Implementation of an electronic checklist in the ICU: Association with improved outcomes

(1)

HAL Id: hal-01791633

https://hal.archives-ouvertes.fr/hal-01791633

Submitted on 15 Apr 2019

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Implementation of an electronic checklist in the ICU:

Association with improved outcomes

Gary Duclos, Laurent Zieleskiewicz, Francois Antonini, Djamel Mokart,

Veronique Paone, Marie Helene Po, Coralie Vigne, Emmanuelle Hammad,

Frederic Potie, Claude Martin, et al.

To cite this version:

Gary Duclos, Laurent Zieleskiewicz, Francois Antonini, Djamel Mokart, Veronique Paone, et al..

Im-plementation of an electronic checklist in the ICU: Association with improved outcomes. Anaesthesia

Critical Care & Pain Medicine, Elsevier Masson, 2018, 37 (1), pp.25-33. �10.1016/j.accpm.2017.04.006�.

�hal-01791633�

(2)

Original

Article

Implementation

of

an

electronic

checklist

in

the

ICU:

Association

with

improved

outcomes

Gary

Duclos

a

,

Laurent

Zieleskiewicz

a

,

Franc¸ois

Antonini

a

,

Djamel

Mokart

b

,

Ve´ronique

Paone

a

_,

_Marie

_He´le`ne

_Po

a

_,

_Coralie

_Vigne

a

_,

_Emmanuelle

_Hammad

a

_,

Fre´de´ric

Potie´

a

,

Claude

Martin

a

,

Sophie

Medam

a

,

Marc

Leone

a,

*

a

Serviced’anesthe´sieetdere´animation,Aix-Marseilleuniversite´,hoˆpitalnord,Assistancepublique–HoˆpitauxdeMarseille,13015Marseille,France b

Serviced’anesthe´sieetdere´animation,institutPaoli-Calmettes,13015Marseille,France

1. Introduction

Infectiouscomplicationsremaina frequentissueinintensive careunits(ICU),withtheuseofinvasivedevicesrepresentingthe mainriskfactor [1,2].Guidelinesaimtoimprovepractices,and thus reduce therisk of healthcare-relatedcomplications [3–6]. However,theadherencetoguidelinesisoftensuboptimal[7–9].In apreviousstudy,weobservedacompliancerateof24%,anditwas greatly reduced in patients requiring more than three clinical recommendations[10].

A number of strategies have beenimplemented to improve guideline adherence. For this purpose, a systematic electronic

checklistwasintroducedinourICU.Checklistsarewidelyusedin theoperatingroom,andtheirimplementationwasreportedtobe associatedwithareducednumberofperioperativecomplications

[11–14]. Some studies have reported encouraging ﬁndings regarding the impact of checklist implementation in the ICU

[15–18].Inmoststudies,interventionstargetspeciﬁcsubgroupsof patientsordiseases[19–21],buttheirimplementationresultsin improvedadherencetoguidelines[22,23].

Ourhypothesiswasthattheimplementationofanelectronic checklist wouldreducetherateofhealthcare-relatedinfections. The ﬁrst goal of this study was to determine the effects of electronic checklist implementation on the rate of nosocomial ventilator-associatedpneumonia(VAP).Thesecondarygoalswere toassessthedurationsofICUstay,invasivemechanicalventilation, centralvenouscatheterisation,andurinarycatheterisationbefore and after checklist implementation. In another analysis of the

ARTICLE INFO

Articlehistory:

Availableonline10July2017

Keywords: Checklists Nosocomial Infection Compliance Guidelines ABSTRACT

Objective:Toassesstheimpactofanelectronicchecklistduringthemorningroundson

ventilator-associatedpneumonia(VAP)intheintensivecareunit(ICU).

Patientsandmethods:Weconductedaretrospective,before/afterstudyinasingleICUofauniversity

hospital.AsystematicelectronicchecklistfocusingonguidelinesadherencewasintroducedinJanuary

2012.FromJanuary2008toJune2014,wescreenedpatientswithICUstaydurationsofatleast48hours.

Propensityscore-matchedanalysiswithconditionallogisticregressionwasusedtocomparetherateof

VAPandnumberofdaysfreeofinvasivedevicesbeforeandafterimplementationoftheelectronic

checklist.

Results:Weanalysed1711patients(beforegroup,n=761;aftergroup,n=950).TheratesofVAPwere

21%and11%in thebeforeand aftergroups,respectively(p<0.001).Inpropensity-score matched

analysis(n=742ineachgroup),VAPoccurredin151patients(21%)duringthebeforeperiodcompared

with 72 patients (10%) during the after period (odds ratio [OR]=0.38; 95% conﬁdence interval

[CI]=0.27–0.53).TheaftergroupshowedincreasesinICU-freedays(OR=1.05;95%CI=1.04–1.07)and

mechanicalventilation-freedays(OR=1.03;95%CI=1.01–1.04).

Conclusion:Inthismatchedbefore/afterstudy,implementationofanelectronicchecklistwasassociated

withpositiveeffectsonpatientoutcomes,especiallyonVAP.Furtherprospectivestudiesareneededto

conﬁrmtheseobservations.

C 2017Socie´te´ franc¸aised’anesthe´sieetdere´animation(Sfar).PublishedbyElsevierMassonSAS.All

rightsreserved.

* Correspondingauthorat:Serviced’anesthe´sieetdere´animation,hoˆpitalNord, chemindesBourrely,13015Marseille,France.

E-mailaddress:marc.leone@ap-hm.fr(M.Leone).

http://dx.doi.org/10.1016/j.accpm.2017.04.006

(3)

completedchecklist,wedeterminedthecomplianceforeachitem toevaluateguidelineadherenceinourunit.

