A i icial In elligence and Managemen Con ol
in Hospi als: A C i ical Re iew and Concep ual
In eg a ion
Yous a Nassou
PhD in Managemen Sciences, ENCG de Tange , Mo occo
Kama Moukadem
PhD in Managemen Sciences, ENCG de Tange , Mo occo
Abs ac
This a icle p o ides a c i ical e iew o he li e a u e on he in eg a ion o
a i icial in elligence (AI) in o hospi al managemen con ol sys ems (MCS).
The apid de elopmen o AI in heal hca e has c ea ed signi ican
oppo uni ies o enhancing o ecas ing, op imizing esou ce alloca ion, and
imp o ing cos ing accu acy. Despi e hese ad ances, he adop ion o AI
wi hin inancial and s a egic con ol amewo ks emains limi ed.
The e iew syn hesizes empi ical indings and concep ual con ibu ions
published be ween 2020 and 2025. I examines applica ions such as
p edic i e analy ics o pa ien lows, digi al win simula ions o esou ce
op imiza ion, and Time-D i en Ac i i y-Based Cos ing (TDABC) o mo e
p ecise inancial moni o ing. These indings a e analyzed in ela ion o
es ablished managemen con ol amewo ks, including he Balanced
Sco eca d (BSC) and Pe o mance Managemen Sys ems (PMS).
The li e a u e e eals a s ong con e gence on AI’s echnical po en ial,
pa icula ly in imp o ing p edic i e accu acy and ope a ional e iciency.
Howe e , di e gences emain ega ding he ex en o which hese ools a e
in eg a ed in o go e nance and inancial con ol sys ems. Me hodological
limi a ions, including eliance on single-si e s udies, absence o causal
designs, and agmen ed app oaches o p edic ion and cos ing, es ic he
s eng h o cu en e idence.
This a icle p oposes a concep ual model ha links AI capabili ies wi h MCS
and highligh s go e nance as a key mode a o o ou comes. I emphasizes
ha AI does no eplace managemen con ol bu augmen s i , mo ing om
e ospec i e epo ing owa d p oac i e and simula ion-based go e nance.
By iden i ying heo e ical gaps and ou lining a u u e esea ch agenda, he
s udy con ibu es o a be e unde s anding o how AI can suppo inancial
sus ainabili y and s a egic decision-making in hospi als.
In oduc ion
Hospi als oday ace g owing inancial and
o ganiza ional p essu es due o ising heal hca e cos s,
ch onic disease p e alence, and esou ce cons ain s
[31]. In his en i onmen , managemen con ol sys ems
(MCS) ha e become essen ial ools o aligning s a egy
wi h pe o mance measu emen , budge ing, and
esou ce alloca ion. [10] emphasized h ough he
Balanced Sco eca d ha managemen con ol is no
limi ed o inancial indica o s bu mus in eg a e
pa ien , p ocess, and lea ning pe spec i es o imp o e
o ganiza ional pe o mance. Ye , empi ical s udies
show ha many hospi als s ill s uggle wi h agmen ed
da a and lack he p edic i e ools needed o e ec i e
cos con ol [13].
The de elopmen o a i icial in elligence (AI) has
gene a ed s ong in e es o hospi al managemen . [8]
demons a ed, using a la ge hospi al da ase , ha
machine lea ning models can signi ican ly imp o e he
accu acy o leng h o s ay (LoS) p edic ions, suppo ing
mo e eliable planning o bed occupancy and s a ing.
[16] in oduced he concep o he digi al win in
heal hca e sys ems, showing how simula ion ools can
an icipa e pa ien lows and op imize ope a ing oom
schedules. On he inancial side, [13] con i med ha
Time-D i en Ac i i y-Based Cos ing (TDABC) p o ides
Mo e In o ma ion
How o ci e his a icle: Nassou Y,
Moukadem K. A i icial In elligence
and Managemen Con ol in
Hospi als: A C i ical Re iew and
Concep ual In eg a ion. Eu J Med
Heal h Res, 2025;3(3):218-25.
DOI: 10.59324/ejmh .2025.3(3).33
Keywo ds:
A i icial in elligence,
Hospi al managemen con ol,
Balanced Sco eca d,
TDABC,
Heal hca e go e nance.
This wo k is licensed unde a
C ea i e Commons A ibu ion 4.0
In e na ional License. The license
pe mi s un es ic ed use,
dis ibu ion, and ep oduc ion in
any medium, on he condi ion ha
use s gi e exac 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 license, and indica e i
hey made any changes.
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
219
mo e accu a e cos in o ma ion in heal hca e and
a gued ha digi al echnologies can acili a e i s
implemen a ion by au oma ing ime da a collec ion.
Howe e , he li e a u e also highligh s pe sis en
ba ie s. [17], in hei sys ema ic e iew, iden i ied
ecu ing obs acles such as low da a quali y, lack o
in e ope abili y, and p o essionals’ esis ance o
algo i hmic ools. [14] u he unde lined ha
o ganiza ional cul u e and go e nance s ongly
in luence whe he AI ini ia i es in hospi als succeed o
ail. Beyond hese ope a ional and cul u al ba ie s,[5]
no ed ha mos AI applica ions in heal hca e ha e
ocused on clinical decision suppo and ope a ional
op imiza ion, lea ing he in eg a ion o AI in o
managemen con ol and inancial go e nance
unde explo ed.
