Physician Knowled ge, Perception s, and Attitudes towards Electronic Medical Record Systems in Newfoundland and Labrador by

PhysicianKnowledge,Perceptions,and Attitudes towards

Electronic MedicalRecordSystems in Newfoundland and Labrador

by

©Sara-Lynn HeathKing

Athesis submitted to the Schoolof Graduate Studies

in partialfulfillmentof therequirementsforthedegreeof

Master of Science Facultyof Medicine Division of Community Health and Humanities

Memorial Univers ityofNewfoundl and

April,2012

St. John ' s Newfoundl and

Abstract

Objectives: To illustrate the knowledge,perceptions,and attitudes of Newfoundland and Labrador (NL)physicianstowards electronic medical record (EMR) systems and their use in the practice of health care.

Methods: A self-administeredmail-out surveywas usedto collect information on physician characteristics,computer experience, perceptions about EMR systems, and opinions on acceptable costs of these systems.

Results:Forty percent of eligiblephysiciansresponded .Physicians agreed that an EMR system should be implementedand thatusing an EMR would improve the access to and the efficiency of health care.

Conclusions:The major concernregardingthe useand implementation of an EMR system iscost-related .Examiningpotential subsidy modelsforimplementation and use ofEMR systems for NL physicians shouldbeundertaken.

Acknowledgements

I would like to offer mymostheartfeltgratitudeto mysupervisors and committee member,Dr. Maria Mathews,Dr.JamesValcour,and Dr. Shabnam Asgharifor their encouragement and supportthroughthe completion of mythesis.Without their efforts, support,and advicethis would nothavebeen possible.

I wouldalsolike to acknowledgethe Newfoundland and Labrador Medical Association(NLMA)for their support of this project,the College ofPhysicians and Surgeonsof Newfoundlandand Labradorforsupplying the addresses forthemail-out, and theHealthResearch Unit for their assistance.In addition,Iwould like to recognize theguidanceprovided bymy expert panel,Drs.Doreen Neville,VeereshGadag,and Gerard Farrell.Aspec ial thank you to all the physicianswho participated in thestudy.

Finally,Iwould liketothankmyfamil y and friends for their love, support, and understandingduringthistime,especially mymother,Patricia Heath and myhusband, Robert King.Ilove youallofyousomuch - this would neverhavebeen possiblewithout your patienceandsupport. Iwouldalsoliketo say a special thankyou tomyfather,Stan Heath-Iwishyoucouldhave seenthe completionof thiswork,Iknow how proud you wouldhave been of me.

iii

Table of Contents

Abstract ii

Acknowledgement s... . iii

Table of Contents iv

Listof Tables vii

List of Figures viii

Listof Abbreviations ix

List ofAppendice s x

Chapter I: Introduction... .. I 1.1 ResearchProbl em... .. I 1.2 Research Objectives .

1.3 Rationale ..

... 2

. 3

Chapter 2: Background andLiterature Review..

2.1 Background .

....4 ..4

2.2 Definiti ons ... 5

2.3 Technol ogy Acceptance ModeL... .. 7

2.4 Perceived Usefulness ...9

2.5 Perceived Ease of Use 10

2.6 Attitude Toward Usin g . ...12

2.7 External Variables 13

2.8 Previous Surveys 14

Chapter 3: Methods 17

3.1 StudyDesign 17

3.2 StudyPopulation 17

3.2.1 Eligibility 17

3.2.2 Representativeness of the Sample 18

3.3 SurveyDevelopment 18

3.3.1 Pretesting 20

3.4 Data Collection 20

3.5 Data Management ...21

3.5. 1 Data Quality 22

3.6 Data Analysis 23

3.7 EthicalConsiderations 24

Chapter 4: Results 25

4.1 Survey Response 25

4.2 Representativeness .

4.3 Respondent Characteristics ..

. 26

...26

4.4 Respondent CurrentComputerUsage 28

4.5 EMR Knowledge andUse 31

4.6 RespondentPerceptions ofEMRSystems 36

4.7 RespondentBeliefsinEMRsEffect onthe PracticeofHealth Care 37

4.8 Respondent PerceptionsRegardingthe CostsofEMR System s 38

4.9 OtherComm ents 41

Chapter5: Discussion andConclusio ns 42

5.1 Representat iveness ofthe Sample 42

5.2 CurrentUseand Trainin g withComputers 43

5.3 PhysicianKnowledge andUseofEMRsystems 43

5.3.1 PhysicianPercepti ons of EMR Systems 44

5.3.2 PhysicianBeliefsRegardin gEMRs Effect on thePractice of HealthCare 45

5.4 PhysiciansPerceptions ofEMR Costs 46

5.5 StudyStrengths 48

5.6 StudyLimitations 48

5.7 Conclusions 49

5.8 Recomm endations 49

Referenc es 51

Appendices 61

List of Tables

Table 1:Comparisonof physicians andsample frame to assessrepresentativeness ofthe

study sample 26

Table2:Characteristics of physiciansin the study 27

Table3:Computercharacteristics ofphysician sin the study 29 Table 4:Frequency ofcomputer usetoperformtasksin practiceamongphysicians in NL ...30

Table5:EMR perceived knowledge and use forphysiciansinNL.. 33 Table6: PhysicianopinionsregardingEMRsystems in general 34 Table 7:Physician opinionsregardingthe movementtowardsusing EMRsystems in

clinicalpractice 35

Table8:Physician opinionswithgeneralstatementsaboutEMRsystems 36 Table9:Physicianopinionswith the beneficialeffectof EMRsystems onthe practice of

health care 37

Table10:Physicianopinionsregarding set upcosts ofEMRsystems 39 Table11:Physicianopinionsregarding ongo ing costs ofEMRsystems 40 Table12:Physicianopinionson this study and questionnaire 41 Table13:Survey question sources,modifications, andstudy objectives 66 Table 14: Numberof respond entsfor generalquestions aboutEMR knowledge 74 Table 15:Number of respondentsforthe EMR perceptionquestions 75 Table16:Number of respondentsforthe EMRcostingquestions 75

Table17:Surveyvariablesand coding scheme 76

vii

List of Figures

Figure I:Techn ology Acceptance Model 9

Figure 2:Study sampleand response rate 25

EHR

EMR

PHR

Infoway

NL

NLCHI

NLMA

NPS

OECD

TAM

Listof Abbreviations ElectronicHealth Record

Electronic Medical Record

Patient Health Record

Canad ianHealth Infoway

Newfoundlandand Labrador

Newfoundlandand LabradorCentreforHealthInform ation

Newfoundlandand Labrador Medica lAssociation

National Physician Survey

Organisation for Economic Co-operationand Development

TechnologyAcceptance Model

ix

Responses and MissingDataforSpecificQuestions 74 SurveyVariables andCodingScheme... .. 76 Data Coding/Cleaningfor Multiple Response s 82

List of Appendices

Survey 62

SurveyCreation . 66

Information Letters 71

Appendix A:

Appendix B:

Appendix C:

Appendix D:

AppendixE:

AppendixF:

AppendixG:

Appendix H:

NLMALetter ofSupport ..

HumanInvestigation Committee ApprovalLetter...

.. 73

.. 84

1.1ResearchProblem

With the creationof Canada Health Infoway (Infoway), in2001, Canada'shealth care system hasbeenmovingtowardsthe creationof anelectronic healthrecord(EHR) (Boonstra & Broekhu is, 2010). Infowayisfundedbythefederalgovernmentand jointly investswith the provinces and territoriestoincreasethedevelopmentand adoption of EHR projectsinCanada.Itisbelievedthat EHRscan improve on the current paper-based health care system by enablinganeasytoreadand accessiblehealthrecord thatintegrates allaspects of an individu al' s care(Infoway, 2005,Shachak,Hadas-Dayagi,Ziv,&Reis, 2008) .An important aspect of anEHR istheportionofthehealth recordassociatedwith physic ian care;thisiscalledthe electronic medical record (EMR).

AnEMR system willenable physiciansto send and receiveinformationfrom the EHRin relation to theirpatients'healthinformation , and thusbeable to offer amore completeview ofa patient' shealth condition.WhileInfoway states thatalmost 50%of Canadian swill have accessto coreelements of anEHRsystem in mid-2011(Infoway, 200Ia),EMRadoption hasbeen slow and,when implemented,generallyunder-used.As of March20 II,Infowayhas reportedan estimated thatacross Canada 49%ofEHR system elementsare available, these elements includebutare not limited to EMR systems, diagnosticimaging,druginformation systems,and registries(Infoway,2011).

The province of Newfoundl and and Labrador (NL),through the Newfoundlandand LabradorCentre for HealthInformati on(NLCHI) ismovingforwardwith the goalof implementin g anEHR for theprovince.NLCHIisattemptingthisimplementation throughtheuse andcreationofmanyinformationtechnologies,includingaunique personalidentifier/clientregistry,diagnosticimaging/picturearchivingand

communications systems,and projects in tele-health,pharmacy, and primaryhealthcare information systems.Anevaluation ofapilot EMRsystem implementationin St. John' s, NL (Neville, Caison,& Farrell,2007) wascompl etedbytheeHealth ResearchUnit, Faculty of Medicine,Memorial Universityof Newfoundland.