2. Materialandmethods 2.1. Checklistdatacollection

The electronic checklist combines the clinical recommenda-tions initially extractedfromour previous one-daysurvey [10]

(Fig.1).Inthesecondstep,amulti-professionalgroupincluding physicians,certifiedregisterednurseanesthetists(CRNA),nurses, andphysiotherapistsdiscussedeachrecommendationaccording to our local experience and challenges. The process required 3months.Oncethe14recommendationshadbeendefined,they weremadeavailableviaanelectronicprogramaccessibletoallICU membersviaourintranet.Afterspecifictrainingusingelectronic tablets, four CRNAs (responsible for nursing care in our ICU) completedtheelectronicchecklistsforeachpatientat07:00.The electronicchecklistswereimmediatelyavailabletoallofourstaff viaourintranetoncomputers,smartphones,andtablets,among others.Duringthemorningrounds,thenurses,CRNAs,orresidents reiteratedtheunresolveditemsfromtheelectronicchecklists.

Theitemsontheelectronicchecklistcoveredinvasivedevices (centralvenouscatheter,arterialcatheter,peripheralvenousline, urinarycatheter)andtheirindications,VAPprevention (head-of-bed elevation, tube cuff pressure control, ventilation pressure<30cm H2O, daily sedation interruption), physical

rehabilitation(enteralfeeding,enteralroute,transferfrombed to armchair), and miscellaneous care (contention, type of mattress,antithromboticprophylaxis).

We usedafour-gradeclassiﬁcationfor eachinvasivedevice: ‘‘present and needed,’’ ‘‘present and in excess,’’ ‘‘present and probably in excess,’’ or ‘‘none.’’ ‘‘Present and in excess’’ was reportedbytheCRNAsduringthe07:00round.‘‘Presentandlikely

inexcess’’correspondedtoanimplementeddevicewithoutany justiﬁcation.Thedecisiontoattributeagradeforeachitemwasleft atthediscretionoftheCRNAs.Then,itwasdiscussedwiththestaff, includingphysiciansandnurses,orduringICUdailyrounds.The usualrecommendationswereclassiﬁedas:‘‘done,’’‘‘notdone,’’or ‘‘notapplicable.’’

2.2. Patients

From January 2008 to December 2014, we conducted a retrospective, before/after study in a 15-bedICU of a 968-bed universityhospital(HoˆpitalNord,Marseille,France).Ourchecklist wasintroducedonJanuary1,2012.Thebeforeperiodrangedfrom January1,2008,toDecember31,2011.Theafterperiodranged fromJanuary1,2012toDecember31,2014.

2.3. Datacollection

Demographic and clinical features, including gender, age, simpliﬁedacutephysiologyscore2(SAPS2),reasonforadmission, useofantibioticsintheﬁrst48hours,andimmunologicalstatus (neutropenia, immunosuppressive therapeutics), werecollected from our electronic database. We also collected the types of invasivedevicesused(venouscentralcatheter,urinarycatheter, tracheal intubation).We recorded the duration of use of each invasive device during the ICU stay, the duration of invasive mechanicalventilation,andthedurationofICUstay.Wenotedthe reintubationrateandtheICUmortalityrate.Thenumbersof ICU-freedaysand invasivedevice-freedays (mechanicalventilation, centralvenouscatheter,urinarycatheter)werereportedfromday of admission or of exposure today 28 or theday of death, as applicable.

Nosocomial infections included VAP, catheter-related infec-tions, bacteraemia, and urinary catheter-related infection, as

PART I. INVASIVE DEVICE

Use of central venous catheter: YES NO USE IN EXCES

Hemodynamic instability (use of catecholamine) Use of venotoxic treatment

Impossible use of peripheral venous line Use of parenteral feeding

Use of arterial catheter: YES NO USE IN EXCES

More than 3 blood samples Daily blood sample

Use of peripheral venous line : YES NO USE IN EXCES

more than 3 days old

Use of urinary catheter : YES NO USE IN EXCES

Renal dysfunction Anatomic issue Woman

PART II. RESPIRATORY CHECK

Head of bed elevation : DONE NOT DONE NOT APPLICABLE

Instable trauma rachis Pelvis trauma

Tube cuff pressure control: DONE NOT DONE NOT APPLICABLE

Ventilation pressure < 30cmH2O DONE NOT DONE NOT APPLICABLE

Tidal volume: _____ ml Controled ventilation mode Spontaneous assisted breathing mode T Tube

Use of sedative treatment YES NO USE IN EXCES

Daily stop prescribed

Acute Distress Respiratory Syndrom Intra cranial hypertension

PART III. NURSING

Physical contention YES NO NO BUT NEEDED USE IN EXCES

Physical contention prescribed Use of a pharmacological treatment

Use of therapeutic mattress YES NO NO BUT NEEDED USE IN EXCES

ICU stay < 3 days

Use of Thrombophylaxsis YES NO NO BUT NEEDED USE IN EXCES

Active bleeding

Use of mechanical compression system Use of compression stocking PART IV. REHABILITATION

Transfert from bed to arm-chair DONE NOT DONE NOT APPLICABLE

Transfert prescribed Traumatic anatomic issue Muscular hypotonia Bedsore preventing sitting

Enteral feeding DONE NOT DONE NOT APPLICABLE

Upper digestive surgery Extubation form less than 48h Other counter indication

Oral route used DONE NOT DONE NOT APPLICABLE

Fig.1.Checklistusedduringstudy.

G.Duclosetal./AnaesthCritCarePainMed37(2018)25–33 26

(4)

definedaccordingtointernationaldefinitions[24](AppendixA). Onlyepisodeswithmicrobiologicalidentificationwerereported. Weusedourlocaldataextractedfromaregionaldatabaseinwhich all patients admitted to ourICU for more than 48hours were registered, and each of the previous nosocomial infections prospectively reported. The patients with data missing were excludedfromtheanalysis(Fig.2).

