incidence and preventability of adverse events requiring intensive care admission: a systematic...
TRANSCRIPT
Incidence and preventability of adverse events requiringintensive care admission: a systematic reviewjep_1612 1..15
Annemie Vlayen ••,1 Sandra Verelst MD,2 Geertruida E Bekkering PhD,3 Ward Schrooten MD PhD MSc,4
Johan Hellings PhD5 and Nerée Claes MD PhD4
1Researcher, Hasselt University, Faculty of Medicine, Diepenbeek, Belgium2Intensivist, Emergency Department, Katholieke Universiteit Leuven, Leuven, Belgium3Researcher, Belgian Center for Evidence Based Medicine, Katholieke Universiteit Leuven, Leuven, Belgium4Professor, Hasselt University, Faculty of Medicine, Diepenbeek, Belgium5Professor, Hasselt University, Faculty of Medicine, Diepenbeek, Belgium and Delegate Director, ICURO, Brussels, Belgium
Keywords
adverse events, intensive care units,medical audit, medical errors, medicalrecords, patient safety
Correspondence
•• Annemie VlayenHasselt UniversityFaculty of MedicineAgoralaanBuilding D3590 DiepenbeekBelgiumE-mail: [email protected]
Accepted for publication: 2 November 2010
AbstractRationale, aims and objectives Adverse events are unintended patient injuries or com-plications that arise from health care management resulting in death, disability or pro-longed hospital stay. Adverse events that require critical care are a considerable financialburden to the health care system, but also their global impact on patients and society isprobably underestimated. The objectives of this systematic review were to synthesize thebest available evidence regarding the estimates of the incidence and preventability ofadverse events that necessitate intensive care admission, to determine the type and conse-quences [mortality, length of intensive care unit (ICU) stay and costs] of these adverseevents.Methods MEDLINE (from 1966 to present), EMBASE (from 1974 to present) andCENTRAL (version 1-2010) were searched for studies reporting on unplanned admissionson ICUs. Several other sources were searched for additional studies. Only quantitativestudies that used chart review for the detection of adverse events requiring intensive careadmission were considered for eligibility. For the purposes of this systematic review, ICUswere defined as specialized hospital facilities which provide continuous monitoring andintensive care for acutely ill patients. Studies that were published in the English, Dutch,German, French or Spanish language were eligible for inclusion. Two reviewers indepen-dently extracted data and assessed the methodological quality of the included studies.Results A total of 27 studies were reviewed. Meta-analysis of the data was not appropriatebecause of methodological and statistical heterogeneity between studies; therefore, resultsare presented in a descriptive way. The percentage of surgical and medical adverse eventsthat required ICU admission ranged from 1.1% to 37.2%. ICU readmissions varied from0% to 18.3%. Preventability of the adverse events varied from 17% to 76.5%. Preventableadverse events are further synthesized by type of event. Consequences of the adverse eventsincluded a mean length of ICU stay that ranged from 1.5 days to 10.4 days for the patient’sfirst stay in ICU and mortality percentages between 0% and 58%.Conclusions Adverse events are an important reason for (re)admission to the ICU and aconsiderable proportion of these are preventable. It was not possible to estimate an overallincidence and preventability rate of these events as we found considerable heterogeneity. Todecrease adverse events that necessitate ICU admission, several systems are recommendedsuch as early detection of patients with clinical instability on general wards and theimplementation of rapid response teams. Step-down or intermediate care units could be auseful strategy for patients who require monitoring to avoid ICU readmissions. However,the effectiveness of such systems needs to be investigated.
jep_1612
Journal of Evaluation in Clinical Practice ISSN 1356-1294
© 2010 Blackwell Publishing Ltd, Journal of Evaluation in Clinical Practice 1
1
2
3
4
56789
10
11121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859
33
IntroductionTo date, there is insufficient evidence on the causes of adverseevents and to what extent patients are harmed. Several interna-tional organizations are putting efforts to create awareness to thisissue and put forward recommendations and solutions in order toreduce patient harm [1–3].
The consequences of patient harm have been estimated by Zhanand Miller on an excess length of stay attributable to medicalerrors of 2.4 million hospital days, which account for $9.3 billionexcess charges, and 32 591 attributable deaths in the USA annually[4].
Several national studies describe the use of medical recordreview to measure the occurrence of adverse events in hospitals[5–13]. The large variation in the incidence of adverse eventsamong these studies in different countries may be explained eitherby true differences in patient safety of the different health caresystems, or by methodological differences between the studies[14]. Despite the awareness that a substantial number of adverseevents are detected among unexpected intensive care unit (ICU)admissions, little is known about the epidemiology of these events.We conducted a systematic review to synthesize the best availableevidence regarding the estimates of the incidence and preventabil-ity of adverse events requiring unplanned ICU (re)admission. Inaddition, we synthesized the types and consequences of theseevents, including mortality rates, length of ICU stay and directmedical costs.
Methods
Review protocol
The review protocol is available from the authors and the JoannaBriggs Institute [15] and details the predefined criteria for thisreview.
Eligibility criteria
This review only considered quantitative studies on the incidenceof adverse events requiring unplanned ICU (re)admissions inpublic or private general hospital settings. Only study designswhich used chart review were included. An adverse event wasdefined as ‘an event that results in unintended harm to the patientby an act of commission or omission rather than by the underlyingdisease or condition of the patient’ [11,16]. Unplanned ICU admis-sions were defined as all patients unexpectedly admitted to theICU from a lower level of care in the hospital. Unplanned ICUreadmissions were all patients discharged from the ICU to a lowerlevel of care in the hospital that had an unplanned return to theICU. Studies that reported on patients admitted on neonatal ICUsor transfers from outside hospitals were excluded.
Primary outcomes were the number of unplanned ICU (re)ad-missions (measured as a proportion, compared with number ofICU admissions), the number of adverse events requiring ICUadmission (measured as a proportion, compared with the numberof ICU admissions) and the number of preventable adverse events(measured as a proportion, compared with the incidence rate).
Secondary outcomes were the type of event, attributable causes,location and subdivision by provider of care, consequences (mor-tality rates, harm, length of ICU stay), direct medical costs ofadverse events and kappa-coefficients (k) of the chart reviewmethods.
Search strategy
We completed searches in MEDLINE (from 1966 to present),EMBASE (from 1974 to present) and CENTRAL (version1-2010) in January 2010. The search strategy combined selectedMeSH terms and free-text terms to identify quantitative studieson the incidence of adverse events requiring unplanned ICU(re)admissions in public or private general hospital settings(Appendix I). Only studies using chart review and published inthe English, Dutch, German, French or Spanish language wereincluded.
The Journal of Intensive Care Medicine, Quality and Safety inHealth Care, International Journal for Quality in Health Care, theproceedings of the International Symposium on Intensive Care andEmergency Medicine (Critical Care) were searched for relevantpapers or conference abstracts. Reference lists of retrieved paperswere screened for new studies. In addition, we focused on nation-wide studies that used chart review for the detection of medicalerrors [5–9,11–13,17]. ISI Web of Knowledge, grey literature inOpenSIGLE, Google, AHRQ PSNet [18] and the Institute forHealthcare Improvement [19] were searched for relevant studies.Authors of relevant papers were contacted regarding any furtherpublished or unpublished work.
Study selection and appraisal
Two reviewers (AV, SV), using standardized screening forms,independently performed the initial scan of titles and abstracts ofall retrieved citations, and applied the inclusion criteria. During thescreening process, the reviewers decided to no longer apply theexclusion criterion on the definition of adverse events as fewstudies reported on the used definition. Both reviewers docu-mented the reasons for study exclusion. Full-text copies of allpotentially relevant studies were obtained and further checked forinclusion (AV, SV). Two independent reviewers (AV, SV) assessedthe included studies for methodological quality according to pre-determined criteria (Table 2). Any discrepancies between review-ers were resolved by discussion.
Data extraction
Data extraction was completed independently by two reviewers(AV, SV), using a standardized data collection form. Data on studydesign, chart review, incidence of adverse events requiring ICUadmission, preventability, outcome, location, provider of care andtype of event were extracted. Data to perform subgroup analysiswere collected on the presence of a medical emergency team,patient characteristics and characteristics of the ICUs. The datawere checked for any discrepancies and were then collated. Anydiscrepancies identified were resolved through discussion untilconsensus was reached among the review authors.
