Sick assessment and prognosis - MarselisborgCentret

Sick leave due to neck or shoulder pain:

Interventions, MRI assessment and prognosis

PhD dissertation

Line Thorndal Moll

Health

Aarhus University

Department of Public Health,

Section for Clinical Social Medicine and Rehabilitation

2018

Supervisors

ClausVintherNielsen,MD,PhD,professor

DepartmentofPublicHealth,SectionforClinicalSocialMedicineandRehabilitation

AarhusUniversity,Denmark

MereteLabriola,OT,MPH,PhD

DepartmentofPublicHealth,SectionforClinicalSocialMedicineandRehabilitation


OleKudskJensen,MD,PhD

SpineCentre,DiagnosticCentre,SilkeborgRegionalHospital

8600Silkeborg,Denmark

ChristinaMalmoseStapelfeldt,MHSc,PhD

DepartmentofPublicHealth,SectionofClinicalSocialMedicineandRehabilitation


BeritSchiøttz‐Christensen,MD,PhD,professor

SpineCentreofSouthernDenmark,HospitalLillebaeltMiddelfart

InstituteofRegionalHealthResearch

UniversityofSouthernDenmark,Odense,Denmark

Evaluation Committee

Professor Vivi Schlünssen (chairman and moderator of the defence)

Section for Environment, Occupation and Health, Department of Public Health

Aarhus University

Bartholins Allé 2, bygning 1260

DK-8000 Aarhus C, Denmark

Professor Kirsten Fonager

Klinisk Institut, Aalborg Universitetshospital

Hobrovej 18-22

DK-9000 Aalborg, Denmark

Professor Pierre Côté

Science Building - Room 2031

North Oshawa

2000 Simcoe Street North

Oshawa, ON L1H 7K4

University of Ontario, Canada

This dissertation is based on the following studies

Study1.LineThorndalMoll,OleKudskJensen,BeritSchiøttz‐Christensen,ChristinaMalmoseStapelfeldt,DavidHøyrupChristiansen,ClausVintherNielsen,MereteLabriola.ReturntoWorkinEmployeesonSickLeaveduetoNeckorShoulderPain:ARandomizedClinicalTrialComparingMultidisciplinaryandBriefInterventionwithOne‐YearRegister‐BasedFollow‐Up.JOccupRehabil2018Jun;28(2):346‐356.doi:10.1007/s10926‐017‐9727‐9.

Study2.LineThorndalMoll,MortenWasmodKindt,ChristinaMalmoseStapelfeldt,TueSecherJensen.DegenerativefindingsonMRIofthecervicalspine:aninter‐andintra‐raterreliabilitystudy.AcceptedforpublicationinChiropracticandManualTherapies.AcceptancereceivedonAugust6th2018.

Study3.LineThorndalMoll,AnneMetteSchmidt,ChristinaMalmoseStapelfeldt,MereteLabriola,OleKudskJensen,MortenWasmodKindt,TueSecherJensen,BeritSchiøttz‐Christensen.Predictionofworkparticipationwithin2yearsinsicknessabsenteeswithneckorshoulderpain:thecontributionofdemographic,patient‐reported,clinicalandimaginginformation.Thefinaldraftisapprovedbyallauthors.Themanuscriptwillbesubmittedafterassessmentofthedissertation.

ThestudiesarefoundinAppendices1‐3.

Acknowledgements

ThisdissertationcomprisestheworkofmyPhDstudieswhichwereperformedbetween2015and2018whileemployedattheDiagnosticCentre,SilkeborgRegionalHospital,inclosecollaborationwithandkindlyhostedbyDEFACTUM,CentralDenmarkRegion.

Iamgreatlyindebtedtoanumberofpeople,andespeciallytomysupervisors,Merete,Christina,Ole,BeritandClaus,forgivingmethespaceandopportunitytoundertakethisperiodofscholarship.Ihavelearnedsomuchfromallofyou.MysincerethankstoClausandUlrichwhohavesupportedtheprojectinbothtangibleandintangibleways,withoutwhichtheprojectwouldnothavebeenpossible.IalsoowemygratitudetoTueandMortenwhoinvitedmeintotheworldofmagneticresonanceimaging(MRI)andtaughtmenumerousthings,notonlyregardingMRI,butalsoaboutunderstandingacademicissuesandorganisationalcontexts.

ThestudysampleoriginatesfromarandomisedcontrolledtrialwhichwasplannedandconductedbeforeIstartedmyPhDstudies.Iamgratefultothestudyparticipantsandtoallthecliniciansandresearchersinvolvedwholetmeusesomeofthedatathatformedthebasisofthisdissertation.

AspecialthankstoDEFACTUMforwelcomingmeintotheresearchenvironment,andtomembersoftheDepartmentforDiagnosticImaging,SilkeborgRegionalHospital,foryourhospitalitywhileperformingtheMRIassessments.

Thankstomyfellow'sprouts'CharlotteMariaandCecilieforcontributingtoaworkingenvironmentwithbothacademicdiscussionsandlaughter.AspecialthankstoAnneMetteforcountlessthings,butmostofallforyourfriendship.

IalsoowemygratitudetoKirstenFonagerandPierreCotéwhoagreedtoevaluatemywork.Iamhumbledandhonouredforthetimeandeffortyouhavespentonmywork.

Lastbutnotleast,Iwouldliketothankmyparentsforalwayshavingbelievedinme.Thankstomyfamilyandclosefriendsforyoursupportandencouragementalongtheway.Inparticular,thankstoJesper,Jakob,EmilandIda.Thegreatestprivilegeinmylifeissharingeachdaywithyou.

LineThorndalMoll,September2018

Funding

TheworkwassupportedbyDEFACTUM,AarhusUniversityDenmark,TrygFoundation,DanishRheumatismAssociation,AaseandEjnarDanielsenFoundation,andHelgaandPeterKorningFoundation.

List of abbreviations AC:AgreementbyChance

AUC:AreaUndertheCurve

BI:BriefIntervention

CI:ConfidenceInterval

CNFDS:CopenhagenNeckFunctionalDisabilityScale

DASH:DisabilitiesoftheArm,ShoulderandHand

DREAM:DanishRegisterforEvaluationofMarginalisation

GP:GeneralPractitioner

IQR:InterQuartileRange

Κ:Kappa

LBP:LowBackPain

MCIC:MinimallyClinicallyImportantChange

MDI:MultidisciplinaryIntervention

NPV:NegativePredictiveValue

NRS:NumericRatingScale

OA:ObservedAgreement

PPV:PositivePredicitiveValue

RCT:RandomizedControlledTrial

RTW:ReturnToWork

TP:TenderPoint

SD:StandardDeviation

s‐WPS:successfulWorkParticipationScore(i.e.WPS≥75%inWeeks30to104)

u‐WPS:unsuccessfulWorkParticipationScore(i.e.WPS<75%inWeeks30to104)

VESC:VertebralEndplateSignalChanges

WPS:WorkParticipationScore

Concepts used in this dissertation

MagneticresonanceimagingAnimagingtechniquewhichdoesnotinvolveionizing

radiation.Itisbasedonthefollowingsteps:1)apatientisplacedinamagnet,2)aradio

waveissentinandturnedoff,3)thepatientsemitsasignalwhichisreceivedand4)the

receivedsignalisusedforreconstructionofthepicture(1).

ModicchangesSee'vertebralendplatesignalchanges'.

MultidisciplinaryThistermdescribestheinvolvementofatleasttwodifferent(health

care)professionalswhoofferdistincttherapeuticmodalities.

MyelopathyPathologicchangeofthespinalcord,oftencausedbycompressiondueto

spinalcanalstenosis.

NeckpainPainlocatedintheposteriorregionofthecervicalspine,fromthesuperior

nuchallinetothefirstthoracicspinousprocess.Itmayrefertothehead,theshoulder,the

anteriorchestwall,ortheupperextremity.