2.4. Statisticalanalysis

StatisticalanalyseswereperformedusingIBMSPSSV24.0(IBM Corp.,Armonk,NY).Theﬁrstanalysisincludedallofthepatients duringeachperiod.Continuousdataareexpressedasmeansand standard deviations or medians withinterquartiles. Qualitative data are expressed as absolute numbers and percentages. Comparisons between the continuous data periods were per-formedusingStudent’sttestorWilcoxon’stestaccordingtotheir distributionandwiththechi-squaretestforcategoricalvariables. Apropensityscore-basedapproach wasusedtomatchpatients between the periods. Propensity score was derived with the followingcovariates: age,SAPS2, reasonforadmission (trauma, medical,scheduledoremergentsurgery),sex,andimmunological status.A1:1matchingalgorithmwithoutreplacementwasused withamaximumrangeofstandarddeviationof20%(AppendixB). Outcomeswerecomparedbetween1444matchedpatients(722in each group) using conditional logistic regression. Results are presented as odds ratios (OR) with 95% conﬁdence intervals (95% CI). All comparisons were two-tailed and P<0.05 was requiredtoexcludethenullhypothesis. Toreducetheeffectof earlydeath,wecomputedthenumbersofICU-freedaysanddevice (trachealintubation,centralline,urinarycatheter)-freedays(from inclusiontoday28).

Asthiswasanobservationalretrospectivestudy,accordingto Frenchlegislation(articlesL.1121-1paragraph1andR.1121-2, PublicHealthCode),neitherinformedconsentnorapprovalfrom theethicscommitteewasneededtouseroutinedataatthetimeof theanalysis.Patients ortheir families wereinformedthatdata

couldbeusedduringtheirICUstay.Datawerestockedandtreated accordingCNILauthorisationnumber588909v1.

3. Results

Duringtheentirestudyperiod,3050patientswereadmittedto ourICU formorethan 48hours, including1816patients inthe beforegroupand1234patientsintheaftergroup.Amongthese patients,1730reportswereavailableintheregionaldatabaseon nosocomialinfections(RegionalSurveillanceNetworkof Health-care-acquiredInfections,C-CLINSud-Est).Unavailabledatawere explainedbyincompletereportsduringthestudyperiod.Ofthe 1730availablereports,19patientswereexcludedduetomissing data(incompleteorerroneousforms).Finally,761and950patients wererespectivelyincludedinthebeforeandaftergroups,(Fig.2). Wealsoanalysed10,007electronicchecklistsofthe950patientsin the after group. Central venous catheter, arterial catheter, and urinary catheter uses were reported as definitely in excess or probablyinexcess(noclearindication)in16%,7.6%,and27%ofour patients, respectively. Among the patients eligible for each guideline, head-of-bed elevation,daily interruption ofsedation, armchairtransfer,andenteralfeedingnon-compliancewerefound in1.2%,32%,28%,and21%oftheelectronicchecklists,respectively. Univariate analysis showed imbalances between the two groups(Table1,Fig.3).Thepatientsintheaftergroupwereolder and had higher severityscoresthan thosein thebefore group. Immunosuppressionwasmore frequentin theaftergroup.The casemixdifferedbetweenthetwoperiods.Theratesofpatients requiringemergencyandscheduledsurgeryincreasedintheafter period, whereas the rates of medical and trauma patients decreased(Table1).Therewerenosignificantdifferencesbetween analysed patientsandpatients withmissingreportsin termsof demographic characteristics,case mix, or death (Table 1). Data regardingthedurationofICUstayofthepatientswhowerenot analysedweremissingfromthebeforeperiod.Wedidnothave accesstospecificICUdurationdataastheinstitution’scomputer program savedonly totallengthofhospitalisation stayforthis

3050 paents

(ICU stay > 48h from 2008 to 2014)

«Before » period

1816 paents

«Aer » period

1234 paents

772paentsincluded 958 paents included

761paentsanalyzed 950 paents analyzed

10007 checklists analyzed 276 reports missing 8 paents excluded 1044 reports missing 11 paents excluded

(5)

period.Exposuretomechanicalventilationdecreasedduringthe afterperiod,whichwasduetopatientsadmittedafteremergency surgery.In contrast,exposure tourinarycatheter useincreased duringtheafterperiod.Theratesofreintubationwere20%inthe beforegroupand13%intheaftergroup(P<0.001)(Table1,Fig.3). The durations of useof each invasive device are shown in

Fig.3.ThedurationofICUstaywassigniﬁcantlyreducedin the after group (1518 days vs. 1115 days in the before group, P=0.005)(Fig.3).Thedurationofinvasivemechanicalventilation decreasedfrom7.413daysinthebeforegroupto6.010days (P=0.01)intheaftergroup(Fig.3).Thedurationsofexposureto urinary catheter and central venous catheter use were similar betweenthetwogroups(Fig.3,Table2).

Therates of VAP were21% and11% in thebefore and after groups,respectively(P<0.001).Centralvenouscatheter-related infection,bacteremia,andcatheter-relatedurinarytractinfection ratesdidnotdifferbetweenthetwogroups.Themortalityrates weresimilarinthetwogroups(23%vs.21%,respectively,P=0.6) (Table1).TheoutcomesareshownaccordingtotheyearinTable3,

Fig.4,andFig.5.

In thepropensity-basedscore, we matched722patients for eachperiod(Table3).Conditionallogisticregressionindicateda reductionintheriskofVAPassociatedwiththechecklistperiod (OR=0.38; 95% CI=0.27–0.53; P<0.001). Secondary outcome analysisshowedincreasesinICU-freedays(168.5vs.197.5; P<0.001) and ventilator-free days (21.57.8 vs. 236.6;

Table1

Comparisonofpatientscharacteristicsaccordingtotheperiodgroup.