The literature suggests that clinicians andstaff aremorelikelytoaccept andcontinue touse anEMRsystemif,prior to implementation,expectationsare clear and realistic and thatphysiciansand otherstaffhave thenecessary skills tousethe system.

1.2Resear chObjectives

Thepurpose of thisdescriptive study isto examine physicians'attitudestowards EMRsystems, in health care.Using a self-administered mail-out survey, theresearch obje ctivesare:

I.Todescribephysicians' current use of and theirtrainingin theuse ofinformation technologies.

2.Todescribe phys icians'knowledgeofEMRsystemsand theirperceptions of the effectofthese systems on the practice ofhealth care.

3.To describephysicians'perceptionsof acceptablecosts forthe implementation and maintenance of EMRsystems.

1.3 Rationale

The movetowards the use of EMRsystemsrepresentsa considerableinves tmentof resources,notonlyfunding for the implementationof these systems butalso thetimefor physicians andother clinical stafftoleamtouse and incorpora teEMRsystems intotheir practiceof health care(ShachaketaI.,2008). For 2011,Infowa y (2001b)hasplanned an investmentof$380 million to assist withtheimplementationofE MRsystems for Canadian physicians. Given the investment,it is important to examine thereasonsforthe slow uptakeofEMRsystems across Canadaand howthesebarriersmaybe overcome.

The findingsof this studywill help with theplannedimplementationof an EMR system in theprovince ofNL. The study willdescribe the current level ofcomputerskills thatphysicians possess and their expectationsand percep tionsofEMRsystems in their practiceofheaIth care.Studyresults willassist the Newfoundland and LabradorMedical Association (NLMA)and the provincialgovernment todesignan appropriate implementationprocess for the province anddevelopstrategiesto address physician concerns.Itwillalsoprovide a baselinelevel of physicians'computer technology skills and knowledge for futureevaluationsof theeffectofEMRsystemimplementations in the province, aswellasproviding local evidenceto beusedin nation-wide comparisonsof EMR implementation.

Chapter 2: Backgroundand LiteratureReview 2.1Background

The movetowardstheimplementationand useofEMR systemsis growing in the world.This growth isbased on themanypotentialbenefitsof these systems ,whichhave suggested an improvementin thequality of patientcareand safetyand addressingthe barriers around implementation and useof thesesystems(Denomme, Terry,Brown, Thind,&Stewart, 2011;Garrido,Jamieson,Zhow,Wiesenthal ,&Liang,2005;Hillestad eta!.,2005; Loomis,Ries,Saywell,&Thakker,2002;Simon,Rundall,&Shortell,2005).

Uptakeofthese systemsis thusvery important.

InCanada,in 2007,approximately9.8%of physiciansrelied solelyon anEMR system intheir practice (NationalPhysicianSurvey [NPS], 2007),thisnumber roseto 16.1%in 2010(NPS,2010).In compari son,Simon et a!. (2005) stated that inthe United States 20%-25% of physicianorganizationshaveadopted EMRs,whileJohnston, Leung, Fung Kam Wong,and Ho(2002,p.42)cited rates of"90 %in the UK,84%in New Zealand,70%in Denmark,60%inSweden and 40%in the Netherl ands"and 30%of individualphysicianpracti sesin HongKong.Ludwick,Manca,andDoucette (2010) found that EMR adoption inCanada(26%) and theUnited States(24-28%)waslow for general practice physicians,and forall physicianswhen comparedto other Organisation for Econo micCo-operationand Development(OECD)countries(80%to 99%).While Canada ismakin gprogress on the implementationanduse of EMRsystems,itisnot at thelevelneededto see the benefits thatthese systems can bring topatienthealth.

A clear message from the literature is that in order to improve EMR usagethereneeds to be greater understanding of the reasonsfor physicians'uptake (orlackthereof)to improve the usage ofEMRsystems (Aydin&Forsyth,1997;Boonstra&Broekhu is, 2010;Gadd &Penrod,2001;Joos,Chen,Jirjis,& Johnson,2006;Rose, Schnipper,Park, Poon,Li,& Middleton,2005).Thisimproved understanding will help ensuresuccessf ul adoptionof EMRsbyphysician and other clinicians.Loomiset al. (2002,p.640) stated thattheimportant differences between EMR usersand non-usersare:"(1)lessperceived needfor EMRs;(2) greater concernsabout EMRdata entry;(3)lessconfidencein the securityandconfidentialityofEMRs;and (4) more concernsabout thecost for installation andongoinguse ofEMRs."

2.2 Definitions

Theuse of informationtechnology systems in health care hasintroduced anew set of terminology.Thesenew health informationterminologies include: electronichealth record (EHR),personalhealthrecord (PHR) and electronicmedical record (EMR),allof which tend tobe usedinterch angeablyalthoughthey havedifferencesin theirdefinitions.

Hodge (2011)explainsthat alot of the confusion betweenthe threeterms is duetothe twoideas,thecompletenessof the informationand the custodianofthehealth inform ation.Heprovidesthefollowing definitions:

ElectronicMedical Record-a partialhealthrecord underthe custo dianshipof a healthcare provider(s) that holds a portion of the relevanthealth information aboutaperson overtheirlifetime.This is often describedas aprovider-centri c or

health organization-centric health recordof aperson...We also have software productscalled Electronic Medical Records (EMRs).TheseEMR product sare primarilyused by physiciansintheir office or in anout-patient clinic. The term EMR hastraditionally notbeen used to describe softwareproductsmarketed at other points ofhealthcare serviceinCanada.(e.g. hospital,continuing care,public health ,mentalhealth and so on).

Electro nicHealthRecord -a complete healthrecordunderthe custodian ship of a healthcareprovider(s) that holds all relevanthealthinformationaboutaperson over their lifetime. Thisis often described as aperson-centrichealth record, which canusedbymany approved health care providers or health care organizations.

Persona lHealthRecord-a completeorpartialhealthrecordunderthe custodianshipofa person(s) (e.g.a patient or family member)thatholdsallora portion ofthe relevanthealth informationaboutthat person over their lifetime. This is alsoa person-centric health record.(Hodge,20 11, para.8)

Thissurvey used the definition createdby The Institute of Medicine'andusedby Simonetal(2005):

...electronically storedinforma tionaboutan individual' slifetimehealth status and healthcare.Itreplaces the papermedicalrecord astheprimaryrecordof care, meeting all clinical,legal, andadministrativerequirements. An [EMR] system

ITheInstitute of Medicine is anAmerican not-for-profit,government-independent organizationwhose purposeistoprovide adviceon issues relatedto biomedicalscience,medicine,and health.

provides remindersandalerts,linkages with knowledge sources for decision support,and dataforoutcomes research and improvedmanagement of health care delivery. (p.631)

2.3 Techn ology Accepta nceModel

The technology acceptance model (TAM),developedin the1980sbyDavis (Davis, 1986, ascitedin Davis,Bagozzi,&Warshaw, 1989),isusedfor researchintothe use and acceptance ofinformation systems(Chismar&Wiley-Patton , 2003; Hu,Chau, Liu Sheng,&Tam,1999; Seeman&Gibson,2009)with the aimof describingthe factors associated with information technologiesacceptanceand intentionsto usebyindividuals (Holden & Karsh, 2009;Malhotra & Galletta, 1999). Thegoalof TAM is"to providean explanation ofthedeterminants of computeracceptance thatis general,capable of explaining userbehaviour across a broad rangeof end-user computingtechn ologies and userpopul ations, whileat the same time beingbothparsimoniousand theoretically justified" (Dav iset aI.,1989,p.985) .

TAM waschosenfor thisstudybecauseof itswidespreaduse in theliterature.TAM hasbecome a well-accepted modelforassessingtheimplementation and use of informati ontechnology in thehealth carefield (Holden&Karsh, 2009; Yarbrough&

Smith,2007). Holden and Karsh (2009) summarizes fifteenpreviouspapers ,which have used TAM toassessahealth caretechnology;ofthese,tenfocusedon physiciansasall or partof thestudy populat ion.Yarbrough andSmith(2007)alsosummarizes eighteen studies on physiciantechnology acceptancefor a varietyoftechnolo gies; of these,half

usetheTAM as the modelapplied and have study populationsconsistingofphysicians and residents.Thismodel allows for the complexities ofhealth care organizatio nsand prov idesa startingpoint to address the problems around uptake of information techn ologiesin health care.