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd2
12345678910
1112131415161718192021222324252627282930313233343536373839404142434445464748495051525354
555657585960616263646566676869707172737475767778798081828384858687888990
4 4
Data synthesis
All primary outcomes were first presented descriptively, usingdata as reported in the paper (i.e. mean, SD, median, percentiles,range). Subsequently, appropriateness of meta-analyses on primaryoutcomes (proportions) was examined. To do so, for each study theproportion was expressed as a logit [20], as the distribution of logitsis more likely to be normal than the distribution of proportions.Heterogeneity was assessed by graphic inspection of forest plotsand by measuring the degree of inconsistency in the studies’ results(I2) [21]. stata statistical software was used for all analyses.
Secondary outcomes were reported in a descriptive way.A subgroup analysis based on population, country of study,
sample size, method of screening and the use of definitions onadverse events was performed to identify and assess the heteroge-neity between the studies.
Subgroup analysis was pre-specified in the study protocol.
Results
Results of the search
The initial database search identified 1100 unique citations (sixstudies were duplicate) and 16 additional studies were identifiedthrough other sources, which were screened on title and abstract. Atotal of 1033 studies were excluded. Full-text copies of 83 poten-tially relevant studies were retrieved and reviewed and of these, 54studies were excluded.
In total, 29 studies met the inclusion criteria for this review(Fig. 1). Two studies were considered duplicate studies [22,23] asthey reported on data already included in the review. These werereviewed together with the studies first published [24,25].
Study characteristics andmethodological quality
Characteristics of the included studies are presented in Table 1. Allstudies used a retrospective study design with chart review inwhich the method of record selection was consecutive. A total of10 391 patient records were derived from the included studies,with the median number of records reviewed per study of 385(interquartile range 71–497). An overview of the methodologicalquality and the use of definitions for adverse events is presented inTable 2.
Primary outcomes
Data on primary outcomes are presented in Table 3. They weresorted by the providers of care and population that were studied:anaesthetic/surgical care, general wards and ICUs.
Primary outcome data were divided into three categories: inci-dence of adverse events requiring unplanned ICU admission, ICUreadmissions and preventability of adverse events.
In total, 20 studies reported on unplanned ICU admissionsbecause of adverse events. In almost every study, an unplannedICU admission was counted as an adverse event. Nine studies
Figure 1 Flow diagram of study selection(based on PRISMA Statement, 2009). AEs,adverse events; ICU, intensive care unit.
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 3
123
4
56789
1011121314151617181920212223242526272829303132
333435363738394041424344
Tab
le1
Cha
ract
eris
tics
ofin
clud
edst
udie
s
Aut
hors
Pub
licat
ion
date
Cou
ntry
Stu
dype
riod
(mon
ths)
No/
type
ofho
spita
lsM
etho
dof
char
tre
view
No.
ofre
cord
sre
view
edP
rovi
der/
popu
latio
nO
utco
mes
Bar
nes
and
Hav
ill[2
6]19
80N
ewZe
alan
d60
1ge
nera
lhos
pita
l,2
priv
ate
hosp
itals
Not
men
tione
d11
0A
Ana
esth
esie
UIA
=A
E,
mor
talit
y
Bui
stet
al.
[27]
1999
Aus
tral
ia12
1te
achi
ngho
spita
l(3
00be
ds)
Mul
tista
ge11
2G
WC
ardi
acar
rest
,un
plan
ned
ICU
adm
issi
onA
E,
ICU
read
mis
sion
s,lo
catio
n,m
orta
lity,
LOS
,IC
ULO
SC
habo
yer
etal
.[2
8]20
08A
ustr
alia
81
met
ropo
litan
hosp
ital
(580
beds
)U
nist
age
300
ICU
Gen
eral
ICU
ICU
read
mis
sion
=A
E,
k
Cul
len
etal
.[2
9]19
92U
SA
151
teac
hing
hosp
ital
Uni
stag
e71
A,
SA
naes
thet
ic,
surg
ical
UIA
=A
E,
ICU
read
mis
sion
,lo
catio
n,ty
peof
even
t,ha
rm,
mor
talit
y,LO
S,
ICU
LOS
Dar
chy
etal
.[2
3,24
]19
9819
99Fr
ance
121
gene
ralh
ospi
tal
(500
beds
)M
ultis
tage
623
GW
All
patie
nts
adm
itted
toM
,S
ICU
from
GW
AE
,pr
even
tabl
eA
E,
loca
tion,
type
ofev
ent,
mor
talit
y,IC
ULO
S,
cost
s,w
orkl
oad
Dow
ney
and
O’C
onne
ll[3
0]
1996
Aus
tral
ia19
1un
iver
sity
-affi
liate
dpa
edia
tric
tert
iary
refe
rral
hosp
ital
(280
beds
)
Not
men
tione
d35
P-A
Pae
diat
ricpo
st-a
naes
thet
icU
IA=
AE
,pr
even
tabl
eA
E,
loca
tion,
type
ofev
ent,
mor
talit
y,TI
SS
,IC
ULO
S
Dun
net
al.
[31]
2006
Aus
tral
ia72
1pa
edia
tric
refe
rral
hosp
ital(
250
beds
)M
ultis
tage
1612
PP
aedi
atric
:de
aths
,cl
inic
alre
ferr
als,
ICU
adm
issi
ons
AE
,ty
peof
even
t,m
orta
lity
Dur
bin
and
Kop
el[3
2]19
93U
SA
181
univ
ersi
tyte
achi
ngho
spita
l(65
0be
ds)
Not
men
tione
d82
MIC
U,
SIC
UM
edic
al,
surg
ical
ICU
read
mis
sion
sIC
Ure
adm
issi
ons
=A
E,
type
ofev
ent,
mor
talit
y,LO
S,
ICU
LOS
ElS
hoba
ryet
al.
[33]
2008
Can
ada
241
Not
men
tione
d25
0S
Lapa
rosc
opic
gast
ricby
pass
AE
,m
orta
lity
End
acot
tet
al.
[34]
2009
Aus
tral
ia12
1re
gion
alho
spita
l(22
0be
ds)
Uni
stag
e38
8IC
UIC
Ure
adm
issi
ons
ICU
read
mis
sion
=A
E
Fran
klin
and
Jack
son
[35]
1983
US
A12
1U
nive
rsity
hosp
ital
Uni
stag
e29
9M
ICU
Med
ical
ICU
read
mis
sion
sIC
Ure
adm
issi
on=
AE
,pr
even
tabl
eA
E,
type
ofev
ent,
mor
talit
y,IC
ULO
SG
upta
etal
.[3
6]20
01U
SA
481
acad
emic
hosp
ital
Not
men
tione
d20
2S
Hip
,kn
eere
plac
emen
t(O
SA
)A
E,
type
ofev
ent
Hal
ler
etal
.[2
2,25
]20
0520
08A
ustr
alia
631
univ
ersi
ty-a
ffilia
ted
hosp
ital
Mul
tista
ge18
8S
Sur
gica
lU
IA,
AE
,pr
even
tabl
eA
E,
type
ofev
ent,
mor
talit
y,LO
S,
k
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Hay
eset
al.
[37]
1996
Aus
tral
ia39
1M
ultis
tage
23S
Lapa
rosc
opic
surg
ical
AE
,lo
catio
n,ty
peof
even
t,m
orta
lity,
LOS
,IC
ULO
SH
eisl
eret
al.
[38]
2009
US
A24
1te
rtia
ryca
reho
spita
lN
otm
entio
ned
736
S-G
Hys
tere
ctom
y+
surg
ery
AE
Kaf
yet
al.
[39]
2006
Can
ada
961
univ
ersi
tyte
achi
ngho
spita
lN
otm
entio
ned
1792
GH
yste
rect
omy
AE
,IC
Ure
adm
issi
on,
type
ofev
ent
Kur
owsk
ian
dS
ims
[40]
2007
Aus
tral
ia72
1te
rtia
rypa
edia
tric
hosp
ital
Not
men
tione
d35
8P
-AP
aedi
atric
post
-ana
esth
etic
UIA
=A
E,
type
ofev
ent
Lehm
ann
etal
.[4
1]20
05U
SA
64
teac
hing
hosp
itals
Mul
tista
ge64
S,
M,
PS
urgi
cal,
med
ical
,pa
edia
tric
AE
,pr
even
tabl
eA
E,
type
ofev
ent,
mor
talit
y,IC
ULO
SM
cGlo
inet
al.