Variables ‘‘Beforeperiod’’ ‘‘Afterperiod’’ P

Allpatients (n=1816) Analysedgroup (n=750) Allpatients (n=1234) Analysedgroup (n=961)

Age(years)(median[Q25–Q75]) 55[39–67] 54[39–67] 56[37–68] 58[42–69] 0.005

Deathoccurring(%) 360(20) 169(23) 257(21) 205(21) 0.6 SAPS2atadmission 40[29–53] 40[15–65] 41[30–52] 41[19–63] <0.001 Males(%) 1323(73) 532(71) 856(69.3) 655(68) 0.23 Immunosuppression(%) – 46(6.1) – 110(11) <0.001 Traumapatients(%) 726(40) 309(41) 432(35) 336(35) 0.009 Medicalpatients(%) 544(30) 232(31) 296(24) 235(24) 0.003 Emergentsurgerypatients(%) 327(18) 126(17) 284(23) 213(22) 0.006 Scheduledsurgerypatients(%) 236(13) 90(12) 210(17) 177(18) <0.001

DurationofICUstay(days) – 1518 118 1115 0.005

Invasivemechanicalventilation(%) – 613(82) – 742(77) 0.026

Re-intubation(%) – 146(20) – 127(13) <0.001

Centralvenouscatheterexposure(%) – 524(70) – 709(74) 0.082 Urinarycatheterexposure(%) – 655(87) – 902(94) <0.001 SAPS2:simpliﬁedacutephysiologyscore2;ICU:intensivecareunit;VAP:ventilator-associatedpneumonia;IMV:invasivemechanicalventilation;CVC:centralvenous catheter;CAUTI:catheter-associatedurinarytractinfection.

Fig.3.Comparisonofpatientsandoutcomeaccordingtotheperiod.SAPS2:simpliﬁedacutephysiologyscore;ICU:intensivecareunit;IMV:invasivemechanicalventilation; CVC:centralvenouscatheter;UC:urinarycatheter;VAP:ventilatorassociatedpneumonia;CAUTI:catheterassociatedurinarytractinfection.

(6)

P<0.001)duringthechecklistperiod.Theoccurrenceratesofdeath, CVC infection, bacteraemia, and catheter-associated urinary tract infection(CAUTI)weredifferentbetweenperiods(Table4). 4. Discussion

Inourcohort,theimplementationofanelectronicchecklistwas associated witha signiﬁcantreduction in the rateof VAP. The electronicchecklistseemedtoinﬂuenceguidelineadherencewith regardtodailysedationinterruptionandhead-of-bedelevation,

and improved the global management of invasive devices, resultinginlowernosocomialinfectionratesandshorterperiods ofinvasivedeviceexposure.

AdherencetoguidelinesisacrucialissueintheICU.Application of care bundles was shown to be associated with improved performance[4,6,25–27].However,mostguidelinesarebasedon targeted strategies with speciﬁcgoals [1,28–31]. Routinely, we observedalowrateofcompliancetoguidelinesinpatientseligible foratleastthreeclinicalrecommendations[10].Itseemsthatthe morecomplexthecarestrategy,themoreomissionsoccur.Thisis

Table2

Comparisonofoutcomedependingoftheyear.

Year Variables Year Variables

2008 n=234 VAP(%) 59(25.2) 2011 n=230 VAP(%) 48(20.9) CAUTI(%) 19(8.1) CAUTI(%) 5(2.2) Bacteraemia(%) 8(3.4) Bacteraemia(%) 19(8.2) Death(%) 50(21.3) Death(%) 54(23.5)

DurationICUstay(days) 16.117.4 DurationICUstay(days) 14.418.1 IMVexposure(days) 8.613.3 IMVexposure(days) 7.515 UCexposure(days) 8.49.7 UCexposure(days) 9.215.2 CVCexposure(days) 7.410.0 CVCexposure(days) 7.811.1 2009 n=139 VAP(%) 23(16.5) 2012 n=375 VAP(%) 60(16) CAUTI(%) 2(1.4) CAUTI(%) 9(2.4) Bacteraemia(%) 7(5) Bacteraemia(%) 18(4.8) Death(%) 29(20.9) Death(%) 91(24.2)

DurationICUstay(days) 1518.3 DurationICUstay(days) 11.312 IMVexposure(days) 6.511.2 IMVexposure(days) 6.410.2 UCexposure(days) 8.014 UCexposure(days) 8.811.6 CVCexposure(days) 5.711.9 CVCexposure(days) 7.211.4 2010 n=147 VAP(%) 25(17) 2013 n=398 VAP(%) 34(8.5) CAUTI(%) 3(1.4) CAUTI(%) 10(2.6) Bacteraemia(%) 11(7.5) Bacteraemia(%) 18(4.4) Death(%) 36(24.5) Death(%) 73(18.5)

DurationICUstay(days) 15.319.7 DurationICUstay(days) 11.317.9 IMVexposure(days) 6.311.6 IMVexposure(days) 6.111.5 UCexposure(days) 7.712.8 UCexposure(days) 7.812.1 CVCexposure(days) 6.910.6 CVCexposure(days) 6.410.7 2014 n=188 VAP(%) 11(5.2) CAUTI(%) 2(1) Bacteraemia(%) 8(4.4) Death(%) 41(22)

DurationICUstay(days) 8.238.2 IMVexposure(days) 4.56.4 UCexposure(days) 6.47.0 CVCexposure(days) 5.88.3

VAP:ventilator-associatedpneumonia;CAUTI:catheterassociatedurinarytractinfection;ICU:intensivecareunit;IMV:invasivemechanicalventilation;UC:urinary catheter;CVC:centralvenouscatheter.