Other models that could have been usedto examine EMRuse include,butare not limitedto,thetheory of reasoned action model,the theory of planned behaviour ,TAM 2, and theuniversaltheor y of acceptanceand use oftechn ology.Thesemodelshavebeen comparedand discussedinrelationto TAMandeachother in theliterature (Chuttur, 2009; Hold en&Karsh , 2009;I-Iu et aI., 1999 ; Venkatesh, Morris,Davis,&Davis,2003;

Yarbrough & Smith,2007). Additionaly,TAM was chose n becauseit is a general but inform ationtechn ology specific modelwhich has been usedin manydifferent popul ations, and thephysician populationinspecific. Ithas also been show n tobe a good predic torofphysicianintention to use and accept technology,includingEMRs.

TAMisillustrated in Figure 1. The aspects of the model thatwere usedtoorgani ze theliterature (shownin the dotted box) are perceived usefulness,perceive dease ofuse, attitudetowardusing and external variables.External variablesinfluence users'ideas about using a system (perceived usefulnessand ease of use).Perceived ease ofuse can influencepercept ion ofusefulness.Both sets of percep tio ns influ ence attitudes toward use,whichisbelievedto influe ncebehavioralintentio n touse, which in turn influence actualuse;perceivedusefuln ess canalsoinfluence behaviouralintentionto use (Burton- Jones& Hubona, 2003).

- - - -- - - -- I

I i ii i i ii i i

!

I~.~_.~-.~_._.~_-~ ~ ~ _ ~__~~._._ ~

Figure 1: Technology Acceptance Model

Reprint ed by permission,Davis, FD., Bagozzi,RP.,&Warshaw,PRoUser Acceptance ofComp uter Techn ology: AComp arisonofTwoTheoretical Models,ManagementScience,volume35,number8, August,1989.Copyright 1989,theInstitute/ orOperationsResearchand theManagement Sciences, 7240 ParkwayDrive,Suite300,Hanover,Maryland21076USA.

• Thedolledlineillustratesthe partsofthemodelusedinthis study.

2.4 Perceived Usefulness

Perceivedusefulness isdefinedas "the degreeto whichapersonbelievesthat using a particul ar systemwouldenhancehis orher jobperformance"(Daviset aI.,1989,p.985).

This includeshowphysiciansperceive EMRsystemsin general,theirexperienceswith EMRs,andwhat physiciansbelievethebenefitsof anEMR wouldbe.

Theperceived benefitsof theuse of EMRsystemshavebeenwelldocum ented and mainly stemfrom the waycomput er systemsstore andstructurepatientchartinformation.

EMRs have beenshown toproduce animprove ment in the quality and continuity of patientcareby allowingforcompl ete and legibledocumentation,reducing medicalerrors and repetition of tests;improvingaccesstopatientrecords,evidence-basedliterature,and communication betweenphysicians.Studiesexamining perceivedusefulness employed a varietyof methodsincludingsurveys (LoomisetaI.,2002; Simonet aI.,2005),qualitative

methods (i.e. focus groupsand interviews) (Rose et aI.,2005),evaluation of administrativedatabases (ClaytonetaI., 2005;GarridoetaI., 2005),and literature reviews (OECD, 2010;Retchin,1999).

2.5 Perceived Ease ofUse

Perceivedease of useisdefinedbyDavis etal. (1989,p.985) as"the degree to which a person believesthat using aparticular system would befree from effort."Ease of use includessystemdesign and usabil ity,the abilityof physicianstousethe system and incorpora teitintotheir workflow, and physicians'perception sof thesecurityofthe EMR systemand their control over the system and the data.

System designand usability are keyconcerns for potential users of EMR systems, including physician s.Themain system concernsincludethe appropriateness and user- friendliness of softwaredesign (Brown, 2005; Clayton et aI., 2005; Hodge,2002;

Johnston et aI.,2002;Joos etaI.,2006; Loomis etaI.,2002;Retchin,1999;Rose et aI., 2005;Santiage,Li,Gagliano, Judge, Hamann,&Middleton, 2006; Teach & Shortliffe, 1981;Youn g,1984) aswellasthelearningcurveand thetrainingrequired to usethe system(Gamm, Barsuki ewiczm,Dansky,&Vasey,1998;Hodge, 2002;Johnstonet aI., 2002; Gadd & Penrod,200 I;Joos etaI., 2006;Kaelber,Greco,& Cebul,2005;Ludwick et aI.,2010; Santiageet aI., 2006; Teach&Shortliffe, 1981;Terry,Giles,Brown,Thind,

&Stewart,2009;Young,1984).Other system concerns revolvearoundsecurity and

confide ntialityof informationheldin theseEMRsystems (Boonstra & Broekhuis,2010;

Claytonet aI.,2005;Johnston et aI., 2002; Joos et aI., 2006; Loomis et aI., 2002;Retchin,

1999; Terry et aI.,2009) and thelack of technica lstandards for EMRsystems(Claytonet aI.,2005; Ludwick etaI.,2010;DECO,2010;Retchin,1999;Young,1984).

Design concern linksinto physicians 'concernsoverinteraction with theirpatients, beingcomfortableusing the system during encountersand theirown levelofkeyboarding and computer skills(Boonstraand Broekhuis,2010;Cork,Detmer,& Friedman,1998;

Johnston et al.,2002;Joos etaI.,2006; LoomisetaI.,2002;Simonet aI., 2005; Terry et al.,200 9).Designissueslead intoconcernsoverthe EMRsystems' potenti alfor change to physicians'currentworkprocesses(Boonstra & Broekhui s,20 10; Brown,2005;Gadd

&Penrod ,2001;Greiver,Barnsley,Glazier, Moineddin ,&Harvey,2011;Johnston et aI., 2002;Ludwick etaI.,2010; Rose et aI.,2005;TerryetaI.,2009; Young,1984);and also potenti alchanges inthe interaction betweenphysicians and patient swith theintroduction ofacomputer during the encounter(Gadd&Penrod,2001;JohnstonetaI.,2002;

Ludw icketaI.,2010;Simonet al., 2005).Sittig,Fuperman,and Fiskio(1999) state that it isimportantfor systemdesignerstotakeinto account howphysicians will beusingthe system.Brown(2005)emphasizes thisby suggesting that physician reluctancetousethe system is oftena resultof the systembeing overlydesigned,thus makingthe system less intuitiveto usethan a paperrecord.Additionally,physicianconcernabout howpaper recordswill be convertedto an electronicformatand hownoteswill be entered (Payne, tenBroek, Fletcher,&Labu guen,2011) havebeenidentified asa potential barriers to EMR use (Chisolm,Purnell,Cohen,&McAlearney, 2010;Clayton et al.,2005;Loomis etaI.,2002).

II

AsRose et al.(2005)andAydinand Forsyth (1997) haveboth stated,for EMR systems tobeof usetophysicians, physicia ns mustbe comfortable with the system and confident in their ability touse the systemto perform their jobs. Healthtechnologymust beuserfriendly and meet standardsand technological requirements for systems to be adoptedand used.Physician acceptanceisnot only closelylinkedtosystemdesign and usabilit y,butalso tothephysician'scomfortlevelin usingcomputer s.Laerum , Ellingsen,and Faxvagg, 200 1(2001)stated thatcomputerliteracyand changesto work- flow were possiblereasonsforthelack ofEMR usage.

2.6 Att itude Toward Using

Attitudetowardusing, asDavis et al. (1989) definesit,isaffectedbybothperceived usefulnessand perceived ease of use.Inadditio n,attitude towardusingincorporatesthe inputof theuserintothe selectionor creationofthe systemandthe feeling of

"voluntariness" theuser experiences in choosing,implementing, and using anEMR (ClaytonetaI.,2005;Joos et aI., 2006; Loomis et aI., 2002; Young 1984). Attitude towardusingis alsoaffectedby concernsof ownershipand securityandany"Big Brother"control ofthe system (Loomis et aI.,2002; SimonetaI.,2005)and beliefin the evidenceof thebenefits of EMRuse (LoomisetaI.,2002;Teach&Shortliffe,1981;

Yarbrough&Smith,2007).

Physician resistancetowardusing anEMR is acommonl y expressed barrier toEMR systemimplementation (Boonstra&Broekhui s, 2010;Gadd & Penrod ,2001;Johnston et aI.,2002;Joos et aI.,2006; SimonetaI.,2005;Young, 1984).Physician attitudes playa

vital role for acceptanceof computersystems; physicianwork process and their attitudes towardsinform ationtechnologies areshown tobeimportantconstraints (Young, 1984).

Aydinand Forsyth (1997)andCork etal.(1998) both state thatmuch researchhas examinedphysicians' attitudes towards computersystems. Teachand Shortliffe(198 1, p.542)expressedit best whensaying,"despite thepromise of medicalcomputing innovations ,manyhealthcareprofessionalshave expressedscepticismabout the roleof the computeras an aid toclinicians".Tocounterthisresistance ,research hasshown that having strong leadership ora 'champion' for theimplementationof thesystem can positively affect the adoptionofE MR(Hing,Curt, & Woodwell,2007; Ludwick et al., 2010;Terry et aI.,2009).

