incidence and preventability of adverse events requiring intensive care admission: a systematic...

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

m100

New Stamp23

m100

New Stamp2

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

m100

New Stamp1

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



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

m100

New Stamp16

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



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



1

234

56789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051

52

53545556575859

m100

New Stamp17

m100

New Stamp18

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



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

m100

New Stamp19

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



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



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

m100

New Stamp20

m100

New Stamp21

m100

New Stamp22

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



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



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



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



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.





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.)



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.


q18 AUTHOR: Please define IOM.


q20 AUTHOR: Please confirm the structure of Table 5 is 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

incidence and preventability of adverse events requiring intensive care admission: a systematic...

Documents