[42]
1999
UK
61
teac
hing
hosp
ital
(702
beds
)M
ultis
tage
563
GW
Une
xpec
ted
deat
hsan
dre
ferr
als
from
GW
AE
,IC
Ure
adm
issi
on,
prev
enta
ble
AE
,ty
peof
even
t,m
orta
lity,
ICU
LOS
Oka
for
[43]
2009
Nig
eria
481
univ
ersi
tyte
achi
ngho
spita
lN
otm
entio
ned
26S
,A
Pos
t-su
rgic
al,
anae
sthe
ticU
IA=
AE
,ty
peof
even
t,m
orta
lity,
ICU
LOS
Pie
rcy
etal
.[4
4]20
06A
ustr
alia
33
acut
eho
spita
lsN
otm
entio
ned
165
S,
AP
ost-
surg
ical
,an
aest
hetic
UIA
=A
E,
prev
enta
ble
AE
,ty
peof
even
t,m
orta
lity,
Sat
yaw
anet
al.
[45]
2006
Mum
bai
61
gene
ralt
each
ing
hosp
ital
Not
men
tione
d76
S,
AP
ost-
surg
ical
,an
aest
hetic
UIA
=A
E,
ICU
disc
harg
es,
mor
talit
y,IC
ULO
SS
tew
art
and
Voss
[46]
1997
Aus
tral
ia16
1te
achi
ngho
spita
lM
ultis
tage
44C
CU
ICU
read
mis
sion
sto
aco
rona
ryca
reun
itIC
Ure
adm
issi
on=
AE
,pr
even
tabl
eA
E,
type
ofev
ent,
mor
talit
y,IC
ULO
SS
wan
net
al.
[47]
1993
Can
ada
121
Mul
tista
ge26
5S
Pos
t-su
rgic
alU
IA=
AE
,pr
even
tabl
eA
E,
type
ofev
ent,
mor
talit
y,IC
ULO
SW
olff
[48]
1995
Aus
tral
ia12
1M
ultis
tage
497
GW
All
inpa
tient
sA
EW
olff
[49]
1996
Aus
tral
ia38
1ru
ralb
ase
hosp
ital
Mul
tista
ge14
65G
WA
llin
patie
nts
AE
Yehi
aet
al.
[50]
2002
New
Zeal
and
721
Not
men
tione
d17
SH
aem
odia
lysi
sA
E,
type
ofev
ent
ICU
,int
ensi
veca
re;M
,med
ical
;CC
U,c
oron
ary
care
unit;
GW
,gen
eral
war
ds;A
,ana
esth
etic
;S,s
urgi
cal;
P,pa
edia
tric
;G,g
ynec
olog
ical
;UIA
,unp
lann
edIC
Uad
mis
sion
saf
tera
naes
thes
ia/s
urge
ry;
AE
,ad
vers
eev
ent;
LOS
,le
ngth
ofho
spita
lsta
y;k,
kapp
ava
lue
for
inte
robs
erve
rre
liabi
lity;
TIS
S,
Ther
apeu
ticIn
terv
entio
nS
corin
gS
yste
m.
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
reported on adverse events on ICU resulting in ICU readmission.Most of these studies made a comparison with ICU discharges.Only nine studies investigated the preventability of the adverseevents.
Because of considerable heterogeneity, we refrained frompooling the results. For all primary outcomes, the degree of incon-sistency I2 was 99.9%. It may be misleading to quote an average
value for the primary outcomes. As an illustration of the hetero-geneity in results, Table 4 provides the ranges of the primaryoutcomes (proportions).
A subgroup analysis based on the population (i.e. surgical,general wards and ICUs) and study characteristics (i.e. country,sample size and outcome definitions) could not clarify the causesof heterogeneity.
Table 2 Methodological quality and definitions for adverse events of included studies
Authors
Method ofselectingmedical records Sample size
Method ofscreening
Kappa value forinter-observeragreement
Explicit definitionsfor adverse event
Barnes and Havill [26] A B C C NBuist et al. [27] A C B C Y(2)Chaboyer et al. [28] A B B C Y(1)Cullen et al. [29] A C B C Y(2)Darchy et al. [23, 24] A B A C Y(1)Downey and O’Connell [30] A C C C Y(3)Dunn et al. [31] A A A C Y(1)Durbin and Kopel [32] A C C C NEl Shobary et al. [33] A B C C NEndacott et al. [34] A B B C Y(1)Franklin and Jackson [35] A B B C NGupta et al. [36] A B C C Y(2)Haller et al. [22, 25] A B A A/B* Y(1)Hayes et al. [37] A C A C Y(2)Heisler et al. [38] A B C C Y(2)Kafy et al. [39] A A C C NKurowski and Sims [40] A B C C Y(3)Lehmann et al. [41] A C A A† Y(1)McGloin et al. [42] A B B C Y(1)Okafor [43] A C C C NPiercy et al. [44] A B C C Y(3)Satyawan et al. [45] A C C C Y(2)Stewart and Voss [46] A C A C NSwann et al. [47] A B A C Y(3)Wolff [48] A B A C Y(1)Wolff [49] A A A C Y(1)Yehia et al. [50] A C C C N
*Haller 2008: the kappa value for inter-observer agreement was 0.44 for incidence and 0.23 for preventability. The study in 2005 did not mention kappavalues.†Lehmann: kappa value for preventability.Methods of selecting medical records were A = adequate [selection method is correct and clearly described (at random or consecutive)];B = inadequate (selection method is described, but no correct method is applied); C = unclear (selection method is not mentioned). Sample size: thenumber of records reviewed was A = >1000; B = >100 and <1000; C = < 100. Method of screening was A = adequate (the medical records wereassessed using a multistage method: a primary assessment by trained clinician(s) using clear criteria. Each record that was positive for one or morecriteria was then reviewed, independently, by (two) clinicians whether or not an adverse event occurred. The assessment of causation andpreventability of adverse events was performed using a clear scale. When disagreement concerning the presence of adverse events and prevent-ability prevailed, an independent assessment was performed; B = less adequate [the medical records were assessed using a unistage or amultidisciplinary method: a primary assessment by trained clinician(s) using clear criteria. The records were reviewed by one or more clinician(s)whether or not an adverse event occurred]; C = unclear (there was insufficient information about the method of chart review). Kappa value forinter-observer agreement was A = adequate (k > 0.40); B = inadequate (k < 0.40); C = unclear (kappa values were not mentioned).Explicit definitions on unplanned ICU admissions, adverse/critical events or serious complications were mentioned: yes (Y); no (N). Y(1): studies thatapplied the following or a similar definition of adverse events as defined by the IOM or Wilson: ‘an event that results in unintended harm to the patientby an act of commission or omission rather than by underlying disease or condition of the patient’. Y(2): studies that defined an adverse event, criticalevent or serious complication as an unplanned admission to an ICU. Y(3): studies that suggest that an unplanned intensive care admissions may resultfrom unexpected events.ICU, intensive care unit; IOM, ••.
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd6
1
234
56789
101112131415161718192021222324252627282930313233343536373839404142434445464748495051
52
53545556575859
Tab
le3
Prim
ary
and
seco
ndar
you
tcom
es
Aut
hors
Par
ticip
ants
Sam
ple
size
Prim
ary
outc
omes
Sec
onda
ryou
tcom
es
Pro
vide
r(p
opul
atio
n)
Sur
gica
lpro
cedu
res/
hosp
itala
dmis
sion
s(s
tudy
perio
d)R
ecor
dsre
view
edIC
Uad
mis
sion
s(s
tudy
perio
d)
Unp
lann
edIC
Uad
mis
sion
sIC
Udi
scha
rges
AE
sre
quiri
ngIC
Ure
adm
issi
ons
AE
sre
quiri
ngIC
Uad
mis
sion
Pre
vent
able
AE
s
Mea
nle
ngth
ofIC
Ust
ay(d
ays)
for
AE
sM
orta
lity
Bar
nes
and
Hav
ill[2
6]A
8317
3pr
oced
ures
110
1929
110
––
110
(5.7
)–
–12
(10.
9)C
ulle
net
al.