Predict/predictionWhenused,thesetermsdescribetheabilitytoforecasttheprognosis.

Prognosis/prognosticfactorPrognosisdescribesthepredictedcourseofagivenstate.

Forinstance,itmaydescribethepredictedcourseofsickleave.Aprognosticfactorisa

variablewhichisassociatedwithprognosisbutwithoutmakinginferencesaboutcausation

(2).

RadiculopathyAclinicalassessmentrequiring1)painradiatingfromthenecktothe

upperextremityand2)oneormorepositiveclinicalfindings:diminisheddeeptendon

reflexes,decreasedmusclestrength,dermatomalsensorydeficitsorpositiveforaminal

compressiontest.

Reliability"Theextenttowhichscoresforpatientswhohavenotchangedarethesamefor

repeatedmeasurementunderseveralconditions"(3).

ReturntoworkAsaconcept,themeaningof'returntowork'isliteral,i.e.resumptionof

paidworkafteraperiodofsickleave.Whenusedasanoutcome,'returntowork'willbe

operationalisedintherelevantmethodssection.

RotatorcuffdisordersSee'subacromialimpingementsyndrome'.

SickleaveAbsencefromwork(part‐timeorfull‐time)duetoworkdisability.Peopleon

sickleaveareconsideredasbeingtemporarilyworkdisabled(asopposedtohaving

permanentworkdisability).

Signal‐to‐noiseratioAmeasureusedinMRIwhichcomparesthetruesignal(reflecting

anatomy)tothelevelofbackgroundnoise.Ingeneral,alowsignal‐to‐noiseratioresultsin

grainyimages.

ShoulderpainPainlocatedaroundtheshoulder.Itmayoriginatefromtheneck,fromthe

joints(glenohumeral,acromioclavicular,orsternoclavicular),orfromsofttissuesadjacent

totheshoulder.

SubacromialimpingementsyndromeCompressionofthesupraspinatustendonbetween

thehumeralheadandtheacromion,coracoacromialligamentandtheinferiorborderofthe

acromioclavicularjoint.Itisaclinicalsignseenindifferentstagesofrotatorcuffdisorders

whichconstitutethemostcommonreasonforshoulderpain(4).Rotatorcuffdiseasemay

includetendinitis,bursitis,tendonchangesorbonydegeneration(5).

VertebralEndplateSignalChangesSignalintensitychangeinvertebralbodymarrow

identifiedonMRI.Howtodistinguishbetweentype1andtype2isdescribedinAppendix4

(theevaluationmanualforStudy2).Type1isassociatedwithhistologicalsignsofendplate

fissuring,degeneration,andgranulationtissuesalongwiththemarrowbeinginfiltratedby

multiplesmallbloodvessels(6).Type2isassociatedwithhistologicalsignsofchronic

repetetivetraumaoftheendplatesandthemarrowbeingreplacedbyabundantfat(6).

WorkdisabilityAsituationwhereaworkerisunabletostayatorreturntoworkbecause

ofaninjuryordisease(7).

WorkoutcomesResearchoffersnumerousdifferentoutcomesregardingsickleaveand

returntowork;forinstancedurationofsickleave,timetofourcontinuousweeksofself‐

support,workstatusatagiventimepoint(atwork:yes/no),workparticipationscore

(WPS)etc.Asfaraspossible,thespecificdefinitionsarepresentedinreferencetothe

relevantcitations.Butwhennecessary,theterm'workoutcomes'isusedasacommon

term.Forinstance'improvedworkoutcomes'willbeusedifdurationofsickleaveis

reducedinsomestudieswhileinothers,workparticipationscorewasincreased.Thisalso

applieswhenreferringtoStudies1and3(theformerusedtimeto4consecutiveweeksof

self‐supportandthelatterWPS).Insuchcases,theterm'workoutcomes'willalsobeused

asacommonterm.

WorkparticipationscoreAfractionyieldingscoresbetween0%and100%.The

numeratorcomprisesthenumberofworkingweekswhilethedenominatorcomprisesthe

totalnumberoffollow‐upweeks(8).

Figures

Figure1.Thearenainworkdisabilityprevention

Figure2.Flowchart,Study1

Figure3.Flowchart,Studies2and3

Figure4.Fractionstillonsickleave,Study1(Kaplan‐Meier)

Tables

Table1.Overviewofstudies,designs,samplesizes,outcomevariablesandstatistical

analyses

Table2.ContactswiththeSpineCentre,Study1

Table3.MRIfindingsandcorrespondingclassifications,Studies2and3

Table4.Baselinecharacteristics,Study1

Table5.Inter‐raterreliabilityestimates,Study2

Table6.Un‐weightedkappaestimatesforneuralforaminalstenosis.Allimagesvs.only

imageswithobliqueslicesavailable,Study2

Table7.Intra‐raterreliabilityestimates,Study2


Table9.Crudeoddsratiosforworkparticipationscore≥75%(s‐WPS),Study3

Table10.Predictivevalues,specificity,sensitivityandAUCofModels1‐4,Study3

Table11.Cross‐tabulationsshowingthedistributionbetweentheclassificationofchance

andactualWPSoutcomes,Study3

Table12.OverviewofregistrationmechanismsinDREAMandpotentialtypesof

misclassification

SUMMARY

Table of contents Summary ............................................................................................................. 14 Background...............................................................................................................................................................14 Methods.......................................................................................................................................................................14 Results.........................................................................................................................................................................15 Conclusions................................................................................................................................................................15

Dansk resumé ...................................................................................................... 16 Baggrund....................................................................................................................................................................16 Metode.........................................................................................................................................................................16 Resultater...................................................................................................................................................................17 Konklusioner.............................................................................................................................................................17

Background .......................................................................................................... 18 Theburdenofneckandshoulderpain..........................................................................................................18 Workdisability,sickleaveandreturntowork..........................................................................................18 Sickleaveduetoneckorshoulderpain........................................................................................................19 Benefitsofwork......................................................................................................................................................20 Thearenainworkdisabilityprevention(stakeholders)......................................................................21 Interventions............................................................................................................................................................22 AssessmentofcervicalspineMRI....................................................................................................................23 Prognosis....................................................................................................................................................................25

Aims .................................................................................................................... 27 Aim1............................................................................................................................................................................27 Aim2............................................................................................................................................................................27 Aim3............................................................................................................................................................................27

Methods .............................................................................................................. 28 Study1,theRCT......................................................................................................................................................29 Design,settingandparticipants......................................................................................................................29 Theclinicalinterventionforallparticipants.............................................................................................31 Thebriefintervention(BI).................................................................................................................................32 Themultidisciplinaryintervention(MDI)...................................................................................................32 NestedRCT................................................................................................................................................................33 Baselinevariables..................................................................................................................................................33 Outcomes...................................................................................................................................................................34 Analyses......................................................................................................................................................................35

Study2,thereliabilitystudy..............................................................................................................................36 Sampleanddatacollection................................................................................................................................36 Readers.......................................................................................................................................................................37 Evaluationmanual,pilotingandworkstations.......................................................................................38 Variables....................................................................................................................................................................38 Analyses......................................................................................................................................................................39

Study3,theprognosticstudy............................................................................................................................41 Designandparticipants......................................................................................................................................41 Outcomevariable:Workparticipationscore(WPS)..............................................................................41 Prognosticvariables.............................................................................................................................................41

SUMMARY

Analyses......................................................................................................................................................................42 Nomenclature..........................................................................................................................................................44

Context:TheDanishLabourMarketandTheDanishSicknessBenefitAct..................................44 Ethicsapproval........................................................................................................................................................46

Results ................................................................................................................. 47 Study1.........................................................................................................................................................................47 Study2.........................................................................................................................................................................49 Study3.........................................................................................................................................................................51