Table3

Characteristicsandoutcomesofpatientsdependingofmatching.

Variables Beforematching Aftermatching

Beforegroup (n=750) Aftergroup (n=961) Beforegroup (n=722) Aftergroup (n=722) Age(years)(median[Q25–Q75]) 54[39–67] 58[42–69] 55[40–67] 58[41–69]

Deathoccurring(%) 169(23) 205(21) 167(23) 151(21)

SAPS2(atadmission) 40[15–65] 41[19–63] 40[29–54] 41[31–52]

Males(%) 532(71) 655(68) 510(71) 490(68)

Immunosuppression(%) 46(6.1) 110(11) 46(6.5) 44(6.1)

Traumapatients(%) 309(41) 336(35) 296(41) 279(38.5)

Medicalpatients(%) 232(31) 235(24) 219(30) 191(26.5)

Emergentsurgerypatients(%) 126(17) 213(22) 126(17.5) 135(18.5) Scheduledsurgerypatients(%) 90(12) 177(18) 81(11) 117(16)

ICU-freedays 19[10–23] 22[17–24] 19[10–23] 22[17–24]

Invasivemechanicalventilation(%) 613(82) 742(77) 600(83) 552(76.5) Ventilation-freedays 25[19–27] 26[21–28] 25[19–27] 26[21–28]

Re-intubation(%) 146(20) 127(13) 143(20) 95(13)

Centralvenouscatheterexposure(%) 524(70) 709(74) 514(71) 528(73)

CVC-freedays 24[19–28] 25[21–28] 24[19–28] 25[21–28]

Urinarycatheterexposure(%) 655(87) 902(94) 647(89.5) 666(92)

(7)

whyourelectronicchecklistaimedtocoverawidespectrumof clinicalareas.

Whereas the implementation of checklists indicated its efﬁciency in the operating room [11–13], its use in the ICU remains to be deﬁned. Previous single-center studies showed improved guideline compliance and decreased use of invasive

devices after checklist implementation [15,16]. Conroy et al. reported a validation process using the Delphi methodfrom a multidisciplinaryworkgroupoflocalexperts[32,33].Thismethod servedtoselectthemostrelevantitems.Inourinstitute,a multi-professional group including CRNAs, nurses, physicians, and therapistsdevelopedaglobalchecklist. Therationalewasbased

Fig.4.Incidenceofnosocomialinfectionaccordingtotheyear.VAP:ventilatorassociatedpneumonia;CAUTI:catheterassociatedurinarytractinfection.

Fig.5.Invasivedevicedurationexposureaccordingtotheyear.ICU:intensivecareunit;IMV:invasivemechanicalventilation;CVC:centralvenouscatheter;UC:urinary catheter.

Table4

Resultsofoutcomeanalysesofmatchedcohort.

Outcome ‘‘Before’’group (n=722) ‘‘After’’group (n=722) OR(95%CI) P Death(%) 167(23) 151(21) 0.93[0.78–1.12] 0.47 VAP(%) 151(21) 72(10) 0.38[0.27–0.53] <0.001 CVCinfection(%) 4(0.5) 3(0.3) 0.42[0.09–2.73] 0.42 Bacteraemia(%) 45(6.2) 32(4.4) 0.69[0.43–1.11] 0.13 CAUTI(%) 28(3.9) 20(2.8) 0.73[0.39–1.26] 0.24 Re-intubationrate(%) 143(20) 95(13) 0.76[0.61–0.95] 0.02 ICU-freedays 16(8.5) 19(7.5) 1.05[1.04–1.07] 0.001 Ventilator-freedays 21.5(7.8) 23(6.6) 1.03[1.01–1.04] 0.001 CVC-freedays 21.5(7.7) 22.3(7) 1.01[1.00–1.03] 0.03 UC-freedays 20.5(7.8) 21.3(7) 1.01[1.00–1.03] 0.06

VAP:ventilator-associatedpneumonia;IMV:invasivemechanicalventilation;ICU:intensivecareunit;CVC:centralvenouscatheter;UC:urinarycatheter;CAUTI: catheter-associatedurinarytractinfection.Continuousdataarereportedasmean(SD).

(8)

on the resultsof a large one-daysurvey [10]. In all cases, the contents of checklists should generate discussion about the processesofcarewithallmembersofthecareteam[17,34].

Inouropinion,checklistsshouldbewidelyadopted.Weused electronicsupport,whichwaseasilyaccessiblebytheentirestaff viaourintranet.ApreviousstudybyKastrupetal.showedthat adherencetoguidelinescouldbeimprovedbyusinganelectronic tool[35].Weissetal.showedthatpromptingchecklistuseduring the morning rounds was associated with better compliance

[18]. We did not assess therate of daily useof our electronic checklist. However, our system collected more than 10,000 checklists during thestudy period, generating an efﬁcient tool forassessingourpractice.

The findings regarding use of the electronic checklist were encouraging. After its implementation, the incidence of VAP statistically decreased. As compared with other studies, this incidenceseemedmoderatelyhigh[1–3].Thiscouldbeexplained by the inclusion of patients hospitalised for at least 48hours, trauma patients with head injuries, and patients developing complicationsafterchestsurgery.Exposuretoinvasivemechanical ventilationmayalsohavebeendecreasedduetotheincreasein post-surgical patientadmission. During thestudy period,there were no changes in rates of CAUTI, bloodstream infection, or catheter-relatedinfection.Astheseveritystatus ofourpatients increased, thelack ofincreasesin thesespecificinfections may reflectimprovedpractice.Moreover,CAUTI andcatheter-related infectionrates werealready low in thebefore period,and any influenceontheelectronicchecklistcouldbemaskedbyalackof studypower. We conductedseveral previous studiesregarding urinary catheter use in the field [36]. Thus, improvement was probablydifficult becauseefforts had previously beenmade to reduce the duration of urinary catheterisation. These findings highlightapossiblelinkbetweenthelocalcultureoftheICUand qualityofcare[37,38].