2.7ExternalVariables

External variables are those thatinfluence users'perceptions(seesections 2.4 and 2.5)about the system, including things suchas practicesize,system costs,and system and userissues, aspreviously discussedinperceived usefulness andeaseofuse sections.

Practice sizehas beenshownto be a factor inthe implemen tationandadoptionof EMRsystems(HingetaI.,2007;Miller, HiIllman,&Given,2004;Retch in,1999;Simon et al., 2005).Therehasbeen alack of research intotheadoptionof EMR systemsin small clinics,with mostresearchbeing conducted on hospitalimplementationsor clinics associated with hospitals (Keshavjee,Troya n, Holbro ok,&VanderMolen,2001).Simon et al. (2005) found that larger groupswould bemorelikelythansmall clinics to adopt EMR systems.The reasonsforthis could bethe associated costsof EMR systemsand the

13

perceiveddifficulties of adoptionof EMRsystemsinto current practice workflow and workloads.

Systemcostshave been identified as apotentialbarriertothe adoptionand useof EMRsystems,with the cost of systemsincludingsoftware,hardware, supportand maintenance,train ing of physiciansandstaff,and initialproductivit yloss (BoonstraetaI., 2010;Johnston etaI.,2002; Loomiset aI.,2002;LudwicketaI.,20I0;Retchin ,1999;

Simon et aI.,2005;Terry etaI.,2009;WangetaI.,2003) . Theliterature estimates that the costofEMR systems (inUS $)rangefrom $1,600 to $10,000 perphysicianfor software costsalone,basedonUS systemsand studies(Brown,2005; Wangct aI.,2003).

Littlejohn s,Wyatt,&Garvican (2003) estimated thecost ofanEMR to be approximately

$50 million forlargehospital. Greiver etal.(2011)and Terry,Chevendra, Thind, Stewart,Marshall,&Cejic(2010)alsohavefound an increasein uptakebased on reimbursement orsubsidies.There isalsoevidencethat EMRsystemswilleventually produceeconomicbenefitsthat aftera fewyears of use wouldoffsetthe initial costof setup(Brotzma n,Gusc,Fay,Schellhase,&Marbella,2009;Wanget aI.,2003).

2.8Prev iousSurveys

A number of researchershave previouslysurveye d physiciansto examinefacilitators and barriersto EMR use.Loomis etal. (2002) completeda cross-section almail out survey offamily physicianstodetermine any differencesinattitudes, beliefs,and demographiccharacteristicsbetween EMR users and non-users.Loomiset al. (2002) foundthat there was adifference inattitudesand perceptionsof usersand non-usersof

EMRsystems.He found thatnon-usersperceivedlessneed forEMRs, hadmore concerns aboutdata entry, had less confidence in the systems securityandwere more concerne daboutassociated costs than users.

Simonet al. (2005) collected data on characteristics of medicalgroups,including yearsofpractice, size of practice,information technology use, and external incentives . These data were collectedthrough structuredinterviews.Simon et al.(2005)foundthat knowin g organiza tionalcharacteristicscouldhelp withtheadoptionof EMRsystems, for example thatfinancial incentiveswould benefit all medical groups, butespeciallysmaller groups.

To evaluate physicians' attitudes towards computer-based clinicaldecision aidsCork etal. (1998)developed a questionnaire. This survey instrumentwas designedtoinclude measures of computeruse not included in most priorstudies,and to specificallyaddress the rolesandactivities of physicians. Cork et al. (1998) found that computer use and knowledgewas relatedto respondents'training and self-reported skilllevel.

Laerum and colleagues (Laerum et aI.,2001; Laerum & Faxvagg, 2004) created and used questionnairesto investigate and compare the use of EMR systemsin a hospital settinganda task-orientedevaluation.Generaltasks related to physician work were assessedalongwithcomputer literacy and user satisfaction, both surveys were validated.

Laerum etal.(200 I) foundthat in generalphysicia ns use EMR systemsfor lesstasksthan they could beusedfor. Laeurm and Faxvagg (2004) found thatthetestedquestionnaire providesreliable resultswithrespect to clinicalwork andEMRsystems.

Other studies have focused on specific sections onthetopics thatthe studies mentioned above have covered. Krall (1995)and Miller et al.(2004) havelinked

15

physici ans ' currentcomputer usage to theiracceptance and use ofEMRs .Attitudesand perceived effects havebeen linked toacceptanceand use of EMRsystems in previous studies(AI Farsi & West ,2006;Ford,Menachemi,& Phillips,2006; Gadd&Penrod, 200 1;Littlej ohn s et aI.,2003;Mush am,Ornstein,&Jenkins, 1995).Specific demograph ic variab les,suchasage andpractice size, have been linked to acceptanceand usage ofE MRsystemsin practice("Physicianuse ofEMRs",2005; Ford etaI.,2006;

Miller,West,Brown,Sim,& Ganchoff,2005;Simo net aI.,2005).

The knowl edgebase ofinformationtechnology use andacce ptancesurrounding healthprofession als, spec ifica llyphysicians, is evergrowi ngand allowsfor a more thorou gh andcomprehe nsiveimplementation plan.Todate,thisinformationhasnot been collect edfromphysicians inNL despit etheplansto adoptEMRs into thehealthcare systems.

Chapte r3: Methods 3.1Study Design

Thisdescripti ve studyexplores physicians' attitudesand perception s ofE MR systems . Aself-administeredsurvey(AppendixA) wasusedtocollectinform ation about physicians 'currentcomputerskillsand training and theirperception s and knowledgeof EMRsystems and their effecton thepractice ofhealthcare.Physicianswere surveyed between September 2007and December 2007.

3.2Study Population

The study populationincluded allgeneral practiti oners/familyphysicians and specialists regi steredwith the CollegeofPhysicians and SurgeonsofNL as of July 31, 2007,includingadministrati ve andteaching physician sfora totalof 1083physician s.

3.2.1 Eligibility

Tobe eligible toparticipate inthestudy,physiciansmustberegistered with the RoyalCollege of PhysiciansandSurgeonsof Newfoundlandand Labrador asof July 31, 2007 and notbe aparticipant inthePilot EMR Implementation Evaluation Study being conducted by the eHealthResearch Unit(phys icians at theNewfoundlandDrive Family Practice, Family Practice Unitat theHealthSciencesCentre,includingthe Shea Heights site).

Alsoexcludedfrom the surveywereresident sand trainees,and thosewho returned surveys which did not havethedemographi cinform ation or themajority of the EMR

17

knowledgeand usagequestions completed.Additionally, physicianswithan address outsideofNL,those whowere identifiedas nolongerpracticingin theprovince,and thoseidentified as no longer working at thelisted address werealsoexcluded.

3.2.2 Representativenessof theSample

To assesstherepresentativeness ofthe sample, X²tests were usedto compare gender andspeciality, forthe samplepopulation tothetotalpopul ationofphysiciansin NL.

Physicianpopulation characterist icswereavailable fromthe2007 National Physician Survey:NL Demographics (NPS,2007).Thesedatawereused as it is representativeof the study population at thetime thisstudywas conducted.

3.3 Survey Development

The most commonmethod found in theliterature forstudying physician acceptance and use of EMRsystems was through the use of surveys(CorketaI., 1998;LaerumetaI., 2001; Loomis etaI.,2002;Simon et aI.,2005).Thequestionnaire usedin this studywas developedby selectingquestions from othervalidatedsurveyinstruments(Cork etaI., 1998;Laerum et aI.,2001;Laerum & Faxvagg, 2004;LoomisetaI.,2002) (see section 2.8). In somecases,questions were modifiedand newquestionswere developedto addressthe studyobjectives. Appendix B describesthe specific question susedto developthesurveyusedin this study.

Thequestionn aire was divided intothree sections:current usage ofcomputers, physicianpercepti ons andattitudestowardsEMRsystemsand theireffectson healthcare, and demographics. These sections andtheirrelated surveyquestions are describedbelow.

Thefirst sectionofthe survey containedquestionsaround physicians'current computerusage. Questionsincluded theownershipand use of computersat home and work ,theuse of computersfor specifictasksintheir practice,previous computer training, and self-reportedrankin g of computerskills. Thesequestions were based on questions from studiesby Corketal.(1998), Laerumet al.(200 I, 2004) and Loomis et al. (2002).

Thesecondsectionofthesurveycontained questionsrelated to knowledge and use of EMRsystems andwas based onquestions from Corketal. (1998),Laerumet al. (200 I, 2004),Loomi s etal.(2002),and Kaelberetal.(2005). Questionsaddressedphysician s ' general thoughtstowards EMRsystems,theirusagein thepracticeof healthcare and the cost of implementation and upkeep of an EMRsystem.Five-point Likert scales were usedtomeasure physicia nattitudes about EMRsystems(where 1wasstrongly disagree and 5was stronglyagree) andtheir usage andtheir effecton thepractice ofhealth care (where Iwashighlydetrimental and 5 was highly beneficial).Thissectionalso containednewly created questions to evaluatethe opinionsaroundgovernmentsubsidyof EMR costs.