Discussion ............................................................................................................ 57 Comparisonwiththeliterature,Studies1and3......................................................................................57 Populationcharacteristics(discussingworkoutcomes)......................................................................57 Populationcharacteristics(discussingtheprognosticmodels)........................................................59 Componentsofinterventions............................................................................................................................61 Work‐relatedandsocietal/legislativefactors...........................................................................................62

Comparisonwiththeliterature,Study2......................................................................................................63 Strengths.....................................................................................................................................................................64 Limitations.................................................................................................................................................................66 Samplesizeconsiderations................................................................................................................................66 Bias...............................................................................................................................................................................67 Confounding.............................................................................................................................................................67 Self‐reportedsickleaveandregisterdata...................................................................................................68 Externalvalidity.....................................................................................................................................................71

Conclusions .......................................................................................................... 73 Implications ......................................................................................................... 74 Forclinicalpractice................................................................................................................................................74 Forresearch..............................................................................................................................................................74

Public health perspectives ................................................................................... 76 References ........................................................................................................... 77 Appendices .......................................................................................................... 92

SUMMARY

Page|14

Summary

Background

Sickleaveduetoneckorshoulderpainisasubstantialburden,bothfortheindividualand

thesociety.Asoneofthestakeholders,healthcareprofessionalsperformorcontributeto:

interventionsaimedatreturntowork(RTW),assessmentofcervicalspinemagnetic

resonanceimaging(MRI),andassessmentofworkprognosis.Theidealcompositionof

interventionshasnotyetbeenestablished.Fewreliabilitystudieshaveassessedseveral

degenerativefindingsoncervicalspineMRI.Nostudieshaveexploredthecontributionsof

demographics,patient‐reported,clinical,andMRIinformationalthoughthesedifferent

typesofinformationareoftenavailablewhenassessingworkprognosis.

Theaimsofthisdissertationwere:1)toevaluatetheeffectofamultidisciplinaryinter‐

ventioncomparedwithabriefinterventioninsicknessabsenteeswithneckorshoulder

pain,2)todeterminetheinter‐andintra‐raterassessmentreliabilityofdegenerative

findingsoncervicalspineMRI,and3)toexplorethedegreetowhichdemographic,patient‐

reported,clinical,andMRIinformationcontributestothepredictionofworkparticipation.

Methods

Inasecondarycaresetting,Study1wasanRCTincluding168sicknessabsenteeswhomet

theinclusioncriteria:age18‐60years,4‐16weeksofsickleaveduetoneckorshoulder

pain,andfluencyinDanish.Exclusioncriteriawere:drugaddiction,primarypsychiatric

disorder,recentorplannedspinesurgery,pregnancy,andspecificmusculoskeletal

disorders.Participantswererandomlyallocatedtothemultidisciplinary(n=85)orbrief

intervention(n=83),andfollowupwas1year.Basedonregistrydata,theprimary

outcomeofRTWwasdefinedas4weeksofself‐supportandanalysedbyCoxproportional

hazardsanalysis.Secondaryoutcomeswerepainanalysedbylogisticregression,and

disabilityanalysedbylinearregression.

Study2wasareliabilitystudyusing50cervicalspineMRIschosenfromthoseavailablein

Study1.Anevaluationmanualwithclassificationsoffindingswascomposed,anda

chiropractor,aradiologistandasecond‐yearresidentofrheumatologyindependently

SUMMARY

Page|15

assessedkyphosis,discheight,disccontour,vertebralendplatesignalchanges,spinalcanal

stenosis,neuralforaminalstenosis,uncovertebralosteoarthritisandzygapophyseal

osteoarthritis.Inter‐raterreliabilitywasassessedpairwiseandforallthreereaders,while

intra‐raterreliabilitywasassessedforthesecond‐yearresidentofrheumatology.

Prevalenceestimatesandun‐weightedkappastatisticswerereported.

Study3wasaretrospectivecohortstudyincludingallthesicknessabsenteesfromStudy1

(n=168).Asuccessfulworkoutcomewasdefinedas≥75%workparticipationscore

(WPS)fromWeeks30to104afterenrolment.Baselinevariableswerecategorizedinto

demographic,patient‐reported,clinical,andMRIdomains.Crudelogisticregression

analyseswereusedtoidentifyprognosticvariables(p<0.2),followedbymultivariable

analysesincludingtheprognosticvariablesinadomain‐wiseorder.Foreachadded

domain,theprobabilityofasuccessfulWPSwasdichotomized(≥50%chanceornot),and

positiveandnegativepredictivevalues,sensitivity,specificityandareaunderthecurve

(AUC)werecalculated.

Results

InStudy1,comparisonofthemultidisciplinaryandthebriefinterventionsyieldedno

statisticallysignificantdifferenceswithrespecttoRTW,painordisabilityat1‐yearfollow

up.InStudy2,mostlysubstantialinter‐raterreliability(K≥0.61)wasfound,andintra‐

raterreliabilityestimateswerehigherforallfindings.InStudy3,demographicandpatient‐

reportedvariablesyieldedapositivepredictivevalueof0.72,asensitivityof0.74,andan

AUCof0.76.Positiveandnegativepredictivevalues,sensitivity,specificityandAUCdidnot

improvenotablybyaddingclinicalandMRIinformation.

Conclusions

Firstly,themultidisciplinaryinterventiondidnotimproveRTW,painanddisability

comparedwiththebriefintervention.InStudy2,theuseofsimple,comprehensible

classificationsofdegenerativefindingsyieldedmostlysubstantialinter‐raterreliability,

andsubstantialtoalmostperfectintra‐raterreliability.Thirdly,clinicalandMRIvariables

providednoadditionalinformationforthepredictionofworkparticipationcomparedwith

onlydemographicandpatient‐reportedinformation.

DANSKRESUMÉ

Page|16

Dansk resumé

Baggrund

Sygefraværpga.nakke‐ellerskuldersmertererenvæsentligbyrdeforbådeindividerog

samfund.Blandtenrækkeaktørerharsundhedsprofessionelleansvarforellerbidragertil:

interventionerrettetmodtilbagevendentilarbejde(TTA),vurderingafMR‐scanningaf

nakkensamtvurderingafprognosenmht.arbejde.Denideellesammensætningafen

interventionrettetmodTTAerendnuikkeetableret.Kunfåreproducerbarhedsstudierhar

vurderetfleredegenerativeMR‐fund.Ingenstudierharvurderetbidragetafdemografisk

oganamnestiskinformation,objektivefundsamtMR‐fund,selvomklinikereoftehar

adgangtildissefiretyperinformationnårprognosenvedr.arbejdevurderes.

Formålenemeddenneph.d.‐afhandlingvar1)atsammenligneenmultidisciplinærmeden

kortinterventionblandtsygemeldtemednakke‐ellerskuldersmerter,2)atvurdereinter‐

ogintra‐læserreproducerbarhedenvedvurderingafMR‐scanningafnakkenog3)at

undersøgeihvilkengraddemografiskoganamnestiskinformation,objektivefundogMR‐

fundbidragertilforudsigelseafprognosenvedr.arbejdsdeltagelse.

Metode

Studie1varetrandomiseret,kontrolleretstudiepåRegionshospitaletSilkeborg.

Inklusionskriteriernevar:alder18‐60år,4‐16ugerssygefraværpga.nakke‐ellerskulder

smerter,oggodedansk‐evner.Eksklusionskriteriernevar:stofmisbrug,primærpsykiatrisk

lidelse,nyligrygkirurgi,planeromrygkirurgi,graviditetogkendtspecifikbevægeapparats‐

sygdom.Ialt168blevinkluderetograndomiserettilentenmultidisciplinær(n=85)eller

kortintervention(n=83).Follow‐upvar1år.Udfraregisterdatablevdetprimære

outcomeTTAdefineretsom4ugersselvforsørgelseoganalyseretvha.Coxregression.De

sekundæreoutcomessmerteintensitetogfunktionsevneblevvurderetvha.hhv.logistisk

oglineærregression.