Analysisofthe10,007checklistsunderlinedthatourpractices canstillbeimproved.Thedailyinterruptionofsedation,theroute offeeding,orarmchairtransferwasappliedinaround70–80%of ourcases.Invasivedevices wereconsidered tobeunjustiﬁedin 7.6%to27%ofcases.Forexample,asinapreviousstudy,wefound about20%ofurinarycatheterusewasinexcess[33].Thephysician inchargecanjustifyortolerateaprocessconsideredinexcessby theCRNAduringtheirmorningrounds.Itshouldbenotedthatthe goalofthechecklist wastointroducediscussionregardingcare quality.Thisdiscrepancyisrepresentativeofadisparitybetween routinepracticeandguidelines.

Ourstudyhadseverallimitations.First,itsretrospectivedesign ledtotheexclusionof severalpatients duetoincomplete data, although demographic characteristics were the same between patientswithmissingreportsandthoseincludedintheanalyses. Thelargenumberofmissingdatacanleadtoselectionbias.We compared some available demographic data and found no significant differences.VAP reduction could be induced by the Hawthorneeffect,butthemaingoalofanelectronicchecklistwas to improve guideline adherence, directly influencing physician practice, and prompting discussion with the medical team. However, with this retrospective study, the staffs were not influenced by the observer effect. Our results reflect real-life practice. Furthermore, analysing device-free days seems more accuratethanthecrude durationofexposureasit removesthe effect of earlydeath during an ICU stay. Second,the case mix evolvedduringthestudyperiodwiththeimplementationofchest surgeryinourhospitalinJune2009andtheconsequentincreasein thenumber ofpost-surgical patients,but thepropensity score-matchedanalysisshouldhavereducedpossibleeffects.Third,the managementofpatientsprobablydifferedbetweentheonsetand completionofthestudyconsideringthelongperiodincluded.

5. Conclusion

In conclusion, the implementation of anelectronicchecklist seemsassociatedwithpositiveeffects onthepatients’outcome. Thedifferenceobservedthetwostudiedperiodsuggestthatuseof a daily electronic checklist may have reduced nosocomial ventilator-associatedpneumonia.Theseresultsrequirea prospec-tiveandmulti-centeredstudyinordertoevaluatetheimpactof checklistuseinICU.

Ethicalstatement

Wefurtherconﬁrmthatanyaspectoftheworkcoveredinthis manuscript that has involved either experimental animals or humanpatientshasbeenconductedwiththeethicalapprovalofall relevantbodiesandthatsuchapprovalsareacknowledgedwithin themanuscript.

Funding

Noﬁnancialsupport.

Disclosureofinterest

Theauthorsdeclarethattheyhavenocompetinginterest. Acknowledgments

Thanks to Dr Suzanna Zgorska-Moussa for her help in the translationofthismanuscript.

AppendixA. Criteriafordeﬁninginfections(basedon internationaldeﬁnitions)[24].

Site Criteriafordiagnosis Catheter-associated

urinarytractinfection

Positiveurineculture(>104

bacteria/ml,nomore thantwobacteria)ANDFever(body

temperature>388C)withoutanyother suspectedinfectionsite

Bacteraemia Atleastonepositivebloodculturecollected duringfeverepisodeexceptedfornegative coagulaseStaphylococcus,Bacillussp., Corynebacteriumsp.,Propionibacteriumsp., Micrococcussp.Forthosemicro-organisms,two positivebloodculturescollectedatdifferentsitesare required

Catheter-relatedinfection Positivebloodculture(collectedbyavenous puncture)

ANDoneofthefollowingcriteria:

1.LocalinfectionANDidentiﬁcationofthesame micro-organisminpusandperipheralblood 2.PositivecultureofcatheterANDidentiﬁcation ofthesamemicro-organisminperipheralblood Ventilator-associated

pneumonia

Evolvingradiologicallungopacity ANDidentiﬁcationofmicro-organismin: 1.Trachealaspiratesthroughendotrachealtube withatleast106

coloniesformingunit/mlOR 2.Bronchial-alveolarlavagewithatleast104 coloniesformingunit/mlOR

3.Protectedbrushwithatleast103_colonies formingunit/mlOR

4.Collectionofabscessorpleuralﬂuid Intra-abdominalinfection Pusorabscesscollectedduringsurgery

ORpositivebloodculturecollectedattheonsetof disease

(9)

AppendixA(Continued)

Site Criteriafordiagnosis

Tissueorskininfection Presenceofatleasttwoofthefollowingsigns: Localpain,tumefaction,localheat,sensibility, redness

ANDatleastoneofthefollowingsigns: 1.Positivecultureofmicro-organismcollectedat thesuspectedsite

2.Micro-organismidentiﬁedinbloodculture

AppendixB. Effectofmatchingonstandardizeddifference.

References

[1]MiettoC,Pinciroli R, PatelN,Berra L. Ventilator associated pneumonia: evolvingdeﬁnitions andpreventive strategies.RespirCare 2013;58:990– 1007.http://dx.doi.org/10.4187/respcare.02380.

[2]VincentJ-L,RelloJ,MarshallJ,SilvaE,AnzuetoA,MartinCD,etal.International studyoftheprevalenceandoutcomesofinfectioninintensivecareunits.JAMA 2009;302:2323–9.http://dx.doi.org/10.1001/jama.2009.1754.