[29]
A,S
1709
3pr
oced
ures
71–
7163
071
–1.
58
(11.
3)D
owne
yan
dO
’Con
nell
[30]
P-A
1497
0pr
oced
ures
3564
0(P
–A)*
35–
–35
(5.5
)7
(20)
4.2
1(2
.9)
ElS
hoba
ryet
al.
[33]
S–
250
837
––
7(8
.4)
––
0[0
]G
upta
etal
.[3
6]S
–20
2–
26–
–26
––
–H
alle
ret
al.
[22]
S44
130
proc
edur
es18
855
5220
1–
–18
3(3
.3)
140
(76.
5)†
–22
(10.
9)H
ayes
etal
.[3
7]S
244
4pr
oced
ures
23–
23–
–23
–10
.43
Hei
sler
etal
.[3
8]S
-G90
3pr
oced
ures
736
–7
––
7–
–0
[0]
Kur
owsk
iand
Sim
s[4
0]P
-A55
196
proc
edur
es35
842
076
––
76(1
8.1)
––
–O
kafo
r[4
3]S
,A
658
1pr
oced
ures
2649
726
––
26(5
.2)
–3.
28
(30.
8)P
ierc
yet
al.
[44]
A–
165
–16
5–
–16
528
(17)
–24
(14.
6)S
atya
wan
etal
.[4
5]S
1317
0pr
oced
ures
7620
476
48–
76(3
7.3)
––
28(3
6.8)
Sw
ann
etal
.[4
7]S
1855
5pr
oced
ures
265
679
34–
–34
(5)
7(2
0.6)
1.5
0(0
)Ye
hia
etal
.[5
0]S
–17
––
–2
––
–B
uist
etal
.[2
7]G
W19
853
adm
issi
ons
112
515
7540
47
(14.
6)–
5�
6.9
35(4
6.7)
Dar
chy
etal
.[2
3]G
W24
555
adm
issi
ons
623
623
6852
1–
68(1
0.9)
35(5
1.5)
6.9
�9.
39
(13.
2)D
unn
etal
.[3
1]P
103
255
adm
issi
ons
1612
–10
66–
–20
7–
–48
(23.
2)K
afy
etal
.[3
9]G
–17
92–
6–
06
––
–Le
hman
net
al.
[41]
S,
M,
P–
6457
2764
––
64(1
.1)
22(3
4.4)
38
(12.
5)M
cGlo
inet
al.
[42]
GW
1563
5ad
mis
sion
s56
3–
98‡
–11
98‡
31(3
1.6)
3.5‡
43(5
0)W
olff
[48]
GW
511
5ad
mis
sion
s49
7–
––
–14
––
–W
olff
[49]
GW
–14
6524
––
24–
––
Cha
boye
ret
al.
[28]
ICU
–30
050
7–
–8
––
––
Dur
bin
and
Kop
el[3
2]IC
U(M
,S)
–82
––
1803
83–
–8.
3�
16.1
(firs
tsta
y)34
(41)
End
acot
tet
al.
[34]
ICU
–38
8–
–38
871
––
––
Fran
klin
and
Jack
son
[35]
MIC
U–
299
512
–36
536
§–
15(4
1.6)
11§
18(5
8)§
Ste
war
tan
dVo
ss[4
6]C
CU
–44
1776
––
44–
15(3
4.1)
13.3
¶4
(9.1
)
*Dow
ney
inve
stig
ated
only
post
-ana
esth
etic
erro
rs.
Ther
ew
ere
1175
paed
iatr
icIC
Uad
mis
sion
sof
whi
ch64
0po
st-a
naes
thes
iaad
mis
sion
s.† H
alle
r:re
port
edin
2005
on18
3A
Es.
The
stud
yin
2008
repo
rted
anA
Era
teof
revi
ewer
Aof
174/
188
and
revi
ewer
Bof
164/
188
(k=
0.44
).Th
epr
even
tabi
lity
rate
sw
ere
129/
188
and
151/
188
(k=
0.23
).‡ M
cGlo
in:8
6pa
tient
sw
ard
patie
nts
wer
eun
expe
cted
lyad
mitt
edto
the
ICU
ona
tota
lof
98oc
casi
ons.
Ele
ven
patie
nts
wer
ere
adm
itted
,one
bein
gre
adm
itted
twic
e.S
urvi
vors
had
asi
mila
rIC
Ust
ay(3
days
)ve
rsus
non-
surv
ivor
s(4
days
).§ F
rank
lin:3
1pa
tient
sw
ere
read
mitt
edw
ithfiv
epa
tient
sre
adm
itted
twic
e.Th
eav
erag
ele
ngth
ofin
itial
adm
issi
onw
as5
days
(ran
ge1–
10da
ys)a
ndth
eav
erag
eof
subs
eque
ntad
mis
sion
(rea
dmis
sion
)was
6da
ys(r
ange
1–41
days
).¶S
tew
art:
aver
age
leng
thof
ICU
stay
ofth
ein
itial
(5.9
days
)an
dse
cond
(unp
lann
ed)
adm
issi
onto
the
coro
nary
care
unit
(CC
U)
(7.4
days
),ex
perie
nced
by39
patie
nts
with
unst
able
angi
nape
ctor
isor
acut
em
yoca
rdia
linf
arct
ion.
Num
bers
(per
cent
ages
).P
erce
ntag
esof
unpl
anne
dIC
Uad
mis
sion
san
dA
Es
are
calc
ulat
edag
ains
tth
enu
mbe
rof
ICU
adm
issi
ons.
Per
cent
ages
ofpr
even
tabl
eA
Es
and
mor
talit
yar
eca
lcul
ated
com
pare
dw
ithth
enu
mbe
rof
AE
sre
quiri
ngIC
U(r
e)ad
mis
sion
.A
,an
aest
hetic
;A
E,
adve
rse
even
t;S
,su
rgic
al;
GW
,ge
nera
lwar
ds;
P,pa
edia
tric
;G
,gy
neco
logi
cal;
ICU
,in
tens
ive
care
unit;
M,
med
ical
;C
CU
,co
rona
ryca
reun
it.
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 7
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Types of preventable event
In this review, it was not possible to classify the adverse events intocategories of attributable causes, as there was not enough infor-mation provided in the studies on the circumstances or contribu-tory factors leading to adverse events.
Types of events were further specified for those studies thatreported on preventability of adverse events (n = 9, Table 5). Thetypes of events were mutually exclusive and could be divided into10 groups: premature discharge on ICU, diagnostic errors, inap-propriate or inadequate treatment, technical error, adverse drugevent, inappropriate intravenous fluid therapy, problems withmedical procedures, problems with anaesthetic or surgical proce-dures, reason not apparent and other. Numbers could not directlybe compared between studies because different populations (inclu-sions and exclusions) were studied.
Preventable adverse events because of
anaesthetic-surgical care
Piercy investigated unplanned ICU admissions after anaesthesia(UIA), as this is a recommended measure by the AustralianCouncil on Healthcare Standards [59]. Only 19.3% of theunplanned admissions contained an anaesthetic contribution to theadmission of which only 5.4% within this group were deemedpossibly preventable. The majority of these events were drug-related and occurred more commonly in elective surgical patients[44]. In contrast, Haller found that 52.2% of UIA patients had anincident or near miss during their procedure of which most werepreventable (74% to 92%) [22]. Downey identified 20% UIApatients in a paediatric cohort whose ICU admission could havebeen prevented by a change in anaesthesia management. Some74.3% of those admissions emerged after elective surgery. Nopattern of preventability could be deduced because of the range inpreventable problems [30]. Finally, Swann identified 12.8% UIAof which 20.6% were deemed preventable. The majority of eventswere related to inappropriate fluid [47].
Preventable adverse events on general wards
Darchy found 68 adverse events, accounting for 10.9% of unex-pected ICU admissions of which 51% was deemed preventable.The majority of preventable adverse events was drug-related [23].In a multi-centre study by Lehmann, 66 (1.2%) patients wereidentified as having an iatrogenic medical event that was theprimary reason for ICU admission. Twenty-two (34%) cases were
thought to be preventable. The majority of events were secondaryto technical error (45%) or were drug-related (33%) [41]. Finally,McGloin reviewed 563 ICU admissions of which 98 (17.4%) wereunexpected admissions. A total of 31 cases were deemed to bepreventable and were due to incorrect or delayed diagnosis ortherapy [42].
