ame ican
College o Su geons, 191(2), 123–130. h ps://doi.o g/10.1016/S1072-7515(00)00352-5
Toh, H. J., Lim, Z. Y., Yap, P., and Tang, T. (2017). Fac o s associa ed wi h p olonged leng h o
s ay in olde pa ien s. Singapo e Medical Jou nal, 58(3), 134–138.
h ps://doi.o g/10.11622/smedj.2016158
Villalobos Din ans, P. (2018). Long- e m ca e sys ems as social secu i y: The case o Chile.
Heal h Policy and Planning, 33(9), 1018–1025. h ps://doi.o g/10.1093/heapol/czy083
Walke , R. J., Segon, A., Good, J., Naga ally, S., Gup a, N., Le ine, D., Neune , J., and Egede,
L. E. (2021). Di e ences in leng h o s ay by eaching eam s a us in an academic medical
cen e in he Midwes e n Uni ed S a es. Hospi al P ac ice (1995), 49(2), 119–126.
h ps://doi.o g/10.1080/21548331.2021.1882238
Walsh, T., Onega, T., and Mackenzie, T. (2014). Va ia ion in leng h o s ay wi hin and be ween
hospi als. Jou nal o Hospi al Adminis a ion, 3(4), 53. h ps://doi.o g/10.5430/jha. 3n4p53
Yang, S., Zhou, M., Liao, J., Ding, X., Hu, N., and Kuang, L. (2022). Associa ion be ween P ima y
Ca e U iliza ion and Eme gency Room o Hospi al Inpa ien Se ices U iliza ion among he
Middle-Aged and Elde ly in a Sel -Re e al Sys em: E idence om he China Heal h and
Re i emen Longi udinal S udy 2011–2018. In e na ional Jou nal o En i onmen al Resea ch
and Public Heal h, 19(19), 12979. h ps://doi.o g/10.3390/ije ph191912979
Zhao, E. J., Yelu u, A., Manjuna h, L., Zhong, L. R., Hsu, H.-T., Lee, C. K., Wong, A. C., Ab amian,
M., Manella, H., S ec, D., and Shieh, L. (2018). A long wai : Ba ie s o discha ge o long
leng h o s ay pa ien s. Pos g adua e Medical Jou nal, 94(1116), 546–550.
h ps://doi.o g/10.1136/pos g admedj-2018-135815
36
Annex 1. Findings om in e iews and ocus g oup discussions
We complemen ou analysis wi h quali a i e in o ma ion o ge a be e unde s anding o he
causes and consequences o long hospi aliza ions. The da a we e collec ed h ough in e iews
(3 pa icipan s) and ocus g oup discussions (6 pa icipan s) wi h heal h expe s in he egion
(Mexico, A gen ina, and Colombia) be ween June and Oc obe 2023. The da a we e analyzed
h ough he me hod o hema ic analysis. This induc i e app oach allowed us o iden i y majo
hemes and pa e ns and concep ualize he in o ma ion in o he s ages o p e-hospi al, in-hospi al,
and pos -hospi al ca e. The con e sa ions p o ided c ucial insigh s in o he di e en inal
componen s along he con inuum-o -ca e e e ed o in ou analysis.
The indings om he in e iews and ocus g oup discussions emphasize he impo ance o
app op ia e p ima y heal hca e and social se ices o p e en ing hospi aliza ions and
s eng hening a pe son’s heal h s a us. The pa icipan s highligh ed ha he inabili y o ensu e
imely access and use o such se ices nega i ely a ec s a pe son’s heal h condi ions, ul ima ely
inc easing he isk o a highe leng h o s ay in he hospi al.
“El cuidado p e io a pisa u gencias/hospi al es undamen al. El que el pacien e acuda de
mane a egula a sus isi as de medicina amilia ambién es undamen al.” – P2, ocus
g oup
“No podemos habla de la demanda de u gencias y de la a ención hospi ala ia si no
hablamos de las en e medades c ónicas que no se p o ocolizan en el manejo ambula o io.”
– P2, ocus g oup
“No enemos plan ampliado de inmunizaciones en pe sonas mayo es, la única acuna
g a ui a es la de in luenza y su cobe u a es muy baja; en onces si no enemos es a egias
de p e ención como la acunación con a neumococo, os e ina, he pes zos e ... es amos
exponiendo a la población mayo , en especial a la más ágil, a hospi alizaciones
ecu en es.” – P1, in e iew
“El hecho de que el adul o mayo no enga sopo e pa a busca a ención médica opo una,
lo hace que llegue al hospi al en una ase más g a e de la en e medad.” – P6, ocus g oup
“La es ancia hospi ala ia la de e mina la es abilidad del pacien e, es deci , su uncionalidad.