[3]MorrisAC,HayAW,SwannDG,EveringhamK,McCullochC,McNultyJ,etal. Reducingventilator-associatedpneumoniainintensivecare:impactof imple-mentingacarebundle.CritCareMed2011;39:2218–24.http://dx.doi.org/ 10.1097/CCM.0b013e3182227d52.

[4]HazamyPA,HaleyVB,TserenpuntsagB,TsivitisM,GiardinaR,KnabR,etal. Effectof2013NationalHealthcareSafetyNetworkdeﬁnitionchanges on centrallinebloodstreaminfectionrates:AuditresultsfromtheNewYork StateDepartment of Health. AmJ Infect Control2015;43:280–2. http:// dx.doi.org/10.1016/j.ajic.2014.11.011.

[5]ChenowethC,SaintS.Preventingcatheter-associatedurinarytractinfections intheintensivecareunit.CritCareClin2013;29:19–32.http://dx.doi.org/ 10.1016/j.ccc.2012.10.005.

[6]NeelakantaA,SharmaS,KesaniVP,SalimM,PervaizA,AftabN,etal.Impactof changesin theNHSN catheter-associated urinarytract infection(CAUTI) surveillancecriteriaonthefrequencyandepidemiologyofCAUTIinintensive care units (ICUs). Infect Control Hosp Epidemiol 2015;36:346–9. http:// dx.doi.org/10.1017/ice.2014.67.

[7]CabanaMD,RandCS,PoweNR,WuAW,WilsonMH,AbboudPA,etal.Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA1999;282:1458–65.

[8]FlodgrenG,ConternoLO,MayhewA,OmarO,PereiraCR,ShepperdS. Inter-ventionstoimproveprofessionaladherencetoguidelinesforpreventionof device-relatedinfections. Cochrane DatabaseSyst Rev2013;3:CD006559. http://dx.doi.org/10.1002/14651858.CD006559.pub2.

[9]QuenotJ-P,MentecH,FeihlF,AnnaneD,MelotC,VignonP,etal.Bedside adherencetoclinicalpracticeguidelinesintheintensivecareunit:theTECLA study. Intensive Care Med 2008;34:1393–400. http://dx.doi.org/10.1007/ s00134-008-1059-y.

[10]LeoneM,RagonnetB,AlonsoS,AllaouchicheB,ConstantinJ-M,JaberS,etal. Variablecompliancewithclinicalpracticeguidelinesidentiﬁedina1-day auditat66Frenchadultintensivecareunits.CritCareMed2012;40:3189–95. http://dx.doi.org/10.1097/CCM.0b013e31826571f2.

[11]JammerI,AhmadT,AldecoaC,KoulentiD,Goranovic´ T,GrigorasI,etal.Point prevalenceofsurgicalchecklistuseinEurope:relationshipwithhospital mortality. Br J Anaesth 2015;114:801–7. http://dx.doi.org/10.1093/bja/ aeu460.

[12]NeilyJ,MillsPD,Young-XuY,CarneyBT,WestP,BergerDH,etal.Association betweenimplementationofamedicalteamtrainingprogramandsurgical mortality. JAMA 2010;304:1693–700. http://dx.doi.org/10.1001/ jama.2010.1506.

[13]HaynesAB,WeiserTG,BerryWR,LipsitzSR,BreizatA-HS,DellingerEP,etal.A surgicalsafety checklisttoreducemorbidity andmortalityina global population. N Engl J Med 2009;360:491–9. http://dx.doi.org/10.1056/ NEJMsa0810119.

[14]WHOGuidelinesforsafesurgery2009:safesurgerysaveslives.Geneva: WorldHealthOrganization;2009[accessedMay11,2015]http://www.ncbi. nlm.nih.gov/books/NBK143243/.

[15]TeixeiraPGR,InabaK,DuboseJ,MeloN,BassM,BelzbergH,etal.Measurable outcomesofqualityimprovementusingadailyqualityroundschecklist: two-yearprospectiveanalysisofsustainabilityinasurgicalintensivecare unit.JTraumaAcuteCareSurg2013;75:717–21.http://dx.doi.org/10.1097/ TA.0b013e327b61829.

[16]DuboseJ,TeixeiraPGR,InabaK,LamL,TalvingP,PuttyB,etal.Measurable outcomesofqualityimprovementusingadailyqualityroundschecklist: one-yearanalysisin a traumaintensive careunit withsustained ventilator-associated pneumonia reduction. J Trauma 2010;69:855–60. http:// dx.doi.org/10.1097/TA.0b013e3181c4526f.

[17]CentofantiJE,DuanEH,HoadNC,SwintonME,PerriD,WaughL,etal.Use ofadailygoalschecklistformorningICUrounds:amixed-methodsstudy. Crit Care Med 2014;42:1797–803. http://dx.doi.org/10.1097/ CCM.0000000000000331.

[18]WeissCH,MoazedF,McEvoyCA,SingerBD,SzleiferI,AmaralLAN,etal. Promptingphysicianstoaddressadailychecklistandprocessofcareand clinical outcomes: a single-site study. Am J Respir Crit Care Med 2011;184:680–6.http://dx.doi.org/10.1164/rccm.201101-0037OC. [19]FrischA,MillerT,HaagA,Martin-GillC,GuyetteFX,SuffolettoBP.Diagnostic

accuracyofarapidchecklisttoidentifydeliriuminolderpatientstransported byEMS.PrehospitalEmergCare2013;17:230–4.http://dx.doi.org/10.3109/ 10903127.2012.744785.

[20]Munoz-PriceLS,DezfulianC,WyckoffM,LenchusJD,RosalskyM,BirnbachDJ, etal.Effectivenessofstepwiseinterventions targetedtodecreasecentral catheter-associatedbloodstreaminfections.CritCareMed2012;40:1464–9. http://dx.doi.org/10.1097/CCM.0b013e39f5b1823.