Studie2varetreproducerbarhedsstudieomfattende50MR‐scanningerafnakkenfra

Studie1.DerblevudarbejdetenmanualmedklassifikationerafMR‐fund,hvorefteren

radiolog,enkiropraktorogenhoveduddannelseslægeireumatologiforetoguafhængige

DANSKRESUMÉ

Page|17

vurderingerafkyfose,discushøjde,discuskontur,Modic‐forandringer,spinalstenose,

foraminalstenose,uncovertebralartroseogfacetledsartrose.Inter‐læserreproducerbarhed

blevvurderetparvisogforalletrelæsere,mensintra‐læserreproducerbarhedblev

beregnetforhoveuddannelses‐lægenireumatologi.Prævalenserafpositivefundsamt

uvægtedekappa‐estimaterblevrapporteret.

Studie3varetretrospektivtkohortestudieomfattendealledesygemeldtefraStudie1(n=

168).Etsuccesfuldtarbejds‐outcome(s‐WPS)blevdefineretsom≥75%arbejdsdeltagelsei

uge30‐104efterinklusioneniStudie1.Baselinevariableblevinddeltifiredomæner

(demografiskinformation,anamnestiskinformation,objektivefundsamtMR‐fund).

Univariatlogistiskregressionidentificeredeprognostiskevariable(p<0.2)ogefterfulgtes

afmultivariatlogistiskregression,hvordeprognostiskevariableblevinkluderet

domænevis.Forhverttilføjetdomæneblevsandsynlighedenfors‐WPSdikotomiseret:≥

50%chancefors‐WPSellerej,ogsensitivitet,specificitet,negativogpositivprediktiv

værdi(NPVogPPV)samtarealunderkurven(AUC)blevberegnet.

Resultater

IStudie1varderingenstatistisksignifikantforskelmellemmultidisciplinærogkortinter‐

ventionmht.TTA,smerterogfunktionsevneved1årsfollow‐up.IStudie2fandtes

overvejendesubstantiel(K≥0.61)inter‐læserreproducerbarhed,mensintra‐læser

reproducerbarhedenvarhøjereforalleMR‐fund.IStudie3gavdemografiskoganamnes‐

tiskinformationenPPVpå0.72,ensensitivitetpå0.74ogAUCpå0.76.Hverkensensi‐

tivitet,specificitet,NPV,PPVellerAUCblevændretvedtilføjelseafobjektiveogMR‐fund.

Konklusioner

FørstogfremmestblevhverkenTTA,smerterellerfunktionsevneforbedretefterden

multidisciplinæreinterventionsammenlignetmeddenkorteintervention.Brugenafenkle

klassifikationerafdegenerativefundbidrogtilovervejendesubstantielinter‐læser

reproducerbarhedogsubstantieltilnæstenperfektintra‐læserreproducerbarhed.Endelig

visteStudie3,atobjektivefundogMR‐fundikkeforbedredeforudsigelsenafprognosen

vedr.arbejdsdeltagelsesammenlignetmedprognose‐vurderingudfrademografiskog

anamnestiskinformationalene.

BACKGROUND

Page|18

Background

The burden of neck and shoulder pain

Neckandshoulderpainarecommonconditions.Inthegeneralpopulation,mostestimates

of1‐yearprevalenceforneckpainrangebetween30%and50%(9)whilecorresponding

measuresforshoulderpainrangebetween5%and47%(10).Thelargevariabilityin

prevalenceestimatesispartlyexplainedbydifferencesinstudypopulationsandcase

definitions(9,10).

Whilepainasaconceptisdefinedasanunpleasantsensoryandemotionalexperience

associatedwithactualorpotentialtissuedamage,ordescribedintermsofsuchdamage

(11),therearedifferentandmorespecificdefinitionsofneckpainandshoulderpain

(4,12).Inthisdissertation,neckpainisdefinedaspainperceivedanywhereinthe

posteriorregionofthecervicalspine,fromthesuperiornuchallinetothefirstthoracic

spinousprocess(12).Neckpainmayrefertothehead,theshoulder,theanteriorchest

wall,ortheupperlimb(12).Shoulderpainisperceivedaroundtheshoulderandmay

originatefromtheneck,anyofthethreejointsadjacenttotheshoulder(glenohumeral,

acromioclavicularandsternoclavicularjoints),oradjacentsofttissues(4).

Thecourseofneckpainisoftenepisodic(9)andhighrecurrencerates(50‐75%within1‐5

years)havebeenseenbothinthegeneralandtheworkingpopulations(13,14).For

shoulderpain,ithasbeenreportedthatapproximately77%havesymptomsafter1month

(15),50%after6months(16)and40%after1year(15,16).So,whilebothneckpainand

shoulderpainwillresolvespontaneouslyforsomepeople,itwillpersist,reoccurorworsen

forothers.Forsome,thepainmaycauseworkdisability.

Work disability, sick leave and return to work

Inthisdissertation,thefollowingdefinitionofworkdisabilityisapplied:"Workdisability

occurswhenaworkerisunabletostayatworkorreturntoworkbecauseofaninjuryor

disease"(7).Itisinfluencedbybiopsychosocial,organizational,legislativeandsocietal

factors(17)andmaybedividedintotwocategories:a)temporalworkdisability(sick

leave)andb)permanentworkdisabilitywhichmaybeassociatedwith(part‐time)workor

BACKGROUND

Page|19

dependenceonbenefits(18).Morespecifically,thefocusofthisdissertationisonsick

leave,i.e.asituationwhereaworkerispartlyorfullyabsentfromworkduetoneckor

shoulderpainandnotconsideredpermanentlyworkdisabled.

Returntoworkwillbeusedbothasaconceptandasanoutcome.Asaconcept,the

meaningisliteral,i.e.resumptionoffull‐orpart‐timepaidworkafteraperiodofsickleave.

Whenusedasanoutcome,returntoworkwillbeoperationalisedintherelevantmethods

section.

Sick leave due to neck or shoulder pain

Manypeoplecananddoworkinspiteoftheirpain(19).Forinstance,ithasbeenreported

thattheannualprevalenceofworkersreportingactivitylimitationduetoneckpainranged

between11%(Norway)and14%(Canada)(20).Butforsome,neckpainnotonlycauses

activitylimitation;itisalsoassociatedwithsickleave.InOntario,Canada,theweighted

prevalenceofneckpaininlost‐timeclaimantswas11.3%(21)whilethemostrecent

Danishrecordshowedthatworkerswithneckpainaccountedfor16%ofsickleavedays

(22).

SimilarnumbersforshoulderpainhavenotbeenestimatedinDenmark,butinternational

evidencesuggestsnotableratesofsickleavetobeassociatedwithshoulderpain(23,24).A

Dutchstudyfoundthatamongworkersconsultingtheirgeneralpractitionersfortheir

initialepisodeofshoulderpain,30%reportedsickleaveduringthe6monthsafterthefirst

consultation(23).Incomparison,aDanishstudyofsecondarycarepatientsundergoing

surgeryforsub‐acromialimpingementsyndromefoundthat30%ofthepatientswereon

sickleaveatthetimeofsurgeryand16%receivedsicknessbenefits1yearaftersurgery.

Thissuggestedaconsiderablyhigherdegreeofsickleaveamongpatientswithsub‐

acromialimpingementsyndromecomparedwiththegeneralpopulation,ofwhom4%

wereonsickleaveatthetimeofthestudy(24).

Whilesickleaveisobviouslyachallengetoanyworker,italsoconstitutesaburdento

societyasillustratedbythequantumofsickleave.Itisassociatedwithahigherriskof

futuredisabilitypension(25,26),ariskthatisincreasedwithadditionalnumbersand/or

durationofabsencesduetosickleave,particularlyformusculoskeletaldisorders(27).