The objec i e o his a icle is he e o e o p o ide a
c i ical e iew o he li e a u e on AI in hospi al
managemen con ol. Ou app oach is in eg a i e and
c i ical, syn hesizing empi ical indings, concep ual
con ibu ions, and managemen amewo ks o
highligh s eng hs, weaknesses, and blind spo s.
Building upon classical amewo ks such as he
Balanced Sco eca d [10] and TDABC [9], while
inco po a ing ecen indings om AI applica ions in
heal hca e [8,16], we p opose a concep ual model ha
links AI capabili ies wi h hospi al MCS.
Theo e ical F amewo k: Hospi al Managemen
Con ol Sys ems
The ole o MCS in hospi als is o p o ide eliable
mechanisms o aligning clinical ac i i ies wi h s a egic
and inancial objec i es. Acco ding o [3], managemen
con ol can be de ined as he p ocess by which
manage s in luence membe s o an o ganiza ion o
implemen s a egies. [10] expanded his pe spec i e
wi h he Balanced Sco eca d (BSC), which in eg a es
ou dimensions— inancial, cus ome , in e nal
p ocesses, and lea ning and g ow h— o cap u e bo h
economic e iciency and se ice quali y. In heal hca e,
he BSC has been widely adop ed as a ool o imp o e
o ganiza ional pe o mance. [2], o example, e iewed
BSC implemen a ions in hospi als and ound ha i
s eng hens pe o mance moni o ing, pa icula ly
when inancial and non- inancial indica o s a e
combined.
Ano he undamen al pilla o hospi al MCS is cos
accoun ing. T adi ional sys ems based on s anda d
cos ing o depa men al alloca ions ha e o en been
c i icized o hei inabili y o e lec he eal
consump ion o esou ces in clinical pa hways. [9]
p oposed he Time-D i en Ac i i y-Based Cos ing
(TDABC) me hod as a mo e accu a e and adap able
app oach. [13] applied TDABC in heal hca e and
demons a ed ha i allows o a mo e p ecise
alloca ion o cos s by p ocedu es and se ices,
p o iding manage s wi h ac ionable in o ma ion o
educe ine iciencies and suppo p icing decisions.
In addi ion o cos ing, budge ing and a iance analysis
emain cen al ools o hospi al con ol. Howe e ,
se e al s udies poin o hei limi a ions in highly
dynamic heal hca e en i onmen s. [1], o ins ance,
showed ha igid budge ing p ac ices in Swedish
hospi als o en con lic ed wi h he need o lexibili y in
esponding o pa ien demand, leading o a gap
be ween s a egic objec i es and ope a ional eali ies.
This inding emains ele an oday, as hospi als
con inue o ace unp edic able a ia ions in admissions
and ea men cos s.
Mo e b oadly, managemen con ol in hospi als mus
be unde s ood as pa o a pe o mance managemen
sys em (PMS). [6] concep ualized PMS as a holis ic
amewo k ha links objec i es, measu es, a ge s, and
eedback loops. In he heal hca e sec o , s udies
con i m he alue o such in eg a ed sys ems. [7], o
example, demons a ed ha when PMS a e linked o
ac i i y-based cos ing da a, hey imp o e anspa ency
and decision-making a he depa men al le el.
Simila ly, [2] emphasized ha balanced sco eca ds,
when in eg a ed in o PMS, enhance accoun abili y by
combining e iciency me ics wi h quali y-o -ca e
indica o s.
Con ibu ions o A i icial In elligence in Heal hca e
and Links o Managemen Con ol
A i icial in elligence (AI) has eme ged as a
ans o ma i e o ce in heal hca e, o e ing ools o
imp o e p edic ion, op imiza ion, and cos
measu emen . While mos applica ions ha e been
s udied in clinical o ope a ional con ex s, hei
implica ions o managemen con ol sys ems (MCS)
a e inc easingly ecognized.
P edic i e analy ics o pa ien low and cos s
A majo applica ion o AI in hospi als is p edic ing
pa ien admissions and leng h o s ay (LoS). [8]
demons a ed ha machine lea ning models
signi ican ly ou pe o m adi ional s a is ical me hods
in p edic ing LoS, enabling hospi als o be e an icipa e
esou ce needs. Imp o ed o ecas ing o LoS and
admissions no only suppo s clinical planning bu also
p o ides managemen con olle s wi h mo e eliable
inpu s o capaci y budge ing and cos a iance
analysis.
[15], in a sys ema ic e iew o LoS p edic ion s udies,
con i med ha AI models (especially g adien boos ing
and deep lea ning) consis en ly achie e highe accu acy
compa ed o logis ic eg ession. They concluded ha
hese ools could be in eg a ed in o hospi al
in o ma ion sys ems o enhance bo h clinical
scheduling and inancial planning.
Op imiza ion o sca ce esou ces
Beyond p edic ion, AI also plays a c ucial ole in
op imizing he alloca ion o esou ces such as beds,
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
220
su gical hea e s, and nu sing s a . [16] highligh ed he
po en ial o digi al win models, which eplica e
hospi al p ocesses i ually and allow manage s o
simula e di e en esou ce alloca ion scena ios. This
ype o simula ion p o ides managemen con olle s
wi h insigh s in o he cos implica ions o al e na i e
scheduling s a egies.
Simila ly, [12] p esen ed he Digi al Supply Chain Twin
pa adigm, showing how simula ion-based decision
amewo ks can imp o e esilience and e iciency.