The final sectionofthe surveycollected demographicdata.Itemsincluded:age, gender,practicesize,communitysize, thenumber ofyears in practice,and area of specialty.Questionsforthis sectioncamefromCorket al.(1998), Kaelberet al. (2005), Laerum et al.(200 1),Simonetal.(2005) andLoomisetal.(2002).Ageand communit y size categorie susedarethe categoriesused in the originalsurvey they weretaken from,

19

andagecategories are the sameasthoseusedin theNeville et al.(2007)study. This section also containedthree open-ended questions allowing physician stoexpress any otherthought s aboutEMRs, theiruseinpracticeandcomments onthis study.

3.3.1 Pretesting

Thesurveywaspre-testedbythreelocalexperts in medical technologies and research methods. As aresult of thepre-test,questionsweremodifiedtomore spec ificallyaddressthecreation of abaselineofNLphysicians'computerskillsandEMR knowledge.Duringthe pre-test,itwasdeterminedthatthe survey could be completedin tenminutes. On the adviceofthis expert panel,apilottestwasnot conducted.

3.4 Data Collection

Mailingaddresses were obtained fromthe CollegeofPhysicians and Surgeonsof NL. Each physician was sent apackage containing the questionnaire, postage-paidpre- addressed return envelope,and a letter explaini ng thepurpose ofthe study.Theletter was signed byDr. G.Farrell,Director of the eHealth Research Unit, MemorialUniversity and bythe studyinvestigator.Theletterinformedphysicians ofthepurpose of thestudy (AppendixC).

The NLMAsupported thisproj ect (Appendix D).TheAssociationincluded informati on aboutthe studyon its websiteandanotice was cmailedout tophysicians, excluding those whoopted out ofthismethodof communicatio n,prior to thesurveys

bein gmailed out.These communications raised awareness ofthe study,assured physiciansthe studywascredibl eand providedabrief overview ofthe study.

Thesurveywasfirst mailedout in September,2007,with a secondpackage sentto non-r espond ents threeweekslater in October,2007.The information letter (AppendixC) usedin the secondpackage wasmodified from the original letterto contai n thank you to anyo newho had previously completed thesurvey and were receivinga follow- up inerror.

Non-respo ndentswere identifi edbyan unique number assigned to each physician and printed oneachsurvey.Thisidentificationnumber was assigned to each physician whowas elig iblefor thestudy by staff at the HealthResearchUnit,Division of Community Health,Faculty of Medicin e.HealthResearchUnitstaff usedthe identifica tion numbers to trackrespondent sand non-respon dents .

Survey result s werekept separatefrom thefilecontaining thephysician contact info rmation(nameand mailing address) . Theuse of an assigned identificationnumber and the separa tefilesensuredthatindividu alphysician surveyresponseswerenot identifi edbythe invest igator.

3.5 Data Managem ent

Surveyresponseswere directly entered into SPSS forWindows,version17.0;where appropriate data were coded/recod ed .Before analysis,the data werecleanedtoident ify andremoveany data entry errors.Theresultsto the open-ended questions were coded intothem es; each theme was assigneda numeric code andthen enteredinto theSPSSfile.

Missingdatawere codedatthedata entrystage.

21

3.5.1 Data Quality

Initialcounts wereconductedto assessthe amountofmissingdata.Onepercent of the datawere missing for the majorityof the questio ns(ranging from 0.2%to 15.9%).

The question sregarding cost of EMRs ("I believethat anaffordable priceperphysicianto set upanEMRsystem is..." and"Iam willingto spend thefollowin g amo unt monthly for ongoing useofanEMR...")hadthehighest amountofmissingdataat 11.5%and 15.9%

respecti vel y.AppendixEprovides detailson thenumber of missing(includingdon't know) for specificquest ions.

Ten percent of surveyswerere-entere d to calculate dataentryerrors rates.To complete the datare-entr y,identificationnumbers for respond entswereentered into excel,randomized,and then thefirst tenpercent (forty)were re-enter edintoSPSS with the number of discrepancies were counted. In2,680variables there werethreeerrors giving an error rate of 0.11%.

Datawerecod eddurin g entry;Appe ndixF shows thecodin gschemefor thesurvey.

Datawere codedasinvalidif anappropriatecatego rycould notbeassignedin consult ation with local experts(Dr.G.Farre1l²,Dr.D. Nevill e",and Dr. V.Gadag", PersonalCommunication,Janu ary 21,2008),thepercent ageof invalidcodingranged from 0.2%to1.0%.Add itionalcoding/cleaningofthedatawereconductedwhen multiplerespon seswereprovidedtoquestions asking for only oneresponse,in these

2Dr. Farrell :Director of the eHealthResearchUnit,Faculty ofMed icine,Memorial Unive rsity of Newfoundland;practicingphysician;EMR advocate

3Dr. Nev ille:Administrative Leadofthe eHealthResearchUnit,Facultyof Medic ine,Mem orialUniversity of Newfoundland;Associate Professorof HealthCarePolicyand Delivery,Division ofCommunity Health and Humanities,Faculty of Med icine,Memorial Universityof Newfoundland

4Dr.Gadag:Professor ofBiostatistics,Divisionof Community Health and Humanities,Faculty of Medicine, Memorial Universityof Newfoundland

cases the data was coded based on therecomm endationsof local experts in the field(Dr.

G. Farrell,Dr.D. Neville and Dr.V.Gaddag,Personal Communication, January21, 2008)andareshow ninAppendix G.Inadditio n,errorswerealso identifi edusing frequenciesand cross-tabul ationsto identifyincorrectorimplausibleerrors. When errors wereidentified,the originalsurveywas consulted.

For the questionto identify respondents'areaofspeciality the original categories were:anaesthe siolo gy,cardiology,critical care, emergency medicine, endocrinology, family medicine, gastroenterology,general internalmedicine ,infectious disease, laboratory medicine,nephrology,neurology, obstetrics/gynaecology, oncology, ophthalmology,orthopaedics,paediatrics,psychiatr y,radiology,rheumatology, surgery, urology,and other.The responsestothisquestion were condensedto two categories

"family physician" and"specia list/other". On therecomm endationof a local expert (Dr.

G.Farrell,Personal Communication,January 21,2008) thosewho selected"family medicine" orwrote "generalpractice" inthe other field werecoded as"familyphysician "

and thosewho eitheridentified as anything outsideofthese were coded as

"specialist/other" .

3.6 DataAna lysis

The statistics program SPSS,version 17.0,wasusedto analyze the data.Given the research objectives,theanalyses were largelylimited todescriptivestatistics(frequencies for categorical dataand meansandstandard deviationsfor ordinal data).The results to the open-ended questions were presented as frequencies, basedon the assignedthemes.

23

Missing,invalid and"don' t know"respon ses were exclud edfrom the analysesoutsideof frequencycounts.

x:

Prior to the analyses,responses tothetwo Likertscale questionswereexamin edto assess the distribution ofthe responses.Using histogram s as wellaskurtos is and skew values for each item , itwas determi nedthat the items werenormall ydistribu ted (excluding the"don' t know"responses).The refore, theseitem swereanalysedasordinal variables, and means andstandard deviations were used(Norman&Streiner,2008).

3.7 EthicalCons idera tions

Thisstudy wasapprovedby the HumanInvesti gation Committee of Memorial University(Appendix H).Alldata were storedonapassword-prot ectedcomputer and all complet ed surveys were stored in asecure room.Result s are presented inaggregate form only,to protect confident iality.

Chapter 4: Results 4.1 Survey Response

There were l083 physic ians listed withtheRoyal Collegeof Physicians and Surgeons of Newfoundl and andLabradoras of August2007.Twenty-nine physicians were excludedsincetheywere participatinginapilot EMR Implementation study being conduct edbythe eHealthResearchUnit. Forty-fivewereexcluded dueto issues with theiraddress, i.e.noneprovided or outofprovince. Of theremaining1009physicians, 409 returned acompl eted survey,giving a responserate of40.5% (see Figure2).

Figure 2: Study sample and respon serate

25

4.2 Representativen ess

The study respondents are representativeofthe complete physicianpopulationofNL for gender, but not whenbrokenoutby specia lty.Thestudysample underrepresented theproportion of familyphysicians, whileover representingtheproportion of specialists (TableI).