BACKGROUND

Page|20

InDenmark,neckpainwasthereasonfor1.5%ofdisabilitypensionsintheyears2010‐

2012(22).Thepercentageattributedtoshoulderpainwasnotestimated,butalarge

population‐basedstudyfoundthatamongpeopleundergoingshouldersurgery,

approximately9%leftthelabourmarketwithinthe2‐yearperiodaftersurgery(28).

Thecostburdenonsocietyishigh:forinstance,forneckpain,theaverageannual

expenditureamountedto917millionDanishkroner(~160millionUSdollars)for

healthcareand2028millionkroner(~353millionUSdollars)forproductivityloss(sick

leaveandprematureretirementintotal)inyears2010‐2012(2012‐equivalentkroner)

(22).HighexpenditureisseeninallOECDcountries(29),e.g.intheUSA,where9%ofthe

totalUSexpenditureonhealthcarewasattributedtoneckandbackpainin2005(30).

Benefits of work

Whilesickleaveisbothanindividualandasocietalburden,returntoworkislikewise

expectedtohavepositiveimplicationsatboththeindividualandthesocietallevels.In

general,havingajobisbetterforhealththannothavingajob(7,19,31),keepinginmind

thatthequalityofthejobandthesocialcontextshouldbetakenintoaccount(19).A

numberofpositiveeffectsofworkhavebeensupportedbyevidence:

Fromapublichealthperspective,workisadeterminantofhealthequalto

education,dietandexercise(7).

Employmentandsocio‐economicstatusarethemaindriversofsocialgradientsin

physicalandmentalhealth(19).

Workcanbetherapeutic(itcanbethepsychosocialvehicleto‘recover’froman

injury/illness(32).

Employmentis(themost)importantmeansofobtainingeconomicresources(19).

Workfulfillsimportantpsychosocialneedsandispivotaltosocialstatusand

identity(19).

Moreover,peoplewithmusculoskeletaldisordershavereportedanumberofmotivators

forreturntowork:apartfromtheevidentfinancialincentives,workalsore‐establishesa

senseofnormality,controlandstructureovertheirlives(33).

BACKGROUND

Page|21

Summingup,sickleaveconstitutesapublichealthproblemand,formostsocieties,there

arestrongincentivesforreducingsickleave.Itistheunderlyingassumptionofthis

dissertationthat‐ifthequalityofthejobandthesocialcontextaretakenintoaccount(19)

‐returntoworkisdesirableforthesicknessabsenteesandforallstakeholdersaffectedby

sickleave.

The arena in work disability prevention (stakeholders)

Sickleavemaybeinitiatedbyaphysicalcondition(e.g.neckorshoulderpain),butthe

courseisofteninfluencedbyanumberofstakeholders.Themultitudeandinterplayof

potentiallyinfluentialfactorsandstakeholderswereillustratedin"Thearenainwork

disabilityprevention"(Figure1(34)).

Theworkerwithaworkdisabilityisinthecentreofthearenawhileeachwingdepicts

influentialsystemsandtheirrepresentativeswhomayallaffecttheprocess.Thebottom

greenwingcomprisespersonalfactorsofthedisabledworker,whiletheredwingcaptures

thedifferentlevelsofaccessiblehealthcareservicesandstakeholdersinthehealthcare

system.Atthetopofthearena,thebluewingillustratesstructurallayersoftheworkplace

systemandtotheright,theyellowwingillustratesinfluentialfactorsandstakeholdersof

thelegislativeandinsurancesystems.Asanimaginaryumbrellainfluencingtheentire

arena,theoverallsocietalcontextwithitscultureandpoliticsisdepicted.

Inbrief,whenassessingthestudypopulationofthisdissertation,sickleavemaybe

triggeredbyneckorshoulderpain(aphysicalfactorofthepersonalsystem),butthe

processofsickleavedependsonaninterplayamongallfoursystemsaswellastheoverall

societalcontext.

Whileacknowledgingthiscomplexityofinfluencesandstakeholders,theoverarching

perspectiveofthisdissertationisthatofthehealthcaresystem(theleftwinginFigure1).

Thehealthcaresystemanditsstakeholdersareinvolvedinmostinterventionsaimedat

returntoworkandsuchinterventionswillbeinvestigated.Sincetheattendingphysician

(theredwingrepresentativenearesttotheworkerinthecentre)isoftenagatekeeperin

sickleavemanagement,thetasksofthephysicianwillalsobeinvestigated.Thesetasks

includeassessmentofcervicalspineMRIandpredictionofworkprognosis.

Interven

Thefoc

painan

intheh

fromth

therape

andthe

theterm

differen

Ifthere

fracture

advicea

approa

ntions

cusofthisd

nddisability

healthcares

heleftwing

euticmoda

enomencla

m'multidis

nt(healthca

earenosym

e,malignan

aboutmain

chwasalso

Figure1

dissertation

youtcomes

systemora

gofthearen

ality(e.g.ex

atureisnot

sciplinary'w

are)profes

mptomsor

ncy),itisre

ntenanceof

ofoundtoi

1.Thearenai

nisoninter

swillalsob

atleastthey

na(Figure

xercises,pa

unanimou

willbeused

ssionalswh

signsofma

ecommend

factivityan

improvew

Page|22

inworkdisab

rventionsa

becovered

yinvolvep

1).Manyin

atienteduca

usinthisfie

d;itdescri

hoofferdis

ajorstructu

edthatclin

ndmoveme

workoutcom

2

ilitypreventio

aimedatim

.Suchinter

participatio

ntervention

ation,orco

eldofresea

besinterve

tinctthera

uralpathol

niciansprov

ent(36).Su

mes(37‐40

on.From(7).

mprovingre

rventionsa

nofhealth

nshaveinc

ognitivebeh

arch(35).In

entionsinv

peuticmod

ogy(e.g.in

videreassu

uchapositi

).

BAC

eturntowo

areoftenem

hcareprofe

cludedmor

haviouralt

nthisdisse

volvingatle

dalities(35

nflammatio

uranceand

ive,reassur

CKGROUND

ork,while

mbedded

ssionals

ethanone

therapy),

ertation,

easttwo

5)).

n,

dpositive

ring

D

BACKGROUND

Page|23

Inadditiontoreassuranceprovidedbyahealthcareprofessional,theinvolvementof

workplacesispivotalininterventionsaimedatreturningpeopletowork.For

musculoskeletaldisorders,moderatequalityevidencesupportsthatworkplace

interventionsimprovetimetofirstreturntoworkandlastingreturntoworkwhen

comparedwithusualcare(41).Thiswassupportedbystrongevidencewhichfound

improvedworkoutcomesafterworkplaceinterventionsinvolvingatleasttwoofthree

components(healthcare,servicecoordination,workaccommodation)(42).Thesame

review(42)foundmoderateevidencethatsuchinterventionsreducedcostsrelatedto

workdisability.

Togetherwithreassuranceandworkplaceinvolvement,thereisstrongevidencetosupport

multidisciplinaryrehabilitationprogrammesimprovingreturntoworkinpeoplewith

persistentmusculoskeletaldisorders(37).Muchoftheresearchhascoveredlowbackpain,

andinchroniclowbackpain,moderateevidencesupportedmultidisciplinarytreatment

beingsuperiortophysiotherapyregardingreturntowork(43).Apositiveeffectonwork

outcomesfollowingmultidisciplinarytreatmentwasalsofoundforback,neckandshoulder

problems(44).Regardingpainanddisabilityoutcomes,multidisciplinarytreatment

outperformedbothphysiotherapyandusualcareinchroniclowbackpain(43).