Al hough hei s udy ocused on supply chains, he
unde lying logic has s ong pa allels wi h hospi al
logis ics, pa icula ly in pha maceu icals and
equipmen managemen , which a e key cos cen e s.
Cos ing and alue measu emen wi h AI suppo
Accu a e cos measu emen is cen al o hospi al
managemen con ol. [13] showed ha Time-D i en
Ac i i y-Based Cos ing (TDABC) in heal hca e p o ides
mo e p ecise cos alloca ions pe pa ien pa hway.
They a gue ha AI ools can au oma e da a collec ion
on ime and esou ce use, educing he bu den o
manual cos acking. By doing so, AI di ec ly enhances
he eliabili y o hospi al cos accoun ing sys ems,
which a e o en c i icized o ou da ed o agg ega ed
da a.
In pa allel, [11] p oposed he concep o alue-based
heal hca e, whe e cos s a e sys ema ically linked o
pa ien ou comes. In eg a ing AI-enabled TDABC wi h
pe o mance measu emen sys ems could s eng hen
hospi als’ abili y o ack alue c ea ion, aligning wi h
he goals o managemen con ol.
Ba ie s and limi s o in eg a ion
Despi e hese p omising applica ions, se e al ba ie s
emain. [17] iden i ied low da a quali y, lack o
in e ope abili y, and esis ance among s a as key
ac o s hinde ing AI adop ion in hospi als. [14]
emphasized ha o ganiza ional go e nance and
cul u e play decisi e oles in whe he AI ools mo e
beyond pilo s o become embedded in decision-making
p ocesses. These limi a ions imply ha , while AI can
echnically enhance MCS, i s ac ual con ibu ion
depends hea ily on go e nance, change managemen ,
and he alignmen o incen i es.
C i ical Re iew o he Li e a u e (2020–2025)
This sec ion c i ically syn hesizes empi ical and
concep ual con ibu ions on AI in hospi als and d aws
explici connec ions o managemen con ol sys ems
(MCS). We s uc u e he e iew a ound con e gences,
di e gences, me hodological limi a ions, and
un esol ed gaps ele an o budge ing, cos ing, and
pe o mance measu emen .
Con e gences: Wha he ield b oadly ag ees on
(a) P edic i e accu acy o pa ien low is consis en ly
highe wi h AI han wi h adi ional baselines.
Mul iple s udies epo ha machine lea ning models
p edic leng h o s ay (LoS) and admissions mo e
accu a ely han classical s a is ical me hods, enabling
be e bed and s a ing planning [8,15]. These esul s
a e epea edly amed as ope a ional enable s o
planning and h oughpu , which a e di ec ly ele an
inpu s o capaci y budge s and a iance analysis in
MCS.
(b) Simula ion and “digi al win” app oaches help
an icipa e esou ce bo lenecks.
Digi al wins a e desc ibed as use ul o es al e na i e
scheduling and con igu a ion scena ios be o e
implemen a ion, educing ial-and-e o on he shop
loo [16]. The same logic—using i ual eplicas o
e alua e h oughpu and esilience—appea s in
ope a ions con ex s and is ans e able o hospi al
logis ics [12]. Fo con olle s, scena io e alua ion
in o ms ex-an e budge ing and sensi i i y analysis.
(c) TDABC p o ides mo e g anula cos in o ma ion
han adi ional alloca ions.
In heal hca e se ings, Time-D i en Ac i i y-Based
Cos ing yields mo e p ecise, pa hway-le el cos
a ibu ion han depa men -le el a e ages, and is
explici ly epo ed o suppo manage ial decisions
[13]. This aligns wi h MCS objec i es o imp o ing
s anda d cos se ing and explaining budge a iances.
(d) Balanced Sco eca d and in eg a ed PMS imp o e
accoun abili y when inancial and non- inancial
indica o s a e combined.
Re iews o BSC in heal hca e conclude ha linking
pa ien , p ocess, lea ning, and inancial pe spec i es
enhances pe o mance moni o ing and accoun abili y
[2]. This con e gence ma e s because AI ou pu s
(p edic ions, ale s) can be embedded as leading
indica o s inside such sys ems.
Di e gences: Whe e indings o emphases do no align
(a) F om ope a ional pilo s o go e nance in eg a ion.
While many a icles show posi i e ope a ional esul s
(e.g., be e LoS o ecas s), se e al e iews highligh
limi ed e idence ha hese ools a e in eg a ed in o
hospi al go e nance and con ol cycles [5,14]. Some
s udies emain echnology-cen ic; o he s emphasize
o ganiza ional ou ines and cul u e. The di e gence
sugges s a ansla ion gap be ween p oo -o -concep
and ou ine MCS use.
(b) Scope o impac measu emen . Ope a ional me ics
(accu acy, wai ing imes, occupancy) a e equen ly
epo ed, bu inancial impac (budge adhe ence, cos
pe case, a iance educ ion) is less consis en ly
measu ed o causally es ablished [5,15]. This c ea es
disag eemen abou he magni ude o AI’s con ibu ion
o managemen con ol.
(c) Da a go e nance e sus speed o deploymen .
Ba ie s such as da a quali y, in e ope abili y, and
explainabili y a e well-documen ed [17], ye some
implemen a ions p io i ize apid deploymen wi h
limi ed model go e nance. Re iews ha emphasize
go e nance cau ion agains scale-up wi hou obus
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
221
da a and model con ols [14]. Thus, he li e a u e
di e ges be ween “mo e as ” case epo s and
“go e n i s ” amewo ks.