Cómo es á uncionando él como indi iduo en cues ión de condiciones mo o as, en cues ión
de como bilidades.” – P4, ocus g oup
“Cualquie cosa que cambia d amá icamen e la uncionalidad de una pe sona (…) es de
iesgo pa a una es adía hospi ala ia mayo .” – P2, in e iew
“No enemos u as de de ección emp ana de os eopo osis y iesgo de caída y ac u a; las
ci ugías o opédicas gene an hospi alizaciones más p olongadas en pe sonas mayo es.” –
P1, in e iew
Mo eo e , inadequa e ehabili a ion and long- e m ca e, including in o ma ion and suppo , and
long hospi al s ays may inc ease he isk o ehospi aliza ion. The ollowing obse a ions
exempli y his poin :
“[Después de una hospi alización,] un ema impo an e es la o ien ación nu icional
adap ada a la pe sona. O o aspec o impo an e es la educación en salud, es deci , un
pacien e con incluso 10 años con hipe ensión no en iende su en e medad, no hemos
sabido in o ma le al espec o.” – P2, ocus g oup
37
“Después de los 70 años en un solo día de hospi alización, si no me mue o, puedo pe de
has a el 3% de la masa muscula o al lo que a a gene a g andes p oblemas de mo ilidad,
dependencia uncional y sob eca ga a los cuidado es amilia es.” – P1, in e iew
The discussions also showed ha hospi al cha ac e is ics and ine iciencies can ha e impo an
epe cussions o a pa ien ’s leng h o s ay in he hospi al. Acco ding o he pa icipan s, he
absence o p o ocols, esou ces, and knowledge a a ious le els can lead o delays in p ocesses
a he beginning, du ing, and a he end o a hospi aliza ion. Wi h ega ds o elde ly pe sons,
common issues seem o pa icula ly e ol e a ound coo dina ion be ween he di e en le els o
heal h ca e, sca ci y o aining in ge ia ics, especially among doc o s wo king in eme gency
se ices, insu icien esou ces o imely and app op ia e ea men o pa ien s, and absence o
clea discha ge p ocedu es.
“(…) La al a de coo dinación es un asun o mayo .” – P1, ocus g oup
“Hay un e a do eno me del médico amilia pa a en ia a los pacien es a 2º o 3º ni el.” –
P1, ocus g oup
“Si el médico amilia no es á capaci ado pa a a ende a los pacien es ge iá icos con
en e medades como diabe es melli us e hipe ensión, las más comunes, cuando llega al 2º
ni el el médico especialis a iene que in e na lo y el pacien e llega con odas las pa ologías
ag a adas y complicaciones.” – P3, ocus g oup
“No exis en p o ocolos de a ención humanizada y di e encial a pe sonas mayo es [en el
caso de u gencias] po la can idad de pacien es que eciben; es o hace que los p ocesos
de admisión hospi ala ia sean la gos y que (…) en los se icios de u gencias se compliquen
o adquie an gé menes opo unis as que hacen que se complique el cuad o inicial, sin
con a odo lo que ocu e en pe sonas mayo es con de e io o cogni i o que gene an
episodios deli an es y e minan siendo inmo ilizados an o ísica como
a macológicamen e.” – P1, in e iew
“Los médicos de U gencias y Hospi alización en su mayo ía no han ecibido capaci ación
en ge ia ía ni cuen an con médicos ge ia as (…), po lo que e minan in e -consul ando a
a ios especialis as (…), lo cual lle a a oma excesi a de labo a o ios y demo as en el
p oceso.” – P1, in e iew
“No enemos ge ia as en los se icios de u gencias en odos los hospi ales como se ía lo
ideal, en onces se a e asando la a ención po que el médico gene al al ez iene miedo
de abo da al pacien e adul o mayo . Si se capaci a a al pe sonal ayuda ía a que la a ención
ue a más e icien e y opo una.” – P5, ocus g oup
“En el ema de ac u a de cade a, e asan mucho desde el diagnós ico y en las á eas
hospi ala ias, la ci ugía. Es o es a ibuible a la al a de conocimien o de que es e
padecimien o es una u gencia y en gene al los médicos espe an a que el pacien e es é lo
más es able posible pa a ope a lo.” – P6, ocus g oup
“Hay hospi ales de 2º ni el que no ienen ecu sos y ienen que espe a a que el 3º ni el
les dé un espacio pa a el diagnós ico, el pacien e puede es a has a 10 días espe ando el
diagnós ico, en luga de ecibi el a amien o.” – P1, ocus g oup
“La al a de insumos necesa ios. Po ejemplo, los pacien es de ac u as de cade a se
quedan mucho iempo hospi alizados po que no hay la ue ca o el o nillo o el medicamen o
necesa io.” – P1, ocus g oup
“Pocos hospi ales ienen p o ocolos de "al a emp ana", po lo que adminis a i amen e
exis en muchas ba e as y p ocesos (…).” – P2, in e iew