Table1: Comp a r isonofphysici an s andsam ple frame toassess representativenessof the studysam ple

Respondents NPS NL 2007 Pvvaluefor'"/! test Gender

Male 274(67.8%) 734(70.6%)

>0.05

Female 130(32.2%) 305 (29.4%)

Sp,eciality

FamilyPhysicians 194 (48.6%) 654(62.9%)

<0.05 Specialist/Oth er 205(51.4%) 385(3 7.1%) NPS=NationalPhysician Survey,NL=Newfoundland and Labrado r

4.3Respondent Characteristics

Table2 describesrespondent characteristics.The majority of respondentswere male (67.8%). Half of the respondentswere in the age category of35-50yearsofage.There was analmostevensplitin thenumber of family physicians (48.6%)andspecialist/other (5 1.4%). The majorit y of respondentspracticedin a community with a populationgreater than 10,000(73.8%)andworked ina small group practice(40.7%). Years of practice rangedfromIto 59 years,with a meanof20 years .

Table 2: Characteri stic sof physician sin the study

Variable n(%

*

Age Category

<35 years of age 49 (12.0)

35-50years of age 207 (50.7)

>50 yearsof age 152 (37.3)

Gender

Male 274(67.8)

Female 130(32.2)

Speciality

Family Physicians 194 (48.6)

Specia list/Other 205 (5 1.4)

Work Setting

Solo practice 81(20.9)

Smallgroup 158 (40.7)

Large group 36 (9.3)

Hospital 92(23.7)

Other 21(5.4)

CommunitySize

<1,000 people 6 (1.5)

1,000-4,999people 40(9.9)

5,000- 10,000people 60 (14.8)

>I0,000 people 299 (73.8)

Yearsof Practice (years)

Range I-59

Mean (sd) 20.0 (11.2)

Median 20.0

*Exceptfor Years of Practice; Variables may addup to less than409 due to missing data

27

4.4 Respondent CurrentComputerUsage

Table3describes respondents' currentcomputerownership and trainin g.The majority (54.0%)reporthaving average computerskills.Alargerproportion of physicians obtained these skillsthroughself-guided learning (67.2%)versusamore formal method oflearning (20.5%)."Other"sourcesoftraining were providedinformally and includedfamily and friends.

The majority of physiciansreported having a desktopcomputerat home (8 1.2%)and atwork(83.9%);among these physicians, 82.2%and 77.3%respectively actually use thesecomput ers. Slightlymorethan seventypercent(72.1%)of physiciansresponded thattheyhad alaptop computerand of these79.3%usedtheirlaptop.Of the42.5%who said they ownedapersonal digitalassistant',72% indicated theyusedit.

Table4illustratessomeof the common uses of computers bythephysiciansin their workplace.Themaj ority of physicians respondedthatthey always used acomput erto obtainthe result s oftests (53.1%), thatthey sometimes used computers to obtainadvice onadiagnosis/therapy (55.2%)and that they neverused a computer to writesick notes (74.4%),order tests (60.8%),refer patients (68.1%)or write prescriptions (82. 1%).Other uses of acompu terinthephysicians 'workplaceincluded:billing,commun icating with patients and their families,and scheduling.The last columnofthistable,"I don't performthistask", allowsus to separate those physicians who donotuse computers for the assigned taskand thosewhose specialitydoesnot require themtoperformthelisted task.

5Persona l digital assistantdoesnot includesmart phones(suchasiPhonesora ndroidp hones)or tabletsas thistechnologywasnot availableat thetime of thisstudy.

Table3:Computerchar acte r istics ofphysicia nsin the study

Self-reportedskill level Veryunsophisticated Unsophisticated Average Sophisticated Verysophisticated

Var ia ble ._. ~-,_n(%)

27(6.7) 83 (20.6) 217(54.0) 62 (15.4) 13 (3.2) Method ofcomputertraining*

No Training

Self-guided learning about computers

Formalworkshops/conferenceson computers (no CMA credit) Formalworkshops/conferences on computers (CMA credit) Formal medical school training in computers Formalcourse(s)in computer science or relatedfield Other

Havea...

Desktopcomputer athome Desktop computer at work Laptopcomputer PersonalDigital Assistant Usc a ...(ofthosewho havethistechno logy)

Desktop computer at home Desktop computer at work Laptopcomputer PersonalDigital Assistant

39 (7.6) 347(67.2) 30 (5.8) 8(1.5) 14(2.7) 54(10.5) 24 (4.7)

332(81.2) 343(83.9) 295 (72.1) 174 (42.5)

273 (82.2) 265 (77.3) 234 (79.3) 126 (72.0) CMA=CanadianMedical Associati on;Variables may add up toless than 409 due to missing data;"Variables mayadd up tomore than 409 due to multiple responses per-ph ysician.

29

Tab le 4:Freq ue ncyof computeruseto perform tasksin practiceamong physiciansinNL

Task

Reviewpatienthistory/recor d~,.

Communic~t;~thcolleague;

'Write sicknotes it"

Obtain adviceona·~pecificpatient'sdiagnosis/therapy Obtainthe resUltS

o f

Variables may addup to lessthan 409duetomissin gdata Never n(%)

101(24.9) 101(25.0) 2%(74.4) 107(26.6) 43 (10.7) 244(60.8) 271(68. 1) 325 (82.1)

Sometimes n(%)

117(28.8) 187(46.3)

'ins:

22-2(55.2) 137(34.0) 41(10.2) 54(13.6) 16 (4.0)

Always n(%)

173(42.6) 105 (26.0) 22(5.'5) 5'S(13.7) 214(53 .I) 54(13.5) 27(6.8) 12(3.0)

I don'fperform this task

n(~)~

15(3.7) 11(2:7) 59(14.8) 18(4.5) 9(2.2)

6 205.5)

4(n m )

43

4.5EMR Knowledge andUse

EMRperceived knowledgeand usefor physiciansin NL are showninTable5.The maj ority (80.1%)of the respondents"agree"or"strongly agree" with thedefinitionof an EMR systemprovided (see section2.2).

An open-endedquestionwasprovidedfor respondentsto express their reasonsfor agreeingor disagreeing with thegiven definition and 27.4%of phys iciansprovidedsome explanation.Of these physicians, 65.2% werein agreement with thedefinition.These physicians feltan EMR system would reduce medical errors,increaseandensurethe lifespan,accuracy,and legibilityof the patientchart. One physiciansummedit upwith thefollowingstatement, "I believeit is simply pathetic thatMedicine,asa discipline,is nottaking advantageof computers.A similarsituation inotherscienceswould be embarrassi ng" .

Thosewho were unsureabout thedefinition(16.1%) ordidnotindicate alevel of agreementwith thedefiniti on (2.6%)stated ignoranceabout these types of systemsor sugges ted thatthesystem mustbeusedin orderfor it to work.Otherreasonsprovided focusedon beingunsure about differentelementsof the givendefinitionandthe perceivedimplication of an improvementin healthcarefrom theuseof thesesystems . Also,concernsabout thepossibility of power outages and computerfailureand theneed to enterallthecurrent patient informationwere cited.

Thosewho disagreed with thedefinition (16.1%)statedconcerns centeredaround the broadness ofthe definition,theuse oftheword"replace" and theissuesthatthey feltthis caused,and the feelingsthat"paper hasbeen usedfor years andworkswell"andthat the

31

EMRwould make healthcaremoredifficulttodeliver, "we should betreatingpatients notcomputers".

When askedif theyhad ever used an EMRsystem46.0%ofphysicians stated that theyhadused anEMR before.Itwas notedhere that some physicianslooked at the MediTech systemas an EMR,whileothersdidnot. The MediTechsystem is anEHR;itis an inform ation systemused in health carefacilities (i.e.hospitals)todocumentand track patienthistory andcare ina comprehensiveand integratedmanner. The definition of EMR systemsusedin this survey was broadenough tothatparticip ants could includethe MediTech system as anEMRsystem.

Respondents were alsoprovidedwithopen-endedquestions to express what theyfelt werethebiggest advantagesand the biggestbarriers tousing anEMRsystem.The responses werecoded into major themes (Table 5).The most commonadvantagegiven was access to patient information and the efficiency this providedforcare; themost common barrier given was technology challenges,suchasuse ofthe EMRsystem, moving paperrecordsto an electronicversion, and generaldiscomfort with using a computer.

At the end of thesurvey,respondentswereprovided with space to add their additionalcommen tsaboutEMRsin general and on themovementtowardsusingthese systems,coded results are found in Table 6 and Table7 respectively. Ofthe409 physician swhorespondedtothe survey,185 (45.2%)supplied a response whenasked about anycomm entstheyhad on EMRs in general,and 174responded(42.5%) witha comm ent whenaskedabout the movement towards theuse of EMRs in clinicalpractice.