[21]ChuaC,WisniewskiT,RamosA,SchleppM,FildesJJ,KuhlsDA. Multidisciplin-arytraumaintensivecareunitchecklist:impactoninfectionrates.JTrauma Nurs2010;17:163–6.http://dx.doi.org/10.1097/JTN.0b013e3181fb38a6. [22]ByrnesMC,SchuererDJE,SchallomME,SonaCS,MazuskiJE,TaylorBE,etal.

Implementation of a mandatory checklist of protocols and objectives improvescompliancewithawiderangeofevidence-basedintensivecare unitpractices.CritCareMed2009;37:2775–81.http://dx.doi.org/10.1097/ CCM.0b013e3181a96379.

[23]HewsonKM,BurrellAR.Apilotstudytotesttheuseofachecklistinatertiary intensivecareunitasamethodofensuringqualityprocessesofcare.Anaesth IntensiveCare2006;34:322–8.

[24]CalandraT,CohenJ.Internationalsepsisforumdeﬁnitionofinfectioninthe ICUconsensusconference.Theinternationalsepsisforumconsensus confer-enceondeﬁnitionsofinfectionintheintensivecareunit.CritCareMed 2005;33:1538–48.

[25]AitkenLM,BucknallT,KentB,MitchellM,BurmeisterE,KeoghSJ. Protocol-directedsedationversusnon-protocol-directedsedationtoreducedurationof mechanicalventilationinmechanicallyventilatedintensivecarepatients. CochraneDatabaseSyst Rev2015;1:CD009771. http://dx.doi.org/10.1002/ 14651858.CD009771.pub2.

[26]YoungquistP, CarrollM,Farber M,MacyD, Madrid P,Ronning J, etal. Implementingaventilatorbundleinacommunityhospital.JtCommJQual PatientSaf2007;33:219–25.

[27]ExlineMC,AliNA,ZikriN,ManginoJE,TorrenceK,VermillionB,etal.Beyond thebundle–journeyofatertiarycaremedicalintensivecareunittozero centralline-associatedbloodstreaminfections.CritCare2013;17:R41.http:// dx.doi.org/10.1186/cc12551.

[28]HootonTM,BradleySF,CardenasDD,ColganR,GeerlingsSE,RiceJC,etal. Diagnosis, prevention,and treatmentofcatheter-associated urinarytract infectioninadults:2009InternationalClinicalPracticeGuidelinesfromthe InfectiousDiseasesSocietyofAmerica.ClinInfectDis2010;50:625–63.

[29]BurryL,RoseL,McCullaghIJ,FergussonDA,FergusonND,MehtaS.Daily sedationinterruptionversusnodailysedationinterruptionforcriticallyill adultpatientsrequiringinvasivemechanicalventilation.CochraneDatabaseSyst Rev2014;7:CD009176.http://dx.doi.org/10.1002/14651858.CD009176.pub2. [30]LeoneM,Albane`seJ,GarnierF,SapinC,BarrauK,BimarM-C,etal.Riskfactors

ofnosocomial catheter-associated urinarytract infectionina polyvalent G.Duclosetal./AnaesthCritCarePainMed37(2018)25–33

(10)

intensivecareunit.IntensiveCareMed2003;29:1077–80.http://dx.doi.org/ 10.1007/s00134-003-1767-2.

[31]DellingerRP,LevyMM,RhodesA,AnnaneD, GerlachH,OpalSM,etal. Survivingsepsiscampaign:internationalguidelinesformanagementofsevere sepsis and septic shock: 2012. Crit Care Med 2013;41:580–637. http:// dx.doi.org/10.1097/CCM.0b013e383af1827.

[32]ConroyKM,ElliottD,BurrellAR.Developingcontentforaprocess-of-care checklistforuseinintensivecareunits:adual-methodapproachto esta-blishing construct validity. BMC Health Serv Res 2013;13:380. http:// dx.doi.org/10.1186/1472-6963-13-380.

[33]ConroyKM,ElliottD,BurrellAR.Validatingaprocess-of-carechecklistfor intensivecareunits.AnaesthIntensiveCare2013;41:342–8.

[34]PronovostP,BerenholtzS,DormanT,LipsettPA,SimmondsT,HaradenC. ImprovingcommunicationintheICUusingdailygoals.JCritCare2003;18:71–5. http://dx.doi.org/10.1053/jcrc.2003.50008.

[35]KastrupM,NoltingMJ,AhlbornR,BraunJ-P,GrubitzschH,WerneckeK-D, etal.Anelectronictoolforvisualfeedbacktomonitortheadherenceto qualityindicators inintensive caremedicine.J IntMed Res2011;39: 2187–200.

[36]LeoneM,PerrinA-S,GranierI, VisintiniP,BlascoV,AntoniniF, etal.A randomized trial of catheterchange andshort course of antibiotics for asymptomaticbacteriuriaincatheterizedICUpatients.IntensiveCareMed 2007;33:726–9.http://dx.doi.org/10.1007/s00134-007-0534-1.

[37]GershengornHB,KocherR,FactorP.Managementstrategiestoeffectchangein intensivecareunits:lessonsfromtheworldofbusiness.PartIII.Effectively effectingandsustainingchange.AnnAmThoracSoc2014;11:454–7.http:// dx.doi.org/10.1513/AnnalsATS.201311-393AS.

[38]NguyenY-L,WunschH,AngusDC.Criticalcare:theimpactoforganizationand management onoutcomes. Curr Opin CritCare 2010;16:487–92. http:// dx.doi.org/10.1097/MCC.0b013e391802833.