Discussion and Implica ions
The concep ual model p esen ed abo e highligh s how
AI can augmen hospi al managemen con ol sys ems
(MCS) by p o iding p edic i e insigh s, g anula cos ing,
and simula ion-based planning. This sec ion discusses
he implica ions o hese indings o hospi al manage s
and con olle s, policymake s, and he academic
communi y.
Implica ions o hospi al manage s and con olle s
Fo hospi al manage s, he in eg a ion o AI in o MCS
p o ides a means o mo e om e ospec i e epo ing
owa d eal- ime, p edic i e decision-making. [8]
showed ha machine lea ning models signi ican ly
imp o e leng h o s ay o ecas s, which can be di ec ly
ansla ed in o mo e accu a e capaci y budge s and
s a ing plans. When such p edic ions a e embedded
in o budge ing p ocesses, con olle s can p oac i ely
manage a iances ins ead o me ely explaining hem
a e he ac .
In e ms o cos ing,[13] demons a ed ha Time-D i en
Ac i i y-Based Cos ing (TDABC) enhances cos accu acy
by cap u ing esou ce use a he pa hway le el. When
pai ed wi h AI o au oma ed ime da a collec ion, his
p o ides con olle s wi h imely and p ecise cos
in o ma ion. This capabili y s eng hens a iance
analysis and suppo s s a egic in es men decisions.
Digi al win echnologies o e addi ional bene i s o
manage s by allowing hem o simula e “wha -i ”
scena ios be o e esou ce ealloca ion [16]. Fo
example, adjus ing he numbe o su gical slo s can be
es ed i ually, wi h p ojec ed impac s on bo h wai ing
imes and cos s. Such ools ans o m he ole o he
con olle om a passi e epo e o an ac i e pa ne
in scena io-based decision-making.
Implica ions o policymake s and heal hca e
go e nance
A he policy le el, he adop ion o AI in hospi al
managemen con ol aises ques ions o da a
go e nance, accoun abili y, and equi y. [17]
emphasized ha low da a quali y and in e ope abili y
ba ie s emain majo obs acles o AI deploymen in
hospi als. [14] u he a gued ha go e nance
s uc u es—such as clea accoun abili y o model
ou comes—de e mine whe he AI p ojec s scale
beyond pilo s. Policymake s he e o e need o es ablish
egula o y amewo ks and unding mechanisms ha
ensu e hospi als can in es in obus da a
in as uc u es and main ain anspa ency in AI-d i en
decision-making.
Mo eo e , embedding AI in o pe o mance
dashboa ds, such as Balanced Sco eca ds, may isk
o e emphasizing e iciency i no coun e balanced wi h
equi y and quali y-o -ca e me ics [2]. Policymake s
mus hus encou age a balanced app oach ha
in eg a es bo h inancial con ol and pa ien -cen e ed
ou comes.
Implica ions o academic esea ch
F om an academic pe spec i e, he p oposed model
unde sco es he need o empi ical s udies ha ace
ull con ol loops— om p edic ion, o decision, o
inancial ou comes. [5] no ed ha mos AI esea ch in
heal hca e emains ocused on clinical o ope a ional
ou comes a he han managemen con ol. Fu u e
s udies should explici ly measu e whe he AI-enhanced
p edic ions educe budge a iances, imp o e
s anda d- s-ac ual cos alignmen , o enhance
Balanced Sco eca d pe o mance.
In addi ion, mo e esea ch is needed on he mode a ing
ole o go e nance. While ba ie s o AI adop ion a e
documen ed [17], ew s udies quan i y how
go e nance p ac ices—such as model moni o ing o
da a quali y assu ance—a ec he inancial bene i s o
AI in eg a ion. Mul i-si e compa a i e s udies could
cla i y hese dynamics and p o ide s onge causal
e idence.
Finally, he in eg a ion o TDABC and AI ep esen s a
p omising bu unde explo ed on ie . [13]
demons a ed TDABC’s po en ial in heal hca e, bu
empi ical wo k linking AI-gene a ed eal- ime da a o
TDABC-based cos sys ems is sca ce. Explo ing his
linkage would di ec ly add ess he needs o hospi al
con olle s o accu a e, imely cos in o ma ion.
Syn hesis
In sum, AI p esen s bo h oppo uni ies and challenges
o hospi al managemen con ol. Fo manage s, i
o e s ools o an icipa e cos s and op imize esou ces.
Fo policymake s, i highligh s he u gency o
go e nance amewo ks ha sa egua d quali y and
equi y. Fo esea che s, i opens a ich agenda ocused
on linking p edic i e analy ics, cos ing, and
pe o mance measu emen in in eg a ed con ol
cycles. The o e a ching implica ion is ha AI does no
eplace managemen con ol—i ans o ms i by
shi ing i om e ospec i e moni o ing owa d
p oac i e, simula ion-based go e nance.
Me hodological limi a ions ha weaken in e ence o
MCS
(a) Unde powe ed causal designs o budge a y
ou comes. Many s udies epo imp o ed p edic ions
o simula ed e iciency bu lack quasi-expe imen al
designs o mul icen e compa isons ha would isola e
e ec s on budge a iance, s anda d- s-ac ual cos
gaps, o BSC a ge s [15]. Wi hou hese, implica ions
o MCS emain sugges i e a he han demons a ed.