So,basedontheliterature,reassuranceprovidedbyhealthcareprofessionals,workplace

involvement,andamultidisciplinaryinterventionareexpectedtoimprovereturntowork,

painanddisabilityinsicknessabsenteeswithneckorshoulderpain.Thereis,however,still

nosilverbulletregardingthedesignofsuchamultidisciplinaryintervention,andsickness

absenteeswithneckorshoulderpainremainlessdescribedintheliteraturecompared

withtheirpeerssufferingfromlowbackpain.

Assessment of cervical spine MRI

Akeyresponsibilityoftheattendingphysicianistoruleouttheexistenceofmajor

structuralorotherspecificpathologies(36).Thisisprimarilydoneonthebasisof

symptomsandsigns(i.e.listeningtothepatient'ssymptomsandperformingaclinical

examination).Inthecaseofneckorshoulderpain,examplesofsuchspecificpathologies

includefracture,inflammation,radiculopathy,myelopathy,androtatorcuffdisorder.

BACKGROUND

Page|24

Accordingtoguidelines,routineimagingisnotrecommendedforneckandshoulderpain,

andimagingshouldonlybeconsideredifithasthepotentialtoaltermanagementand

improveoutcomesofmanagement(45,46).Advancedimaging(e.g.magneticresonance

imaging(MRI))shouldbeconsideredincasesofnoimprovementordeteriorationof

neurologicalstatus,ifthereispersistentseverepainanddisability,andifmalignancyor

othermajorstructuraldiseaseissuspected(45,46).

Despitetheseclinicalguidelinesandarelativelystableprevalenceofneckandshoulder

paininDenmark,thenumberofcervicalspineMRIsperformedhasincreased18%in

Denmarkoverrecentyears(47)comparedtoonly4%increaseintheprevalenceofneck

andshoulderpain(48,49).IncreaseinthenumberofMRIsperformedhasalsobeenseen

elsewhere,e.g.inCanada(50,51).

ThisincreasemaybeattributedtoimprovedaccesstoMRI,alongwithbothpatientsand

physiciansunderstandingthepotentialbenefitsofMRI.PatientsbelieveMRImightreveal

causesfortheirpain(52)whilephysiciansvaluethenon‐invasiveness,theabsenceof

ionisingradiation,andthecapacitytodiscriminatebetweensofttissues(53).Inspecific

casesofneckorshoulderpain,MRImayalsoberequestedwhentheclinicalevaluation

leavesdoubtastowhetherthepainoriginatesfromtheneckorfromtheshoulder,a

discriminationwhichissometimesdifficultbasedonsymptomsandsignsalone(4,54).

ForMRI,asforalltypesofimaging,therearetwomaintasksrelatedtoitsuse:1)

Interpretationoftheimagesand2)Communicationofthefindings–andbothoftheseare

required(55).ToachieveconsistencyinthecommunicationaboutMRIfindings,those

requestingandthoseinterpretingMRIsneedtoagreeaboutboththeinterpretationof

imagesandaboutthelanguageappliedtodescribetheseimages.Theprerequisiteforthis

isreliabilitywhichisdefinedas"theextenttowhichscoresforpatientswhohavenot

changedarethesameforrepeatedmeasurementunderseveralconditions"(3).

AnumberofreliabilitystudiesoncervicalspineMRIhavefoundvaryingdegreesofinter‐

raterreliabilityfordifferentfindings(56‐60).Intra‐raterreliabilityestimateswere

generallyhigherthanthoseforinter‐raterreliability(56‐60).Moststudieshowever,only

assessedreliabilityforafewdegenerativefindingsorcomparedreaderswithsimilar

educationandexperience.

BACKGROUND

Page|25

Onlyonereliabilitystudywasidentifiedwhichassessedseveraldegenerativefindings(61).

Therefore,furtherstudiesareneeded.Moreover,itwasnecessarytodeterminetheintra‐

andinter‐raterreliabilityofcervicalspineMRIassessmentswithinthisdissertation,since

thiswasaprerequisitetoassesscervicalspineMRIsintheprognosticstudy.

Prognosis

Knowledgeaboutworkprognosiscanguideexpectationsforallinvolvedstakeholdersin

theprocessofsickleave.Inneckpain,itisrecommendedthatcliniciansassessprognostic

factorsfordelayedrecovery(36).Andalthoughsomephysiciansexperienceconflicting

demandsfromdifferentstakeholders(62),theassessmentofworkprognosisisa

mandatorytaskwhenrequestedbyauthoritieswhocoordinatereturntoworkand

reimbursesicknessbenefits.Sincecorrectassessmentofworkprognosisnotonlyguides

expectationsbutmayalsohavethepotentialtoidentifyindividualswhoneedspecial

interventionsorcare,thereisaneedforthephysiciansresponsibletobeaswell‐informed

aspossiblewhenmakingtheirassessments.

Knowledgeaboutfactorsinfluencingworkprognosishasemergedoverrecentdecadesof

research.Someexamplesofparametersthatarenegativelyassociatedwithworkoutcomes

include:increasingage(13,64,99,106),psychosocialparameters(e.g.poorrecovery

expectationsorperceivedsocialsupport(63)),increasingsickleaveduration

(13,14,44,64,65),work‐relatedfactors(e.g.lowjobcontrolandhighjobdemands(62))and

societal/culturalfactors(e.g.thegenerosityofthecompensationbenefits(62)).

Butmanyphysicianstypicallycategorisetherangeofavailableinformationdifferently.The

orderofinformationobtainedinaclinicalsettingistypicallyasfollows:1)demographics

(sexandage)arealreadyknownbeforetheclinicalencounterduringwhich2)patient‐

reportedinformationisrecorded.Basedonthis,3)afocusedclinicalexaminationis

performedandinsomecases4)imaging(e.g.MRI)isrequested.Basedonthegathered

information,thephysicianassessesworkprognosiswhenrequested.However,especially

forMRIinformation,evidenceabouttheprognosticvalueforworkoutcomesisscarce.

Despitetheampleliteratureondeterminantsofworkprognosis,nostudieswereidentified

thataddressedtheorderofavailableinformationbyexploringthecontributionof

BACKGROUND

Page|26

demographics,patient‐reportedinformation,clinicalinformationandMRIinformationto

thepredictionofworkprognosis.

AIMS

Page|27

Aims

Thisdissertationseekstocontributenewknowledgeaboutinterventions,MRIassessment

andprognosisinsicknessabsenteeswithneckorshoulderpainseeninsecondarycare.In

doingso,theoverarchingperspectivewasthatofthehealthcaresystemand/orthe

attendingphysiciansincethesearekeystakeholdersintheprocessofsickleave.Applying

thisperspective,theoverallaimwastoevaluatearangeoftasksperformedwithinthe

healthcaresystem,namelyinterventionsaimedatreturntowork,reliabilityofMRI

assessmentsandpredictionofworkparticipation.

Thespecificaimswere:

Aim 1

Toevaluatetheeffectofamultidisciplinaryinterventioncomparedwithabrief

interventionwithrespecttoreturntowork,painanddisabilityinworkersonsickleave

duetoneckorshoulderpain.

Aim 2

Todeterminetheinter‐andintra‐raterassessmentreliabilityofdegenerativefindings

(kyphosis,discheight,disccontour,vertebralendplatesignalchanges,spinalcanal

stenosis,neuralforaminalstenosis,uncovertebralosteoarthritisandzygapophyseal

osteoarthritis)onMRIofthecervicalspine.

Aim 3

Toexplorethedegreetowhichdemographic,patient‐reported,clinical,andMRI

informationcontributetothepredictionofworkparticipationinsicknessabsenteeswith

neckorshoulderpain.

METHODS

Page|28

Methods

Table1presentsanoverviewofthedesigns,samples,outcomevariablesandanalysesin

thisdissertation.Theappliedmethodswereinterwovenwiththechronologyofthestudies.