Preventable adverse events on intensive care units
Two studies reported on the reasons for readmissions on ICU.Franklin identified 36 (12.0%) unexpected ICU readmissions ofwhich 19 (53%) were readmitted because of a recurrence of theiroriginal problem. Detailed analysis revealed that the majority ofpatients might have benefited from a longer ICU stay [35]. Finally,Stewart found 44 (2.5% of ICU admissions) unexpected readmis-sions to a coronary care unit and concluded that 15 (34.1%) caseswere related to incorrect or delayed therapy and thus could havebeen prevented.
Length of ICU stay, mortality and costs
Data reported on length of ICU stay and mortality are presented inTable 3. The mean length of ICU stay reflected the average daysthat patients who experienced an adverse event spent in the ICU.The results varied from 1.5 to 10.4 days for the patient’s first stayin ICU. For the patients who were readmitted on ICU, the ICUlength of stay, including the second stay, varied between 8.3 and13.3 days.
The percentage mortality was calculated as the number ofdeaths compared with the number of adverse events requiring ICUadmission. Mortality percentages varied between 0% and 58%.
Only one study reported on the financial costs of adverse events.In the cohort of 68 unexpected ICU admissions by Darchy, thesepatients accounted for a total of 472 days in ICU, with a meanlength of stay of 6.9 � 9.3 days (range, 1–52 days) and a 13%(9/68 patients) fatality rate. Costs of medical care in the ICU forthese patients were estimated at $688 470 [23].
DiscussionWe conducted a systematic review to investigate the incidence,preventability and consequences of adverse events that necessitateintensive care admission. It was not possible to estimate the inci-dence based on multiple studies because of considerable hetero-geneity. The percentage of surgical and medical adverse eventsthat required ICU admission ranged from 1.1% to 37.2%. ICU
Table 4 Ranges of the primary outcomes(proportions)
Primary outcomeNumber of studiesthat reported outcome
Range
Lower value Upper value
Proportion adverse events (compared withICU admission)
11 0.11 0.37
Proportion ICU readmissions (comparedwith ICU discharges)
5 0 0.18
Proportion preventable adverse events(compared with adverse events)
9 0.17 0.77
ICU, intensive care unit.
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd8
12345
6789
10111213
14
1516171819202122232425262728293031323334353637383940414243
44454647484950515253545556575859606162636465666768697071727374757677
78
79
808182838485
5 5
Tab
le5
Pre
vent
able
adve
rse
even
ts(A
Es)
clas
sifie
dby
type
ofev
ent
Aut
hors
Fran
klin
and
Jack
son
[35]
Sw
ann
etal
.[4
7]D
owne
yan
dO
’Con
nell
[30]
Ste
war
tan
dVo
ss[4
6]M
cGlo
inet
al.
[42]
Dar
chy
etal
.[2
3]Le
hman
net
al.
[41]
Pie
rcy
etal
.[4
4]H
alle
ret
al.
[22]
Pop
ulat
ion
MIC
US
P-A
CC
UG
WM
/SM
/S/P
AS
No.
ofA
Es
3634
3544
9868
64†
165
183§
No.
ofpr
even
tabl
eA
Es
157
715
31*
3522
28‡
129
to15
1C
ateg
orie
sof
type
s
Pre
mat
ure
disc
harg
eon
ICU
15D
iagn
ostic
erro
r12
6In
appr
opria
te/in
adeq
uate
trea
ting
19Te
chni
cale
rror
29A
vers
edr
ugev
ent
Dos
ing
erro
r1
89
5Id
iosy
ncra
ticre
actio
n7
Freq
uenc
yer
ror
2U
nabl
eto
clas
sify
2In
appr
opria
tedr
ug6
11
Inad
equa
tefo
llow
-up
ofdr
ugth
erap
y14
12%
Inap
prop
riate
antip
late
let
oran
ticoa
gula
ntth
erap
y15
Failu
reto
use
prop
hyla
ctic
trea
tmen
t2
Inap
prop
riate
intr
aven
ous
fluid
ther
apy
61
2P
robl
ems
with
med
ical
proc
edur
esR
adio
cont
rast
infu
sion
2H
emor
rhoi
dsc
lero
ther
apy
1P
robl
ems
with
anae
sthe
tic-s
urgi
calp
roce
dure
sS
upra
pubi
ccy
stot
omy
tube
1E
pist
axis
tam
pona
de1
Cho
ice/
appl
icat
ion
anae
sthe
tic/s
urgi
calt
echn
ique
39%
to52
%E
quip
men
t/m
onito
ring
failu
res
1%to
6%P
robl
ems
with
preo
pera
tive
asse
ssm
ent
124
%to
27%
Trau
mat
icin
tuba
tion
with
airw
aybl
eedi
ng1
Rel
ucta
nce
toco
mm
ence
inot
rope
infu
sion
intr
aope
rativ
ely
1
Rel
ucta
nce
toad
min
iste
rC
PAP
imm
edia
tely
post
-ope
rativ
ely
1
Intu
batio
nof
the
right
mai
nbr
onch
us1
Del
ayin
drai
nage
ofbi
late
ralp
neum
otho
race
s1
Inad
equa
tepr
eope
rativ
ere
spira
tory
oper
atio
n1
Asp
iratio
n1
Pul
mon
ary
oede
ma
1R
easo
nno
tap
pare
nt4
Oth
er2
*McG
loin
defin
edsu
bopt
imal
care
asei
ther
non-
reco
gniti
onof
(the
seve
rity
of)a
nab
norm
ality
clea
rlyap
pare
ntfr
omph
ysio
logi
calr
ecor
ding
sor
labo
rato
ryda
ta(d
iagn
ostic
erro
r)or
inap
prop
riate
/in
adeq
uate
trea
tmen
t.† L
ehm
ann:
Cat
egor
ies
for
AE
sar
epr
esen
ted.
Cat
egor
ies
for
prev
enta
ble
AE
sar
eno
tav
aila
ble.
Tech
nica
lerr
orw
asde
fined
asa
med
ical
proc
edur
eev
ent
such
asan
inju
ryoc
curr
ing
durin
gan
oper
atio
nor
blee
ding
.Adv
erse
drug
even
twas
defin
edas
anin
jury
resu
lting
from
com
plic
atio
nsof
noni
nvas
ive
ther
apy.
Dia
gnos
ticer
rorw
asde
fined
asa
dela
yed
orin
corr
ectd
iagn
osis
orth
erap
y.‡ P
ierc
y:on
lyty
pes
ofan
aest
hetic
-rel
ated
prob
lem
sw
ere
give
n.§ H
alle
r:nu
mbe
rsor
perc
enta
ges
wer
egi
ven
for
two
inde
pend
ent
revi
ewer
s.A
tota
lof
105
patie
nts
with
UIA
had
anin
cide
ntor
near
mis
sdu
ring
thei
rpr
oced
ure.
Inne
arly
half
ofth
eca
ses,
ther
ew
asm
ore
than
one
even
tre
port
ed.
Num
bers
orpe
rcen
tage
sw
ere
give
n.C
PAP,
••;
ICU
,in
tens
ive
care
unit;
M,
med
ical
;S
,su
rgic
al;
P,pa
edia
tric
;A
,an
aest
hetic
;C
CU
,co
rona
ryca
reun
it;G
W,
gene
ralw
ard;
UIA
,un
plan
ned
ICU
adm
issi
ons
afte
ran
aest
hesi
a/su
rger
y.
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 9
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
readmissions varied from 0% to 18.3%. Furthermore, the prevent-ability of the adverse events varied from 17% to 76.5%. Conse-quences of the adverse events included a mean length of ICU staythat ranged from 1.5 to 10.4 days for the patient’s first stay in ICUand mortality percentages between 0% and 58%. Authors sug-gested several systems to reduce adverse events that necessitateintensive care admission but the effectiveness of such systems stillneeds to be researched.
We performed a comprehensive search strategy. We searchedseveral electronic sources and put considerable effort in identify-ing studies using alternative sources such as reference lists, patientsafety journals and websites. We also contacted the authors ofrelevant papers. The additional sources yielded nine additionalstudies. The rigorous search strategy strengthens the methodologi-cal quality of our review.