(b) Sho ollow-up and na ow con ex s. E idence
o en comes om single hospi als o sho ime
windows, limi ing ex e nal alidi y o con ol cycles
ha ope a e annually o ac oss ne wo ks [8,16]. Tha
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
222
weakens claims abou sus ained imp o emen s o
budge ing and pe o mance dashboa ds.
(c) Missing linkage be ween cos ing and p edic i e
laye s. Al hough TDABC imp o es cos g anula i y [13]
and AI imp o es o ecas s [8], ew s udies explici ly
connec he wo o show how o ecas ed olumes eed
TDABC-d i en s anda d cos s o olling budge s. This
me hodological sepa a ion hampe s MCS in eg a ion.
Subs an i e gaps and wha hey imply o
managemen con ol
Gap 1 — F om p edic i e accu acy o budge a y alue.
The li e a u e shows ha AI can p edic lows [8,15] bu
a ely es s whe he using hose p edic ions educes
budge a iances o imp o es se ice-line p o i abili y.
Fu u e wo k should es he pa hway: p edic ions →
s a ing/bed decisions → cos a iances wi hin he MCS
cycle [5].
Gap 2 — Go e nance as a mode a o o pe o mance
e ec s. Ba ie s ela ed o da a and cul u e a e well-
documen ed [14,17], bu ew s udies quan i y how
go e nance p ac ices (da a ca alogs, model moni o ing,
accoun abili y) mode a e he impac o AI on MCS
ou comes. This in i es explici mode a o analyses in
mul i-si e se ings.
Gap 3 — Coupling TDABC wi h AI-enabled ope a ional
da a. E idence suppo s TDABC’s supe io i y o cos ing
[13], ye s udies seldom ins umen TDABC wi h eal-
ime imes amps gene a ed by AI/IoT, and hen ack
e ec s on s anda d- s-ac ual cos gaps. Demons a ing
his coupling would di ec ly add ess con olle s’ needs.
Gap 4 — PMS and BSC as ecipien s o AI signals. BSC
usage in hospi als is documen ed [2], bu ew pape s
desc ibe how AI ou pu s become leading indica o s
inside PMS wi h clea a ge -se ing, eedback loops,
and accoun abili y [6]. This limi s ansla ion in o
ou ine con ol p ac ices.
Gap 5 — O ganiza ional design and budge ing
ou ines. Classic wo k shows ensions be ween igid
budge s and dynamic ca e needs [1]. AI could mi iga e
hese ensions ia olling o ecas s, ye obus e idence
connec ing AI adop ion o changes in budge ing
ou ines is sca ce.
Implica ions o a con ol-o ien ed esea ch agenda
1. Design s udies ha ace ull con ol loops.
Mo e beyond accu acy me ics o examine how
p edic ions change decisions and how hose decisions
change budge a iances and BSC a ge s [5,15].
2. Tes go e nance as a mode a o . Compa e
si es wi h s ong da a/model go e nance agains hose
wi hou o quan i y di e en ial pe o mance [14,17].
3. In eg a e TDABC wi h AI da a s eams.
Ins umen TDABC wi h au oma ed ime s amps and
compa e cos ing accu acy and s anda d cos adhe ence
be o e/a e [13].
4. Embed AI signals in o PMS/BSC. Speci y AI-
de i ed leading indica o s, a ge -se ing ules, and
accoun abili y mechanisms wi hin Fe ei a & O ley’s
PMS logic and hospi al BSC p ac ice [2,6].
5. Conside o ganiza ional ou ine change.
Examine whe he AI adop ion igge s shi s om ixed
budge s o olling o ecas s o lexible a iance
h esholds in line wi h documen ed igidi y p oblems
[1].
Concep ual Model: In eg a ing AI in o Hospi al
Managemen Con ol Sys ems
Building on exis ing amewo ks
The concep ual model p oposed in his pape is
designed by combining insigh s om es ablished
amewo ks in managemen con ol and ecen
de elopmen s in heal hca e AI:
• Balanced Sco eca d (BSC) [10]: widely applied
in hospi als o link inancial, pa ien , p ocess, and
lea ning ou comes [2].
• Time-D i en Ac i i y-Based Cos ing (TDABC)
[9]: shown o p oduce mo e accu a e cos alloca ions
pe pa hway and suppo manage ial decision-making
in heal hca e [13].
• Pe o mance Managemen Sys em (PMS)
amewo k [6]: emphasizes he impo ance o
objec i es, measu es, a ge s, and eedback loops in
aligning decision-making.
• AI-enabled ools: including p edic i e
analy ics o pa ien lows [8,15] and digi al wins o
simula ing and op imizing esou ce alloca ion [12,16].
Each o hese amewo ks add esses a speci ic
dimension o hospi al pe o mance. Howe e , cu en
li e a u e shows ha hey a e a ely in eg a ed. The
model p oposed he e aims o combine hem in o a
cohe en a chi ec u e o AI-augmen ed hospi al
managemen con ol.
Desc ip ion o he p oposed model
The concep ual model (Figu e 1, o be de eloped o
he ull a icle) links h ee co e laye s:
1. Da a and p edic ion laye :
o AI collec s and p ocesses da a om elec onic
heal h eco ds, en e p ise esou ce planning sys ems,
and ope a ional da abases.
o P edic i e models (e.g., LoS o ecas ing,
demand o ecas ing) gene a e o wa d-looking insigh s
ha di ec ly eed budge ing and planning cycles [8].