Forthisreason,andforthepurposeofclarity,thestudieswillbepresentedaccordingto

theirchronologyfollowedbyadescriptionofthesocietalandlegislativeconditionsofthe

studies,andtheirethicsapproval.

Table1.Overviewofstudies,designs,samplesizes,outcomevariablesandstatisticalanalyses

Study Design N Outcomevariables Statisticalanalyses

1 RCT 168 Primaryoutcome:4weeksRTWSecondaryoutcomes:painimprovement≥2points(NRS)anddisability(CNFDS)

CoxproportionalhazardsregressionanalysisSurvivalanalysis(Kaplan‐Meier)LogisticregressionanalysisLinearregressionanalysis

2 Reliabilitystudy 50 PrevalenceUn‐weightedkappaestimates

Un‐weightedkappastatistics

3 Prognosticstudy(retrospectivecohort)

168 Workparticipationscore≥75%inweeks30‐104afterbaseline(s‐WPS)SensitivitySpecificityPPVNPVAUC

CrudeandmultivariablelogisticregressionanalysesCalculationsofsensitivity,specificity,PPV,NPVandAUC

AUC:AreaUndertheCurveCNFDS:CopenhagenNeckFunctionalDisabilityScaleNPV:NegativePredictiveValueNRS:NumericratingscalePPV:PositivePredictiveValue

RTW:ReturnToWorkRCT:RandomizedControlledTrials‐WPS:successfulWorkParticipationScoreu‐WPS:unsuccessfulWorkParticipationScore

METHODS

Page|29

Study 1, the RCT

Design,settingandparticipants

FromMay2009toJanuary2014,participantswereincludedinarandomizedcontrolled

trialatTheSpineCentre,SilkeborgRegionalHospital,Denmark.Participantswere

recruitedfromprimarycaresettingsinsevenmunicipalitiescollaboratingwithTheSpine

Centre.Inthesemunicipalities,writteninformationabouttheRCTwasdisplayedinthe

waitingroomsofchiropractors,physiotherapistsandgeneralpractitioners(GPs).TheGPs

werefurtherencouragedtoreferpatientsfulfillingtheinclusioncriteria:age18‐60years,

4‐16weeksofsickleave(part‐timeorfull‐time)owingtoneckorshoulderpain,and

fluencyinDanish.Exclusioncriteriawere:drugaddiction,primarypsychiatricdisorder,

recentspinesurgery,plansforspinesurgery,pregnancy,andinflammatory/otherspecific

musculoskeletaldisorder.

Figure2presentstheflowchartforStudy1.Overall,328patientswerereferredand

screenedforeligibility.Ofthese,160couldnotbeincluded,primarilybecausethesick

leavecriterioncouldnotbefulfilled(n=120).Intotal,168participantswereincludedand

randomlyallocatedtobriefintervention(BI)(n=83)ormultidisciplinaryintervention

(MDI)(n=85).TheyallcompletedtheRCT.

METHODS

Page|30

Figure2FlowchartStudy1

METHODS

Page|31

Theclinicalinterventionforallparticipants

AnoverviewofeventsandcontactswithTheSpineCentreispresentedinTable2.Attheir

firstvisittoTheSpineCentre,alleligibleindividualswereinvitedtoparticipateinthe

study.Iftheyagreed,writteninformedconsentwasobtained.Afterthis,theywereall

examinedbyarheumatologistandaphysiotherapistbothofwhomwerededicatedspine

cliniciansandblindedtothesubsequentrandomization(whichtookplace2weekslater,

seeTable2).Themedicalhistorywasrecordedbytherheumatologist,whoalsoperformed

athoroughclinicalexamination.Inareassuringway,therheumatologistexplainedthe

usuallylimitedcorrelationbetweenpainanddegenerativespinalMRIfindings.

Pharmacologicaltreatmentwasadjustedifneededandlaboratorytestswererequested

whenclinicallyrelevant.Iftheclinicalexaminationrevealedsignsofprimaryshoulder

pathology,ultrasonographywasperformedandincaseofinflammation,asteroidinjection

wasoffered(n=2,oneineachgroup)(66).Iftheclinicalexaminationrevealedsignsof

radiculopathy,participantswereinformedaboutthelikelypositiveprognosisandabout

possiblereferraltoaspinesurgeonifimprovementdidnotoccur.

Table2.ContactswiththeSpineCentre

Baseline:

Clinical

examination

andadvice

2weeks:

Physiotherapist

followupand

randomization

3‐4weeks:

1stmeeting

withcase

manager

3‐6weeks:

Information

aboutMRI

findings

12weeks:

Followup

atthephysio‐

therapist

RTWplan

andregular

meetings

withcase

manager

MDI

group+ + + + + +

BI

group+ + ‐ + + ‐

BI:BriefIntervention

MDI:MultidisciplinaryIntervention

MRI:MagneticResonanceImaging

RTW:ReturnToWork

METHODS

Page|32

Thephysiotherapistperformedastandardizedclinicalexamination.Neuromusculartesting

wasperformedandisometricneckstrengthwasmeasuredexceptinparticipantswith

radiculopathy.Inthiscase,testswereperformedusingtheMcKenziemethodwhichis

widelyusedinneckpainalthoughonlysupportedbymoderateevidenceforlowbackpain

(67).Bothcliniciansencouragedthemaintenanceofeverydayactivities,workandexercise.

Atthefollow‐upvisitafter3‐6weeks,therheumatologistexplainedtheMRIfindingsstill

usingareassuringapproach.Thelastfollow‐upvisitwiththephysiotherapisttookplace

after12weeks.Aneffortwasmadetoensurecoordinationbetweenallrelevant

stakeholdersbysendingcopiesofmedicalrecordswithminimaldelay(fewdays).

Thebriefintervention(BI)

Apartfromtheclinicalinterventiondescribedabove,thoseintheBIgroupreceivedno

furthertreatment.Itwasrecommendedthattheyreturntoworkwhenpossible,andifthey

neededfurthertreatmentoradvice,theywereadvisedtocontacttheirGP.

Themultidisciplinaryintervention(MDI)

Inadditiontotheclinicalintervention,thoseintheMDIgroupwerescheduledformeetings

withindividualcasemanagers.Incollaborationwiththese,thesicknessabsenteesmade

individualRTWplans.Thecasemanagersheldregularmeetingswithamultidisciplinary

team(anoccupationaltherapist,therheumatologist,thephysiotherapist,asocialworker,a

specialistinoccupationalmedicineand,inrelevantcases,apsychologist).Theseteam

memberswereengagedintheRTWprocesswhenrelevant:12participantshad

consultationswiththepsychologistwhilethedatadonotdescribetheinvolvementofthe

otherteammembers.Themultidisciplinaryteamhad1‐2hoursofsupervisionwith2‐

monthintervalstoensurethelargestpossibledegreeofstandardizationinthegivenMDI.

Meetingsbetweentheparticipantsandtheircasemanagerswereheldasoftenasneeded.

Round‐tablediscussionswerearrangedattheworkplacefor19participantsandinthree

morecases,thecasemanagertalkedtotheemployersviatelephone.Itwasuptothe

sicknessabsenteestodecidewhetherworkplacesshouldbeinvolvedintheRTWprocess,

somethingthatmanyofthemwerenotinclinedtodo.WhenRTWwasachieved,thecases

wereclosed.IfRTWwasconsideredunrealistic,alternativeplansweremade,e.g.totake

METHODS

Page|33

upjobssupportedbythesocialsystem.Ifneitherregularworknorjobssupportedbythe

socialsystemwereachievable,thecasewaspassedontotheresponsiblesocialworkerin

themunicipalityofthesicknessabsentee.

NestedRCT

NestedintheRCTwasasmallerRCTcomparingtwodifferentexerciseprograms(68).