We found considerable heterogeneity between the studies,which precluded pooling of study results. For all outcomes, the I2
was close to 100%. As I2 describes the percentage of the variabilityin effect estimates that is due to heterogeneity rather than samplingerror [21], this means high diversity between the studies. Subgroupanalyses based on population, country and methodological qualityof the studies could not clarify heterogeneity. However, care mustbe taken in the interpretation of these analyses, as the includedstudies have small sample sizes and few studies report on the sameoutcomes.
Clinical diversity may be due to population mix (variability intype of participants, e.g. surgical, paediatric, gynecological, andinclusions and exclusions of participants) and the use of differentdefinitions on outcomes. Institutional factors like bed size (e.g.high dependency beds outside ICU), teaching status or staffingmay also play an important role. These clinical characteristicscould not be investigated because of the poor reporting of studyresults. Differences in study design and methods of medical recordreview may explain methodological diversity. Only three studieswere multi-centre studies [26,41,44]. All other studies were casestudies investigating unplanned admissions on one ICU.
The limitations of this review mainly concern the scope of thereview. This review only includes studies that used the investiga-tion of adverse events through chart review. Our strict inclusioncriteria potentially may have caused us to exclude interestingstudies with prospective study designs or studies addressing inci-dent reporting.
The use of the method of medical record review itself might leadto an underestimation of adverse events. The quality of the medicalrecords is often poor as too much information is missing or incom-plete. Because many errors are not (well) documented in themedical records, study results may be an underestimation of thetrue incidence rate of adverse events leading to unanticipatedintensive care admission. In addition, estimates on preventabilityare prone to subjective judgments of the authors. Darchy mentionsthe difficulty in distinguishing between an adverse event and apreventable adverse event after surgery because the surgeon is bothjudge and judged. Haller made in his study a classification ofpreventable complications with a rather moderate level of exactagreement between reviewers of 79.7%, with a kappa = 0.23 (95%CI: 0.16 to 0.29) [22]. In the multi-centre study of Lehmann, thepreventability of the events could not be evaluated with certainty in52% of the cases [41]. Several studies warn that the review of eachcase via the medical records is hampered by the retrospective
nature of the analysis, and by the dependence on the quality of notetaking by anaesthetics, intensivists and nursing staff [30,33]. Oneparticular difficulty that was mentioned was the lack of recordingin the notes of the actual reason for admission to ICU. This mayarise from the reluctance of house staff, perhaps from medico legalreasons, to record the nature of misadventures under anaesthesia[41]. These measurement errors in the measurement of error, asLilford describes [51], can degrade the ability to measure theimpact of interventions or to provide evidence of association orcausality between processes of care and outcome.
Despite the limitations of the available literature, several keyfindings suggest the need for further high-quality research in thedetection and classification of adverse events requiring ICUadmission. The following suggestions can be made for futureresearch.
Consecutive selection of ICU admissions allows that any infor-mation missing in the medical charts can be collected prospec-tively. Patient demographic data, the reason for ICU admission,severity of illness score, for example APACHE II score (AcutePhysiology and Chronic Health Evaluation), hospital outcome(mortality, discharge, transfer) and length of ICU stay could bebetter documented in a follow-up study. In addition, this studydesign allows a better comparison between patients with plannedand unplanned ICU admissions [27,29,38].
A multidisciplinary approach with professional involvement is amain condition in conducting medical chart review. In addition, adoctor-led programme in contrast to nursing-led initiatives mightaffect the acceptance of such a programme in other settings [31].
Several studies have shown some strengths of the method as anauditing tool [31,39,40,42,43,45,47–50]. Continuous medicalrecord review can be a useful strategy in quality improvement [31],as the detection of adverse events, followed by an in-depth analysisof the underlying causes and specific prevention strategies, canreduce the occurrence of adverse in hospitalized patients.However, in this review only one study predominated thisapproach [23]. It is a challenge to better understand the failures inthe organization of care and an important source of informationlies with the health care providers and with the patient. Within thisissue, an important question arises whether it is appropriate tomeasure actual harm (adverse events) or the causes that lead toadverse events (errors). In this context, Brown et al. propose anexcellent framework for study design and measurement of errors[52].
Although higher heterogeneity must be dealt with, multi-centrestudies also allow the possibility to aggregate data and analysepatterns of factors leading to adverse events across different hos-pital settings. This approach demands an appropriate design asthere is an important loss of power that results from greater simi-larity across individuals within a cluster than across individualsbetween clusters [52]. The study of Hillman et al., which investi-gates the impact of a medical emergency team on the incidence ofunplanned ICU admissions in 23 Australian hospitals, applied anillustrative evaluation framework [53].
There is a need for a better standardization and validation ofrecommended improvements aimed at reducing morbidity andmortality arising from unplanned ICU admissions. In this regard,McGloin favours the early recognition and correction of abnor-malities by introducing a rapid response team or medical emer-gency team [42]. Other authors support these findings [27,28].
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd10
123456789
1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859
However, a recent systematic review on this issue showed thatthere is minimal evidence on the effectiveness of these systems[54]. Swann recommends grand rounds (educational training) onthe management of perioperative fluid administration [47]. Inorder to reduce ICU readmissions, Stewart suggests the implemen-tation of formal mechanisms to ensure appropriate pharmaco-therapy when patients are transferred to lower levels of care [46].Several authors suggest from their experience that a progressivecare (‘step-down’) unit may be an effective alternative to earlyICU discharge by reducing the likelihood of ‘premature’ ICUdischarge and, hence, reducing readmissions to the ICU[28,35,47]. However, the effectiveness of any of these systems stillneeds to be confirmed [54].
Conclusions
Implications for practice
The currently available evidence is relatively weak and it is notpossible to estimate the overall incidence and preventability rate ofadverse events that necessitate ICU admission. Variability in meth-odology and definitions, and poor reporting in studies may be themain reasons, so this can be researched in order to improve qualityof care. Hospitals should better document the reasons for admis-sion on ICUs. Several authors recommend early detection ofpatients with clinical instability on general wards and the imple-mentation of rapid response teams. Step-down or intermediate careunits can be a useful strategy for patients who require monitoringto avoid ICU readmissions. However, the effectiveness of any ofthese systems still needs to be confirmed.
Implications for future research
There is a need for further studies on the detection of adverseevents. Planning of future studies should aim to standardize ter-minology and measures of outcomes (standard taxonomy) and toapply more explicit study designs in order to allow for compari-sons across studies. This area of research is important in order toidentify and explain failure of health care systems leading topatient harm, with the ultimate aim to improve the quality andsafety of care.
Acknowledgements
Funding
This review was supported by Limburg Sterk Merk.
Conflict of interests
There were no conflicts of interests.
Contribution of authors
AV conceived the study, carried out the literature searches and isguarantor. AV and GEB designed the protocol, analysed and inter-preted the data. GEB supported with methodological advice. AVand SV selected the studies, assessed the included studies andcollected the data. AV draughted the manuscript. SV contributed to
the writing of the paper. GEB, WS, NC and JH critically revisedthe manuscript for important intellectual content. All authorsapproved the final version to be published.
References1. World Health Organization ••. Available at: http://www.who.int/
patientsafety (last accessed 28 September 2010).2. European Commission ••. Available at: http://ec.europa.eu/health/
ph_systems/patient_safety_en.htm (last accessed 28 September 2010).3. The Joint Commission ••. Available at: http://www.jointcommission.
org/PatientSafety/Solutions/ (last accessed 28 September 2010).4. Zhan, C. & Miller, M. R. (2003) Excess length of stay, charges, and
mortality attributable to medical injuries during hospitalization.Journal of the American Medical Association, 290 (14), 1868–1874.
5. Brennan, T. A., Leape, L. L., Laird, N. M., Hebert, L., Localio, A. R.,Lawthers, A. G., et al (1991) Incidence of adverse events and negli-gence in hospitalized patients. Results of the Harvard Medical PracticeStudy I. New England Journal of Medicine, 324 (6), 370–376.
6. Davis, P., Lay-Yee, R., Briant, R., Ali, W., Scott, A. & Schug, S.(2002) Adverse events in New Zealand public hospitals I: occurrenceand impact. New Zealand Medical Journal, 115 (1167), U271.
7. Michel, P., Quenon, J. L., de Sarasqueta, A. M. & Scemama, O. (2004)Comparison of three methods for estimating rates of adverse eventsand rates of preventable adverse events in acute care hospitals. BritishMedical Journal, 328 (7433), 199.