2. Cos ing and measu emen laye :
o AI-enhanced TDABC acks esou ce use in eal
ime, enabling mo e g anula cos ing pe p ocedu e o
pa ien pa hway [13].
o These cos s a e connec ed o BSC dimensions
( inancial, pa ien , in e nal p ocesses, lea ning),
ensu ing a balanced pe o mance pe spec i e[2].
3. Go e nance and decision laye :
o P edic i e indica o s and cos measu es a e
embedded in PMS eedback loops [6].
o Digi al win simula ions allow manage s o es
“wha -i ” scena ios be o e implemen ing changes [16].
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
223
o Go e nance s uc u es ensu e da a quali y,
model accoun abili y, and alignmen wi h
o ganiza ional s a egy [14].
P oposi ions o u u e esea ch
F om his concep ual model, we de i e se e al es able
p oposi ions:
• P1. In eg a ing AI-based LoS p edic ions in o
hospi al capaci y planning will educe budge a iances
associa ed wi h unexpec ed admissions.
• P2. The combina ion o AI-enabled TDABC and
PMS amewo ks will inc ease he accu acy o
s anda d- s-ac ual cos analysis, imp o ing inancial
con ol.
• P3. Digi al win simula ions embedded in
con ol sys ems will enhance he eliabili y o scena io
planning, leading o mo e e ec i e alloca ion o sca ce
esou ces.
• P4. The impac o AI on hospi al inancial
pe o mance will be posi i ely mode a ed by
go e nance p ac ices, such as da a quali y
managemen and algo i hm accoun abili y.
• P5. Hospi als ha embed AI ou pu s in o
Balanced Sco eca d dashboa ds will achie e highe
alignmen be ween ope a ional decisions and s a egic
objec i es.
Con ibu ion o he model
This in eg a i e model con ibu es o he li e a u e by
explici ly linking ope a ional AI applica ions (p edic ion,
op imiza ion, simula ion) wi h managemen con ol
ools (budge ing, cos ing, PMS, BSC). While p io
s udies o en examine hese componen s sepa a ely,
he p oposed amewo k highligh s hei
complemen a i y and sugges s pa hways o empi ical
alida ion.
Limi a ions and Resea ch Agenda
Al hough esea ch on a i icial in elligence (AI) in
heal hca e has accele a ed since 2020, he cu en
body o e idence e eals signi ican limi a ions ha
es ic i s ansla ion in o hospi al managemen
con ol sys ems (MCS). Recognizing hese sho comings
is essen ial o designing a u u e esea ch agenda ha
can s eng hen bo h academic igo and manage ial
ele ance.
Me hodological limi a ions
O e eliance on single-si e s udies. Many empi ical
s udies ely on da a om one hospi al o heal h sys em,
which cons ains he gene alizabili y o esul s.[8], o
example, used a la ge da ase o imp o e leng h o s ay
(LoS) p edic ion, bu hei indings emain con ex -
speci ic. Simila ly, [16] illus a ed he po en ial o digi al
wins h ough concep ual modeling and simula ion, ye
wi hou la ge-scale, eal-wo ld implemen a ion
e idence.
Lack o causal designs o inancial ou comes. [15]
e iewed LoS p edic ion s udies and ound consis en
imp o emen s in accu acy. Howe e , e y ew o hese
s udies es ed whe he imp o ed p edic ions led o
measu able changes in inancial pe o mance, such as
educed budge a iances o imp o ed cos - o-
ou come a ios. This me hodological gap limi s he
abili y o claim a causal link be ween AI adop ion and
MCS e ec i eness.
F agmen a ion o domains. While TDABC has been
alida ed as a cos ing me hod in heal hca e [13], and
p edic i e AI models ha e been alida ed o
ope a ional e iciency [8], he li e a u e seldom
in eg a es he wo. Cos ing, p edic ion, and go e nance
a e o en s udied in silos, making i di icul o assess
hei combined e ec on hospi al con ol sys ems.
Sho - e m ocus. Mos s udies examine immedia e
imp o emen s (e.g., p edic ion accu acy, sho e
wai ing imes), bu ew analyze long- e m sus ainabili y
o AI in e en ions wi hin annual budge cycles o
mul i-yea s a egic con ol amewo ks [5].
Theo e ical limi a ions
Limi ed in eg a ion wi h con ol sys em amewo ks.
Al hough he Balanced Sco eca d [10] and Pe o mance
Managemen Sys ems [6] a e widely used in
managemen con ol esea ch, AI applica ions a e
a ely heo ized wi hin hese amewo ks. Ins ead, hey
a e o en p esen ed as echnological add-ons a he
han as in eg al componen s o hospi al con ol
sys ems.
Insu icien ocus on go e nance as a mode a o .
Ba ie s such as da a quali y and esis ance o adop ion
a e documen ed [14,17], bu ew s udies measu e how
go e nance p ac ices—such as da a s ewa dship,
algo i hmic audi ing, o accoun abili y s uc u es—
a ec he pe o mance ou comes o AI adop ion.