Table5:EMR perceivedknowledgeand useforphysician sin NL Variable

Agreement withEM R Definition Strong lydisagree Disagree Unsure Agree Strongly agree Have usedanEMRsyste m

No Yes

Biggestadvantage ofanEM Rsystem*

Access andefficiency Improved healthcare Paperless Costsavings Research

Biggestbarrier relatedtoEMRsystems*

Technology challenges Program and changemanagement Fundingand human resources Data concems Other

n(%

20(5.0) 12 (3.0) 48(11.9) 237(58.8) 86(21.3)

218(54.0) 186(46.0)

287(72.5) 64 (16.2) 33(8.3) 9(2.3) 3 (0.7)

173 (34.0) 121(23.8) 115(22.6) 96(18.8) 4(0.8)

EMR=Electronic MedicalRecord; *Variables may add up to more than 409 due to multipleresponsesper-physician ;Categories based on coding of open-endedquestion s

33

As Table6showsgeneralcomments wereprimar ilypositivein natur e (43.8%), other comments were coded intofourspec ific categories:EMRsystem concern s (including topic s of systemdesign andcompatibility,standardsand security), changemanagem ent concerns(including topics relatedthe trainingof phys iciansandstaff tousethesystem and theprocesstotransferpaper records into theEMR), monetaryconcerns (including topic ssuchasfunding andsystem cost) and miscell aneous (this includedmainly commentsabout theneedfor NLto catch up to other countri esand CanadainEMR use).

Table 6:Physician opinions regardin g EM Rsyste ms ingene r a l Commentson EMRsingen er al

EMRsystemconcerns Change managem ent concerns Monetary concerns General comment-positive General comment-negative Miscellan eous

n(%

33 (14.0) 27(11.5) 30 (12.8) 103 (43.8) 24 (10.2) 18(7.7) EMR=ElectronicMedicalRecord;Categories based on coding ofopen- ended questions; Multipleresponses per-ph ysician

Table7showsthe codin gof the responsesto the open-endedquest ionaskingabout themovem en ttowardsusing EMR systems in clinica l practice,74.1%of thosethat respondedwitha comme nt positive in nature.Ofthose who providedpositi vecomments themajority werecod ed asmiscell aneous (59.2%), these comments generallynoted that EMRs were essential for practice,thatweneededto moveaheadwiththeir implementation and that current progress wastoo slow.Of thosewho provided a negative comment ,themajorit y were codedas miscellaneo us(4 7.5%).These centered on a

varietyof top ics rangingfrom the type offacilitya phys icianworkedin, to the level of governmentcontrol, and the stage of the physicians career. Othercommentsof interest from across all levels of comments were focused on the need for consultations with physiciansduring the entirety of the implementation process,technical support and governmentsubsi dy.

Table 7: Physician opinio nsregarding themovem enttoward susin g EM Rsystems inclinical practic e

Variable Classification of Comment

Negative comment Neutral comment Positive comment Positive Comments Coding

System Monetary Changemanagement Miscellaneous Neutral Comments Coding

System Monetary Changemanagement Miscellaneous Negative Comments Coding

System Monetary Changemanagement Miscell aneous

n(01..) 36 (20.7) 9 (5.2) 129 (74.1)

26 (19.0) 15 (10.9) 15 (10.9) 81 (59.2)

2(18.1) 3 (27.3) 3 (27.3) 3 (27.3)

2(5.0) 5 (12.5) 14(35.0) 19 (47.5) EMR=Electronic Medical Record;Categories based on coding of open- ended questions;Multiple responses per-physician

35

4.6 RespondentPer cepti onsofEMRSystems

Table 8 detailsphysician opinionsregarding severalgeneralstatements aboutEMR systems.Statements were rated on a Likert scaleofIto5, where Iwas strongly disagree and5was stronglyagree. The two statementsthatphysic iansagreedorstronglyagreed with had thehighestlevelof agreementwere"EMRsare ausefultoolforphysician s,such as when docum enting patient information" (mean=4.15)and"Physiciansshoulduse EMRsystems" (mean=4.02).Itis also interestingtonotethatphysiciansdisagreedthat

"EMRswill take away from doctor-patient interactions" (mean=2.44).

Table8: Physician opi nionswithgene ra lstatementsaboutEM Rsystems General Statement

Physiciansshoulduse EMRsystems EMRswill improve thequalityof care EMRs area usefultoolforphysicians, suchaswhen docum entin g patient information

EMRswilltake away fromdoctor-patientinteractions EMRs aremore secure thanpaRer records EMRs aretoo expensive EMRswillreducemedical errors AnEMR willincreasephysician workload EMRsare more confidential thanpaper records

EMR=ElectronicMedical Record; Statementswere ratedon a Likertscaleof I to 5, whereIwas strongly disagree and5 was stronglyagree

4.7 RespondentBeliefs inEMRsEffecton thePractic e of Health Car e Table9 detailsphysician opinionsabout theirpercepti onoftheeffect (beneficialor detrim ental) ofEMRsystemsonthepract iceofhealth care.Statementswere ratedon a Likert scale of I to5,where Iwas highlydetr imental and 5washighly beneficial. The twostatements that physiciansfeltwerebenefi cial or highlybeneficialtothepracticeof health carewere"Clinicians'accesstoup-to-d ateknowledge" (mean=4.20 )and"Access tohealth carein remote or rural areas" (mean=4.10).

Table 9:Physici an opinionswith theben efi cial effectofE M Rsyste ms on the practiceof healthcare

The effect that EMRslIlayhave on medicin eand health care inrelatio~l1...,to..._ _ , . - _ ...,. _ Costsofhealth care

Qualityof health care

Accesstohealth care inremoteor rural areas Enjoymentof thepracticeof medicine Person al and professional privacy Doctor-Patientrelati onship Clinicians' access toup-to-dateknowledge Patients' satisfac tionwith the qualityof care theyreceive Roleof governmentin healthcare

Therapportbetweenclinicians and patients

Mean (std deviation) 3.32 (0.88) 3.86 (0.71) 4. 10(0.68) 3.54 (0.84) 3.21 (0.84) 3.18(0.7 1) 4.20(0.64) 3.50 (0.69) 3.14(0.80) 3.20 (0.71) EM R=Electron ic Medica l Record;StatementswereratedonaLikert scaleof I to 5,whereI washighlydetrimental and 5 washighlybeneficial

37

4.8 RespondentPercepti onsRegardingtheCostsofEMRSystems

Physicianswere asked questions about the costof implementingan EMRsystemand itsmonthly cost of maintenance,including question s ongovernmentsubsidy.Table 10 detailsresponses regarding set-up costs and TableIIshows thoseresponsesrelated to monthl ymaintenance costs.

Alargepercent of physicians (4 1.9%) felt thattheinitialsetupofanEMRsystem should costbetween $1,000 and$4,999perphysician.Most(94.9%)felt thegovernment should subsidizethe initialsetup cost.Almost half(4 8.2%)believedthat100%of the set-up costsshouldbe covered by thegovernment.

In relationto ongoingmonthlymaintenance,almost half (4 1.9%) ofphysicians respondedthatthe monthly ongoing costofusing anEMR systemshould belessthan

$50.Alargenumber(87.9%) felt thatthe governmentshouldsubsidize themonthly maintenance fees.Almosthalf (47.0%)believed that the governmentshouldcover the wholeofthemonthlycost.

Table10: Physicianopinions regardingset up eosts of EMR systems

Vari abl e n (%)

Ibelievethatanaffordableprice perphysicianto set up an EMRis...

<$1,000 95(26.4)

$1,000 -$4,999 151 (4 1.9)

$5,000- $9,999 69 (19.2)

$10,000- $20,000 31(8.6)

>$20,000 14(3.9)

Do you believethe governmentshould subsidizethe costof EMR installation?

No 20(5.1)

Yes 376(94.9)

What percentage of the set up costdo you feel thegovernmentshouldcover?

5% 0(0.0)

10% 3 (0.8)

15% 4(1.1)

25% 12 (3.4)

50% 101 (28.3)

75% 65(18.2)

100% 172 (48.2)

EMR=ElectronicMedical Record; Variablesmay add up tolessthan409due tomissingdata

39

144 (4 1.9) 81(23.5) 60 (17.4) 45 (13.1) 14 (4. 1) Table11: Physicianopinionsregarding ongoingcostsof EMRsyste ms

Variable n(%)

Iam willing to spend thefollowingamount monthlyfor ongoing useof an EMR system ...

<$50

$50-$99

$100-$ 149

$150 -$200

>$200

Do you believethe governmentshould subsidizethemonthlycost ofE MR usage?

No 47(12.1)

Yes 342(87.9)

Whatpercentageof the set upcost do you feelthegovernment should cover?

5% 1(0.3)

10% 5(1.5)

15% 1(0.3)

25% 16 (4.8)

50% 107 (32.0)

75% 47 (14.1)

100% 157(47.0)

EMR=ElectronicMedicalRecord ;Variablesmayadduptoless than409dueto missingdata

4.9Other Com ments

At the endofthe survey,respondents were provided withspace to give comment son this study and questionnaire.A total 103 physician s(25.2%)provid ed aresp on setothis questi on (Ta ble 12).Asshown in Table 12,theresponsestothisquesti onweremostl y positi ve in natur e (65%),and were compose dof comments along thelines of good, interestin g, orvaluable.Thosecommenting directl y on the survey indicatedboth positiv e commentsandsugges tionsfor improving the survey. Positi ve commentson the survey were in the veinofresp ond ent s feeling that the survey len gthwasacceptable,thatthe questionscovereda comprehen sivelookat the topic .Thoseindicatinga waytoimprove the surveysuggestedthatthe wordingofcerta in questi onsfavoured EMRsandtheiruse, othersindicatedawishfor thesurveytobeavailabl e online,and othersindicatedaneed to distinguishbetweenhosp ital- based and community-basedphysician s, aswellastheir fee-far-service versussalariedfunding.