8. Schioler, T., Lipczak, H., Pedersen, B. L., Mogensen, T. S., Bech, K.B., Stockmarr, A., et al. (2001) [Incidence of adverse events in hos-pitals. A retrospective study of medical records]. Ugeskr Laeger, 163(39), 5370–5378.
9. Soop, M., Fryksmark, U., Koster, M. & Haglund, B. (2009) Theincidence of adverse events in Swedish hospitals: a retrospectivemedical record review study. International Journal for Quality inHealth Care, 21 (4), 285–291.
10. Thomas, E. J., Studdert, D. M., Burstin, H. R., Orav, E. J., Zeena, T.,Williams, E. J., et al. (2000) Incidence and types of adverse events andnegligent care in Utah and Colorado. Medical Care, 38 (3), 261–271.
11. Wilson, R. M., Runciman, W. B., Gibberd, R. W., Harrison, B. T.,Newby, L. & Hamilton, J. D. (1995) The quality in australian healthcare study. Medical Journal of Australia, 163 (9), 458–471.
12. Zegers, M., de Bruijne, M. C., Wagner, C., Hoonhout, L. H., Waaij-man, R., Smits, M., et al. (2009) Adverse events and potentially pre-ventable deaths in Dutch hospitals: results of a retrospective patientrecord review study. Quality & Safety in Health Care, 18 (4), 297–302.
13. Baker, G. R., Norton, P. G., Flintoft, V., Blais, R., Brown, A., Cox, J.,et al. (2004) The Canadian Adverse Events Study: the incidence ofadverse events among hospital patients in Canada. CMAJ, 170 (11),1678–1686.
14. Zegers, M., de Bruijne, M. C., Wagner, C., Groenewegen, P. P., Waaij-man, R. & van der Wal, G. (2007) Design of a retrospective patientrecord study on the occurrence of adverse events among patients inDutch hospitals. BMC Health Services Research, 7, 27.
15. JBI. The Joanna Briggs Institute ••. Available at: http://www.joannabriggs.edu.au (last accessed 28 September 2010).
16. Aspden, P. C. J., Wolcott, J. & Erickson, S. M. (2004) Patient Safety:Achieving a New Standard for Care. Washington, DC: The NationalAcademies Press.
17. Vincent, C., Neale, G. & Woloshynowych, M. (2001) Adverse eventsin British hospitals: preliminary retrospective record review. BritishMedical Journal, 322 (7285), 517–519.
18. AHRQ Patient Safety Network ••. Available at: http://www.psnet.ahrq.gov/ (last accessed 28 September 2010).
19. IHI. Institute for Healthcare Improvement ••. Available at: http://www.ihi.org/ihi (last accessed 28 September 2010).
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 11
123456789
101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100
101
66
77
88
99
1010
1111
1212
1313
20. Lipsey, M. W. & Wilson, D. B. (2001) Practical Meta-analysis. Thou-sand Oaks, CA: Sage Publications.
21. Deeks, J. J., Higgins, J. P. & Altman, D. G. Analysing data andundertaking meta-analyses. Wiley-Blackwell, editor. Cochrane hand-book for systematic reviews of interventions. 2008.
22. Haller, G., Myles, P. S., Langley, M., Stoelwinder, J. & McNeil, J.(2008) Assessment of an unplanned admission to the intensive careunit as a global safety indicator in surgical patients. Anaesthesia andIntensive Care, 36 (2), 190–200.
23. Darchy, B., Le Miere, E., Figueredo, B., Bavoux, E. & Domart, Y.(1999) Iatrogenic diseases as a reason for admission to the intensivecare unit: incidence, causes, and consequences. Archives of InternalMedicine, 159 (1), 71–78.
24. Darchy, B., Le Miere, E., Figueredo, B., Bavoux, E., Cadoux, G. &Domart, Y. (1998) [Patients admitted to the intensive care unit foriatrogenic disease. Risk factors and consequences]. La Revue deMedecine Interne, 19 (7), 470–478.
25. Haller, G., Myles, P. S., Wolfe, R., Weeks, A. M., Stoelwinder, J. &McNeil, J. (2005) Validity of unplanned admission to an intensive careunit as a measure of patient safety in surgical patients. Anesthesiology,103 (6), 1121–1129.
26. Barnes, P. J. & Havill, J. H. (1980) Anaesthetic complications requir-ing intensive care. A five year review. Anaesthesia and Intensive Care,8 (4), 404–409.
27. Buist, M. D., Jarmolowski, E., Burton, P. R., Bernard, S. A., Waxman,B. P. & Anderson, J. (1999) Recognising clinical instability in hospitalpatients before cardiac arrest or unplanned admission to intensive care.A pilot study in a tertiary-care hospital. Medical Journal of Australia,171 (1), 22–25.
28. Chaboyer, W., Thalib, L., Foster, M., Ball, C. & Richards, B. (2008)Predictors of adverse events in patients after discharge from the inten-sive care unit. American Journal of Critical Care, 17 (3), 255–263;quiz 64.
29. Cullen, D. J., Nemeskal, A. R., Cooper, J. B., Zaslavsky, A. & Dwyer,M. J. (1992) Effect of pulse oximetry, age, and ASA physical status onthe frequency of patients admitted unexpectedly to a postoperativeintensive care unit and the severity of their anesthesia-related compli-cations. Anesthesia and Analgesia, 74 (2), 181–188.
30. Downey, G. B. & O’Connell, A. J. (1996) Audit of unbooked paedi-atric post-anaesthesia admissions to intensive care. Anaesthesia andIntensive Care, 24 (4), 464–471.
31. Dunn, K. L., Reddy, P., Moulden, A. & Bowes, G. (2006) Medicalrecord review of deaths, unexpected intensive care unit admissions,and clinician referrals: detection of adverse events and insight into thesystem. Archives of Disease in Childhood, 91 (2), 169–172.
32. Durbin, C. G., Jr & Kopel, R. F. (1993) A case-control study ofpatients readmitted to the intensive care unit. Critical Care Medicine,21 (10), 1547–1553.
33. El Shobary, H., Backman, S., Christou, N. & Schricker, T. (2008) Useof critical care resources after laparoscopic gastric bypass: effect onrespiratory complications. Surgery for Obesity and Related Diseases,4 (6), 698–702.
34. Endacott, R., Chaboyer, W., Edington, J. & Thalib, L. (2010) Impactof an ICU Liaison Nurse Service on major adverse events in patientsrecently discharged from ICU. Resuscitation, 81 (2), 198–201.
35. Franklin, C. & Jackson, D. (1983) Discharge decision-making in amedical ICU: characteristics of unexpected readmissions. CriticalCare Medicine, 11 (2), 61–66.
36. Gupta, R. M., Parvizi, J., Hanssen, A. D. & Gay, P. C. (2001) Post-operative complications in patients with obstructive sleep apnea syn-drome undergoing hip or knee replacement: a case-control study.Mayo Clinic Proceedings, 76 (9), 897–905.
37. Hayes, C., Ambazidis, S. & Gani, J. S. (1996) Intensive care unitadmissions following laparoscopic surgery: what lessons can belearned? Australian and New Zealand Journal of Surgery, 66 (4),206–209.
38. Heisler, C. A., Weaver, A. L. & Melton, L. (2009) J, 3rd, Gebhart JB.Effect of additional reconstructive surgery on perioperative and post-operative morbidity in women undergoing vaginal hysterectomy.Obstetrics and Gynecology, 114 (4), 720–726.
39. Kafy, S., Huang, J. Y., Al-Sunaidi, M., Wiener, D. & Tulandi, T.(2006) Audit of morbidity and mortality rates of 1792 hysterectomies.Journal of Minimally Invasive Gynecology, 13 (1), 55–59.
40. Kurowski, I. & Sims, C. (2007) Unplanned anesthesia-related admis-sions to pediatric intensive care – a 6-year audit. Paediatric Anaesthe-sia, 17 (6), 575–580.
41. Lehmann, L. S., Puopolo, A. L., Shaykevich, S. & Brennan, T. A.(2005) Iatrogenic events resulting in intensive care admission: fre-quency, cause, and disclosure to patients and institutions. AmericanJournal of Medicine, 118 (4), 409–413.