Fu u e esea ch agenda
Based on hese limi a ions, we iden i y se e al
p io i ies o u u e esea ch. Fi s , u u e s udies
should go beyond single-hospi al case s udies o
include mul i-hospi al o mul i-coun y designs,
enabling s onge ex e nal alidi y. Quasi-expe imen al
designs could es whe he AI-enhanced p edic ions
educe budge a iances o imp o e Balanced
Sco eca d indica o s. Second, esea ch should explici ly
es he causal chain linking AI p edic ions o
manage ial decisions and hen o inancial ou comes.
Wi hou his, he impac o AI on MCS emains
specula i e. Thi d, schola s should combine TDABC
wi h AI-d i en ope a ional da a (e.g., imes amps,
wo kload p edic ions) o es whe he his in eg a ion
imp o es s anda d- s-ac ual cos analysis. Fou h,
compa a i e s udies should analyze how hospi als wi h
s ong go e nance p ac ices (da a quali y
managemen , model moni o ing) achie e di e en
ou comes han hose wi hou . Finally, longi udinal
esea ch is needed o assess he sus ainabili y o AI
in e en ions in annual budge ing and s a egic
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
224
planning cycles, a he han sho - e m ope a ional
bene i s alone.
Syn hesis
In sum, he cu en li e a u e demons a es AI’s
echnical po en ial o p edic ion, op imiza ion, and
cos ing, bu me hodological and heo e ical limi a ions
p e en de ini i e claims abou i s impac on hospi al
managemen con ol. The nex gene a ion o esea ch
mus he e o e adop designs ha a e b oade , deepe ,
and longe - e m, while embedding AI i mly in o
es ablished con ol amewo ks such as BSC, TDABC,
and PMS.
Conclusion
This pape has c i ically examined he con ibu ion o
a i icial in elligence (AI) o hospi al managemen
con ol sys ems (MCS). The e iew showed ha AI
p o ides clea oppo uni ies o enhance o ecas ing o
pa ien lows, op imize esou ce alloca ion, and
imp o e cos ing accu acy h ough TDABC. Howe e ,
despi e hese p omising de elopmen s, cu en
esea ch emains agmen ed, o en limi ed o
ope a ional pilo s, and a ely embedded in o budge ing
cycles o pe o mance managemen amewo ks such
as he Balanced Sco eca d.
The concep ual model p oposed in his a icle
in eg a es AI ools wi h es ablished managemen
con ol amewo ks, highligh ing hei
complemen a i y. I posi ions AI no as a eplacemen
bu as an augmen a ion o managemen con ol,
mo ing om e ospec i e epo ing owa d p oac i e
and simula ion-based go e nance. Impo an ly, he
e iew iden i ied se e al gaps, including he lack o
causal e idence on inancial ou comes, insu icien
in eg a ion o cos ing and p edic i e ools, and he
limi ed ole gi en o go e nance as a mode a o .
The implica ions a e h ee old. Fo p ac i ione s, AI
o e s con olle s ools o an icipa e cos s and ac as
s a egic pa ne s in hospi al decision-making. Fo
policymake s, go e nance and da a in as uc u e
emain c i ical p e equisi es o success ul adop ion.
Fo esea che s, he agenda poin s owa d mul i-si e,
longi udinal, and heo e ically g ounded s udies ha
explici ly connec AI in e en ions wi h inancial and
s a egic con ol ou comes.
Ul ima ely, hospi als will only cap u e he ull po en ial
o AI in managemen con ol i echnological inno a ion
is accompanied by obus go e nance, o ganiza ional
lea ning, and in eg a ion in o es ablished pe o mance
amewo ks.
Re e ences
[1] Aidema k LG. The meaning o balanced
sco eca ds in he heal h ca e o ganisa ion. Financial
Accoun abili y & Managemen . 2001;17(1):23–40.
doi:10.1111/1468-0408.00119.
[2] Ame F, e al. The deploymen o balanced
sco eca d in heal h ca e o ganiza ions: A sys ema ic
e iew. BMC Heal h Se Res. 2022;22:148.
doi:10.1186/s12913-022-07555-9.
[3] An hony RN. Planning and con ol sys ems: A
amewo k o analysis. Bos on: Di ision o Resea ch,
G adua e School o Business Adminis a ion, Ha a d
Uni e si y; 1965.
[4] Bajwa J, Muni U, No i A, Williams B. A i icial
in elligence in heal hca e: ans o ming he p ac ice o
medicine. Fu u e Heal hc J. 2021 Jul;8(2):e188–e194.
doi:10.7861/ hj.2021-0095.
[5] Da enpo T, Kalako a R. The po en ial o
a i icial in elligence in heal hca e. Fu u e Heal hc J.
2019 Jun;6(2):94–98. doi:10.7861/ u u ehosp.6-2-94.
[6] Fe ei a A, O ley D. The design and use o
pe o mance managemen sys ems: An ex ended
amewo k o analysis. Manag Accoun Res. 2009
Dec;20(4):263–282. doi:10.1016/j.ma .2009.07.003.
[7] Hoozée S, B uggeman W. Iden i ying
ope a ional imp o emen s du ing he design p ocess o
a ime-d i en ABC sys em: The ole o collec i e wo ke
pa icipa ion and leade ship s yle. Manag Accoun Res.
2010 Sep;21(3):185–198.
doi:10.1016/j.ma .2010.01.003.
[8] Jain R, Singh M, Rao AR, Ga g R. P edic ing
hospi al leng h o s ay using machine lea ning on a
la ge open heal h da ase . BMC Heal h Se Res. 2024
Jul 30;24:860. doi:10.1186/s12913-024-11238-y.