Tab le12: Physicianopinions on thisstudyand questi onnai r e

67(65.0) 9 (8.7) 18(17. 5) 2 (1.9) 2(1.9) 5 (4.9) 1 C=a__tegon ''''''- --:.:..'~·,_ _1

Generalcomme nt-positive Genera lcomment-negati ve Commentonsurveyformat Comme ntoncost Comme nt onEMR - negative Miscellan eous comments

EMR=electroni cmedi calrecord;Categoriesbased oncodingofopen-ended question

41

Chapter 5: Discu ssionandConclusions

Thisstudy describesthe characte ristics,computertrainingand use,perceptionsabout EMRsystems,andopinionsonacceptablecostsEMRsystemsofphysiciansinNL.This inform ationhasbeen shownin theliterature toindicate which physicians aremorelikely toaccept and use anEMRsystem andtohelp guide implementationplans.

5.1 Representativene ss of the Sample

Surveyresponserate of 40.5%is consideredreasonable forthis study population.

TheNPS(2007), whichsurveyedall physiciansinCanada,hadaresponserateof32.1%

usinga mailedpaper survey.Surveysof small subsetsofphysiciansinNLhadhigher responserates of50.0%(Gates,2004)and 84.0% (MacEa chern,2009);theseratesare thoughttobehigher dueto the smaller size and specializednatureofthese studies populations.

Respondents arerepresentative of the physicianpopulationofN L withregardsto genderbutnotspeciality.Thestudy sample underrepresentedtheproportion of family physicians whileover representingthe proportion of specialists.Thisdifferencecould resultin respondentbias.Since speciali stsinthe provinceare generally locatedin hospital settingsor clinics in largeurban areas,theycouldhavemore knowledgeof differentinformati ontechn ologies and theuse of computersin theirpractice ofhealth care.Thisdifference couldalsoindicate that specialists havemoreinterestin thetopicof EMRsystems.Theresponsebias could bea reflectionof survey respondent fatigue amongfamilyphysicians,who mayhavereceivedmany surveys onthistopic.

5.2 Current Use and Training with Computers

Eightoutoften physicians reported having a computerat homeandat theirplace of work.Themajority of physiciansfeeltheyhave an averageskill levelusingtheir computer; these skillswereprimarily obtained mostly throughself-guided learnin g. These findings aresimilar to thosefoundintheliterature. Cork et al.(1998) reportedthat the respondent s (full-time physician faculty members)to thatstudy,self-rated theircomput er skillsasaverage and that "self-guided learnin g" was the main typeof trainingfor computer skills.In the workplace, physicians'primarilyusetheir computer to obtain patient testresult s and toreview patient histories.Relativelyfewphysician suse their computersto write sick notes, order prescriptions, order tests, ortoreferpatients.These findings sugges t that computersare not well integratedintothepractice ofhealthcare by physiciansinNLand that theyare used topassivelyreceive and review information but notactivelyusedto communica tewith other physicians orresources.

5.3 Physician KnowledgeandUseof EMRsystems

Themaj ority of physicians agreedwith the provided definitionof anEMRsystem (see section2.2).Physicians who agreed with the definitionfelt that anEMRsystem would improve health care practice, while those whodisagreed with the definition expressed astrong dislike ofintroducing computers intothehealth care process.

Lessthan half of thephysicians indicatedthey hadused anEMR system.Thisseems tobe dueto thegeneralconfusionwithregards totheMediTech systemand participants beingunclear ofthedifferences betweenEHRandEMRsystems, with different

43

respondents indicating they had and had not used an EMRsystem but used MediTec h.

Thisdifference in knowledge of EMR systems, and MediTech inparticular, could be a resultof thesurvey sample over representing the proportion of specialistsfor NL; since specialistsgenerally work in hospitals they would have moreknowledgeofMediTech and have used the system more than family physicians in communitypractices and thus treatthe systemas an EMR.

5.3.1 Physician Perceptions ofEMR Systems

The survey results indicate that physiciansare in favour oftheimplementation and use ofEMRsystemsin their practice ofhealth care.Thereis ahighlevel of agreement to statementssuchas: "Physicians should use EMR systems" and"EMRsare ausefultool for phys icians...". These findingsare similar to those foundby Loomiset al. (2002), who surveyed active members of the Indiana Academy of Family Physicians (both users and non-users ofEMRsystems).Loomis et al. (2002) found that themajority ofEMR users and non-users agreed that"physicians should computerize their medical records".

This generally positiveview ofEMRsystemsamong physicians in this study was furtherexpressedin the responsesprovided for thc generalopen-ended questionabout EMRs, wherephysiciansexpressed overwhelminglythat EMRs wouldimprove accessto and the efficiencyof care provided.Thesereasons echoed statements providedbythe survey respondentsinthe Neville et al. (2007) study.Nevilleetal.(2007)conducteda study to assessthe feasibility of implementing an EMR systeminNL,the sampleforthis

study included allstaff(physicians,licensedpractical nurses,officestaff, and resident s) at fourclinicsin the St. John'sarea ofNL.

Whilephysicians perceivedEMR systemspositively,concerns were expressedover their use and implementation. These concerns matched those commonlyexpressed in the literatureand wererelated to system costs(Boonstra&Broekhuis, 20 I0;JohnstonetaI., 2002;Loomis etaI.,2002; LudwicketaI., 2010; Retchin,1999; SimonetaI.,2005;Terry etaI.,2009; Wang ctaI.,2003),changemanagement(Boonstra&Broekhui s,2010;

Clayton etaI., 2005; Gadd&Penrod,2001; Joos etal.,2006; Laerumetal., 2001; Loomis eta!., 2002; Ludwicketa!.,2010;Terryeta!.,2009; Young,1984),and technology concerns (Boonstra&Broekhuis,2010; Johnston etal.,2002;JoosetaI.,2006; Kaelber et aI.,2005;Loomiseta!., 2002; Ludwick et al.,2010;OECD,20 10; Retchin,1999 ; SantiageetaI.,2006; Terry et a!.,2009; Young,1984). Additionally,some physicians indicatedthatthe personaltraits of some (i.e.lengthof time practicingandage of physician) wouldbe a barrier to the implementation anduse ofEMRsystems. These findingsare similarto resultsfrom Boonstra and Broekhuis (2010)and Joos etal.(2006), as wellas other literature ("Physicianuse ofEMRs",2005; FordetaI.,2006;Miller etaI., 2005;SimonetaI.,2005).

5.3.2 Physician Beliefs Regarding EMRs Effect011the Practice of Health Care

Themajority ofphysicians supported themovetowardsusing anEMRsystem, as indicated bytheirresponses to the open-endedquestionssystem. Theyalsoindicated a needtomoveforwards withimplementi nga standardandinteroperableEMRsystem for

45

NL, so that theycouldcatch upand keeppace with therestof Canadaand other countries.Physiciansperceived EMRsystemsas enabling them to accessmore up-to- dateinformationandallow forimproved access forremoteand ruralareas.Perceptions of improved accessforremote and ruralhealth carewere alsofoundbyCork etal. (1998).

Stillas the results ofthis studyand theliteraturedemonstrate,themajorityof benefit s listedfor using EMRsystems are perceptionsand more research isneededto establish actual benefit s from the implementatio nofEMRsystems.ThisisillustratedbyGreiver etal.(2011) whostudied the implementationofEMRs forspecific services usingpay-for- performanceincentivesinOntario.Greiver et al.(2011) found that therewas no significantchange in thepractice ofhealthprovidersrelatedto these serviceseven with thefeaturesthe EMRsystem provided suchasreminders.

5.4 Physician sPer cepti ons ofEMRCosts

Physiciansoverwhelmin gly agreed that the costsof changingfrom paper,training themselvesand their staff, alongwithhardware,maintenance,andsupport costs should be subsidized by government insomeway.The OECD(2010)recently released areportthat statesthatthat subsidiesor grantsarenecessaryfor start-up but doesnotinfluencethe continueduse ofthe EMRsystemand thatfinancing policyneedstobeput intoplace prior to system implementation.The report states thatthe adoption ofEMRsystems is moresuccess ful in countri es wheresubsidies or grantsare usedto "insulate" physician s fromany lossesfromup-front costsandpotentialdecreasein productivity.Ludwicketal.

(2010)found thatdifferentremu neration approacheswereneededfor EMRadoption