42. McGloin, H., Adam, S. K. & Singer, M. (1999) Unexpected deathsand referrals to intensive care of patients on general wards. Are somecases potentially avoidable? Journal of the Royal College of Physi-cians of London, 33 (3), 255–259.
43. Okafor, U. (2009) An audit of unplanned postoperative intensive careunit admissions in Enugu, Nigeria: causes and outcome. SouthernAfrican Journal of Critical Care, 25 (1), 16–19.
44. Piercy, M., Lau, S., Loh, E., Reid, D., Santamaria, J. & Mackay, P.(2006) Unplanned admission to the intensive care unit in postoperativepatients – an indicator of quality of anaesthetic care? Anaesthesia andIntensive Care, 34 (5), 592–598.
45. Satyawan, A. S. V. & Chaudhari, L. (2006) Audit of intensive care unitadmissions from the operating room. Indian Journal of Anaesthesia,50 (3), 193–200.
46. Stewart, S. & Voss, D. W. (1997) A study of unplanned readmissionsto a coronary care unit. Heart & Lung, 26 (3), 196–203.
47. Swann, D., Houston, P. & Goldberg, J. (1993) Audit of intensive careunit admissions from the operating room. Canadian Journal of Ana-esthesia, 40 (2), 137–141.
48. Wolff, A. M. (1995) Limited adverse occurrence screening: an effec-tive and efficient method of medical quality control. Journal of Qualityin Clinical Practice, 15 (4), 221–233.
49. Wolff, A. M. (1996) Limited adverse occurrence screening: usingmedical record review to reduce hospital adverse patient events.Medical Journal of Australia, 164 (8), 458–461.
50. Yehia, M., McDonald, M. & Walker, R. (2002) The management andoutcome of occluded haemodialysis access: a retrospective audit. NewZealand Medical Journal, 115 (1166), U258.
51. Lilford, R. J., Mohammed, M. A., Braunholtz, D. & Hofer, T. P. (2003)The measurement of active errors: methodological issues. Quality &Safety in Health Care, 2 (Suppl), ii8–i12.
52. Brown, C., Hofer, T., Johal, A., Thomson, R., Nicholl, J., Franklin, B.D., et al. (2008) An epistemology of patient safety research: a frame-work for study design and interpretation. Part 3. End points and mea-surement. Quality & Safety in Health Care, 17 (3), 170–177.
53. Hillman, K., Chen, J., Cretikos, M., Bellomo, R., Brown, D., Doig, G.,et al. (2005) Introduction of the medical emergency team (MET)system: a cluster-randomised controlled trial. Lancet, 365 (9477),2091–2097.
54. McGaughey, J., Alderdice, F., Fowler, R., Kapila, A., Mayhew, A. &Moutray, M. (2007) Outreach and Early Warning Systems (EWS) forthe prevention of intensive care admission and death of critically illadult patients on general hospital wards. Cochrane Database of Sys-tematic Reviews, (3), CD005529.
jep_1612
Incidence and preventability A. Vlayen et al.
© 2010 Blackwell Publishing Ltd12
123456789
1011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162
1414
1515
Appendix I MEDLINE search termsThe following terms were used in the MEDLINE search strategy. This search strategy was translated into the other databases using theappropriate controlled vocabulary as applicable.
Search terms
Studies on adverse events1 ADVERSE EVENT2 PREVENTable ADVERSE EVENT3 INCIDENT4 CRITICAL EVENT5 ADVERSE OUTCOME6 ADVERSE EFFECTS (Mesh)7 ADVERSE DRUG EVENT8 MEDICAL ERRORS (Mesh)9 MEDICATION ERRORS (Mesh)
10 DIAGNOSTIC ERRORS (Mesh)11 OR/1-10
Studies on intensive care12 INTENSIVE CARE (Mesh)13 INTENSIVE CARE UNITS (Mesh)14 CRITICAL CARE (Mesh)15 CRITICAL ILLNESS (Mesh)16 ICU17 OR/12-16
Studies on medical record review18 MEDICAL RECORDS (Mesh)19 MEDICAL AUDIT (Mesh)20 MEDICAL RECORD REVIEW21 CHART REVIEW22 TRIGGER TOOL23 OR/18-2224 11 AND 17 AND 23
Studies on unplanned ICU admissions25 UNPLANNED26 UNEXPECTED27 UNANTICIPATED28 OR/25-2729 28 AND 1730 24 OR 29
jep_1612
A. Vlayen et al. Incidence and preventability
© 2010 Blackwell Publishing Ltd 13
AUTHOR QUERY FORM
Dear Author,During the preparation of your manuscript for publication, the questions listed below have arisen. Please attend to these
matters and return this form with your proof.Many thanks for your assistance.
QueryReferences
Query Remark
q1 AUTHOR: A running head short title was not supplied; please check if this one issuitable and, if not, please supply a short title of up to 40 characters that can be usedinstead.
q2 AUTHOR: Please provide title for the corresponding author.
q3 AUTHOR: Please provide the qualification for Annemie Vlayen.
q4 AUTHOR: Zahn et al. has been changed to Zhan and Miller so that this citation matchesthe Reference List. Please confirm that this is correct.
q5 AUTHOR: Reference [59] has not been included in the Reference List, please supplyfull publication details.
q6 AUTHOR: Please check the website addresses/URLs and confirm that they are correctfor References 1–3. (Please note that it is the responsibility of the author(s) to ensurethat all URLs given in this article are correct and useable.) Please provide the onlinearticle titles.
q7 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 5.
q8 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 8.
q9 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 10.
q10 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 12.
q11 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 13.
q12 AUTHOR: Please check this website address/URL and confirm that it is correct forReference 15. (Please note that it is the responsibility of the author(s) to ensure that allURLs given in this article are correct and useable.) Please provide the online articletitle.
q13 AUTHOR: Please check the website addresses/URLs and confirm that they are correctfor References 18 & 19. (Please note that it is the responsibility of the author(s) toensure that all URLs given in this article are correct and useable.)
q14 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 52.
q15 AUTHOR: List up to 10 authors, for more than 10 authors list first three plus et al. forReference 53.
Toppan Best-set Premedia LimitedJournal Code: JEP Proofreader: JasonArticle No: 1612 Delivery date: 3 December 2010Page Extent: 13 Copyeditor: Elaine
q16 AUTHOR: Please confirm the authors and reference citations are correct.
q17 AUTHOR: Please confirm the authors and reference citations are correct.
q18 AUTHOR: Please define IOM.
q19 AUTHOR: Please confirm the authors and reference citations are correct.
q20 AUTHOR: Please confirm the structure of Table 5 is correct.
q21 AUTHOR: Please confirm the authors and reference citations are correct.
q22 AUTHOR: Please define CPAP.
q23 AUTHOR: Please note that this proof exceeds the journal’s free pages allocation (7pp)and will be subject to a charge for excess pages. You have the following options:Remove/edit content from your proof in order to make it 7pp or less; Pay excess pagecharges when the article is published in an issue of the journal.
MARKED PROOF
Please correct and return this set
Instruction to printer
Leave unchanged under matter to remain
through single character, rule or underline
New matter followed byor
or
or
or
or
or
or
or
or
and/or
and/or
e.g.
e.g.
under character
over character
new character new characters
through all characters to be deleted
through letter orthrough characters
under matter to be changedunder matter to be changedunder matter to be changedunder matter to be changedunder matter to be changed
Encircle matter to be changed
(As above)
(As above)
(As above)
(As above)
(As above)
(As above)
(As above)
(As above)
linking characters
through character orwhere required
between characters orwords affected
through character orwhere required
or
indicated in the marginDelete
Substitute character orsubstitute part of one ormore word(s)
Change to italicsChange to capitalsChange to small capitalsChange to bold typeChange to bold italicChange to lower case
Change italic to upright type
Change bold to non-bold type
Insert ‘superior’ character
Insert ‘inferior’ character
Insert full stop
Insert comma
Insert single quotation marks
Insert double quotation marks
Insert hyphenStart new paragraph
No new paragraph
Transpose
Close up
Insert or substitute spacebetween characters or words
Reduce space betweencharacters or words
Insert in text the matter
Textual mark Marginal mark
Please use the proof correction marks shown below for all alterations and corrections. If you
in dark ink and are made well within the page margins.wish to return your proof by fax you should ensure that all amendments are written clearly