[9] Kaplan RS, Ande son SR. Time-d i en ac i i y-
based cos ing: A simple and mo e powe ul pa h o
highe p o i s. Bos on: Ha a d Business P ess; 2007.
[10] Kaplan RS, No on DP. S a egy maps:
Con e ing in angible asse s in o angible ou comes.
Bos on: Ha a d Business School P ess; 2004.
[11] Kaplan RS, Po e ME. How o sol e he cos
c isis in heal h ca e. Ha Bus Re . 2011 Sep;89(9):46–
52.
[12] Longo F, Pado ano A, Umb ello S. The digi al
supply chain win pa adigm o enhancing esilience
and sus ainabili y. Compu Ind. 2023;149:103913.
doi:10.1016/j.compind.2022.103913.
[13] Niñe ola A, He nández-La a AB, Sánchez-
Rebull MV. Imp o ing heal hca e pe o mance h ough
Ac i i y-Based Cos ing and Time-D i en Ac i i y-Based
Cos ing. In J Heal h Plann Manage. 2021
No ;36(6):2079–2093. doi:10.1002/hpm.2999.
[14] Roppel JS, Kanbach DK, K aus S. A i icial
in elligence in heal hca e ins i u ions: A sys ema ic
li e a u e e iew on in luencing ac o s. Technol Soc.
2024;76:102387. doi:10.1016/j. echsoc.2024.102387.
[15] S one K, e al. A sys ema ic e iew o he
p edic ion o hospi al leng h o s ay. J Biomed In o m.
2022 Jul;130:104087. doi:10.1016/j.jbi.2022.104087.
[16] Vallée A. Digi al win o heal hca e sys ems.
F on Digi Heal h. 2023;5:1253050.
doi:10.3389/ dg h.2023.1253050.
EUR J MED HEALTH RES
Volume 3 | Numbe 3 | May-June 2025
225
[17] Ahmed MI, e al. A sys ema ic e iew o he
ba ie s o he implemen a ion o AI in heal hca e.
Cu eus. 2023 May 15;15(5):e44324.
doi:10.7759/cu eus.44324.
[18] Topol EJ. Deep medicine: How a i icial
in elligence can make heal hca e human again. New
Yo k: Basic Books; 2019.
[19] Choudhu y A, Asan O. Role o a i icial
in elligence in pa ien sa e y ou comes: sys ema ic
li e a u e e iew. JMIR Med In o m. 2020 Jul
15;8(7):e18599. doi:10.2196/18599.
[20] Obe meye Z, Emanuel EJ. P edic ing he
u u e—big da a, machine lea ning, and clinical
medicine. N Engl J Med. 2016 Oc 13;375(15):1216–
1219. doi:10.1056/NEJMp1606181.
[21] an de We e ing R, e al. The impac o hospi al
in o ma ion sys ems on managemen con ol: a
sys ema ic e iew. Heal h Policy Technol. 2021
Jun;10(2):100517. doi:10.1016/j.hlp .2021.100517.
[22] Hamami F, Jouini O. A i icial in elligence in
heal hca e ope a ions: A e iew. In J P od Res. 2020
Sep;58(18):5742–5764.
doi:10.1080/00207543.2020.1780531.
[23] Alami H, e al. A i icial in elligence in heal h
ca e: laying he ounda ion o esponsible, sus ainable,
and inclusi e inno a ion. CMAJ. 2020 No
9;192(44):E1254–E1258. doi:10.1503/cmaj.200978.
[24] Rajkoma A, Dean J, Kohane I. Machine
lea ning in medicine. N Engl J Med. 2019 Ap
4;380(14):1347–1358. doi:10.1056/NEJMp1903100.
[25] Elbanna S, Child J. The in luence o
go e nance, s a egy and pe o mance managemen
on hospi al inno a ion. Public Manag Re . 2021
Jul;23(7):1001–1025.
doi:10.1080/14719037.2020.1744365.
[26] Sil a E, e al. A i icial in elligence in
heal hca e: Ba ie s and acili a o s o
implemen a ion. BMJ Heal h Ca e In o m. 2022 Jan
6;29(1):e100495. doi:10.1136/bmjhci-2021-100495.
[27] Buolamwini J, Geb u T. Gende shades:
In e sec ional accu acy dispa i ies in comme cial
gende classi ica ion. P oceedings o Machine Lea ning
Resea ch. 2018;81:1–15.
[28] G eenhalgh T, Papou si C. S udying complexi y
in heal h se ices esea ch: despe a ely seeking an
o e due pa adigm shi . BMC Med. 2021 Ma 10;19:57.
doi:10.1186/s12916-021-01994-7.
[29] I ano D, Dolgui A. A digi al supply chain win
o managing dis up ions. In J P od Res. 2023
Jun;61(11):3552–3567.
doi:10.1080/00207543.2023.2190143.
[30] Williamson B, Eynon R. His o ical h eads,
missing links, and u u e di ec ions in AI in educa ion.
Lea n Media Technol. 2020 Jul;45(3):223–235.
doi:10.1080/17439884.2020.1776889.
[31] Nassou Y. Public Heal h and Economic
Resilience in Mo occo: Towa ds a Sus ainable
Heal hca e Model. In J Sci Res Inno S ud.
2025;4(1):67–77. doi: 10.63883/ijs isjou nal. 4i1.460