Eighty‐threeoftheparticipantswithnonspecificneckpainwereincludedinthenested

RCTandrandomlyallocatedtooneoftwohome‐basedexercisegroups:ageneralphysical

activitygroup(GPAgroup,n=40)andagroupdoingbothgeneralphysicalexerciseand

specificstrengthtraining(SSTgroup,n=43).Nodifferencesinpainintensity(theprimary

outcome)werefoundbetweenthegroups.

Baselinevariables

Baselinevariableswerecollectedfromcomprehensivequestionnairesandfromtheclinical

examinationperformedbytherheumatologistpriortorandomization.

Thequestionnaireswerefilledinbythestudyparticipantspriortotheirfirstmeetingwith

thecliniciansandpriortorandomization.Thesequestionnairescovered

sociodemographics,education('noeducation','briefcourses','skilledworker','education<3

years','education3‐4years,'education>4years','other'),painintensity(11‐pointNumeric

RatingScale)(69),currentsmokingstatus(yes,no)anddisability(CopenhagenNeck

FunctionalDisabilityScale(CNFDS)(70)forthosewithprimarilyneckcomplaintsand

DisabilitiesoftheArm,ShoulderandHand(DASH)forthosewithprimarilyshoulder

complaints(71)).ItfurthercoveredtheÖrebroMusculoskeletalPainQuestionnaire

(ÖMPSQ)(72).TheÖMPSQwasdevelopedasascreeningtoolaimedatidentifying"yellow

flags"(e.g.emotionalstate,copingstrategies,fearavoidancebeliefs,andexpectationsof

returntowork).Thatis,itwasdevelopedasatoolforidentificationofpatientswith

psychosocialriskfactorsforprolongedsickleave(73).

Thebaselinequestionnairealsocoveredhealth‐relatedqualityoflife(Short‐Form36),

musculoskeletalcomorbidity(lowbackpain,legpain),sickleavedurationpriortostudy

enrollment(weeks)(13,14,44,64,65),numberofpreviousabsencesduetosickleave,a

questionaboutworkbeingthecauseforpain('yescompletely','yesbutotherissuesalso

METHODS

Page|34

contributetopain','nothepainiscausedbyotherthings','noIdonotknowthereasonformy

pain'),andpart‐timesickleave('yes','no').

Furtherbaselinevariableswerecollectedfromtheclinicalexaminationperformedbythe

rheumatologist.Thediagnosisradiculopathydenotedaclinicalassessmentbasedon1)

radiatingpainfromthenecktotheupperextremityand2)oneormorepositiveclinical

findings:diminisheddeeptendonreflexes,decreasedmusclestrength,dermatomalsensory

deficitsorpositiveforaminalcompressiontest.Thediagnosisprimaryshoulderdisorder

wasnotbasedonpredefinedcriteriabutontherheumatologist's>30yearsofclinical

experience.Thediagnosiswasappliedwhena)bothsymptomsandsignsunambiguously

correspondedwithshoulderpathologyandb)thisclinicalsuspicionwassupportedby

imaging(ultrasonographyorMRIoftheshoulder).Non‐specificneckpainwasapplied

whenstudyparticipantsreportedneckpainandtheclinicalexaminationrevealednosigns

ofspecificunderlyingpathology(e.g.radiculopathy).

Outcomes

TheprimaryoutcomeofRTWwasdefinedas4consecutiveweeksoffinancialself‐support

exceptforstudyparticipantswhoheldjobssupportedbythesocialsystempriortotheir

sickleave.Inthesecases,4consecutiveweeksofhavingreturnedtothesejobswas

regardedasRTW.DataweresuppliedbytheDanishRegisterforEvaluationof

Marginalization(DREAM).DREAMhasinformationaboutthesourceofincomeforalladult

Danishcitizenswhohavereceivedtransferbenefitsatsomepointsince1991.Weekly

recordingsaremadedescribingwhetheragivenindividualisself‐supportingorreceiving

somekindoftransferbenefits.Thesebenefitsareregisteredbymeansof3‐digitcodesand

orderedhierarchicallywhichimpliesthatiftwodifferentbenefitsarereceivedinthesame

week,thehighest‐rankingcodewilloverwritetheother.Sicknessbenefitshaveahigh

priorityandtheseareonlyoverwrittenbyemigration,agepensions,anddeath(74).Ifno

transferbenefitsarereceived,theindividualisregardedasbeingself‐supporting.Forthe

purposeofthisthesis,self‐supportisregardedasequaltohavingapaidjob.

METHODS

Page|35

Thesecondaryoutcomeswere:

self‐reportedchangesofpainintensity(NumericalRatingScale)forallparticipants.

Thiswascomputedbysubtractingthe1‐yearpainscorefromthebaselinepain

score.

self‐reportedmeasuresofdisabilityasmeasuredbyCNFDS(fornon‐specificneck

painandradiculopathy)andDASH(onlyforprimaryshoulderdisorder).

Dataonthesesecondaryoutcomeswerecollectedbysendingpostalquestionnairestothe

participants1yearafterenrolment.Inthecasesofnoreply,areminderletterwassentand

afinalattempttoachievethefollow‐upquestionnaireswasdonebymeansofatelephone

call.Sincefollow‐updataonDASHwerelargelyincomplete,thisoutcomewasomitted(out

of21participantswithprimaryshoulderdisorder,onlyninerespondedofwhomonewas

intheMDIgroupandeightintheBIgroup).

Analyses

ApowercalculationwascarriedoutpriortotheRCT.Assumingthattherewouldbea15%

differenceinRTWbetweenthegroups,apower(1–β)of70%andasignificancelevelof

5%(α=0.05),astudysamplewith85participantsineachgroupwasrequired.

Baselinevariableswereanalyzedafterexclusionofmissingvaluesusingdescriptive

statistics.Ifdatawerenotnormallydistributed,medianvaluesandinterquartileranges

(IQR)werereported.

Fortheprimaryoutcome,survivalanalysiswasusedtoestimatethetimetoRTW.Ratesof

RTWinthetwointerventiongroupswerecomparedusingCoxproportionalhazards

regressionanalysisafterconfirmingtheassumptionofproportionalhazardsbylog‐minus‐

logplots(resultsnotshown).Deathandemigrationweredefinedascompetingrisks,and

theintention‐to‐treatprinciplewasapplied.Followingthecrudeanalyses,adjustments

weremadeforsex,age,sickleaveduration,clinicaldiagnosesandpart‐timesickleave.

Forthesecondaryoutcomeofpain,theassumptionsforlinearregressionanalysiswerenot

met.Andsinceaminimallyclinicallyimportantchange(MCIC)wasdesired,thiswas

definedasMCIC≥2points(yes/no)(75,76)andlogisticregressionanalysisapplied.

METHODS

Page|36

Crudeoddsratiosandoddsratiosadjustedforsex,agegroupsandbaselinepainintensity

werecalculated.

Forthesecondaryoutcomeofdisability(CNFDS),thedifferencebetweenintervention

groupsat1‐yearfollowupwasestimatedasthedifferenceinmeanCNFDS.Crude

estimatesandestimatesadjustedforsex,agegroupsandbaselineCNFDSscorewere

calculated.Themodelwascheckedbydiagnosticplotsoftheresiduals(notshown).

Finally,owingtoalargenumberofnon‐respondersonthesecondaryoutcomes(n=89),an

analysisofrespondersvs.non‐responderswasperformed.Differenttestswereused

dependingonthetypeanddistributionofthevariable:anunpairedTtest,Chisquaredtext,

Fisher'sexacttestorWilcoxonrank‐sumtest.

ExceptforthepowercalculationwhichwasperformedpriortotheRCT,allstatistical

analyseswereperformedbytheauthorofthisPhDdissertationwhowasnotinvolvedin

theclinicalinterventionsgiven.

Study 2, the reliability study

Sampleanddatacollection

Inagreementwithrecommendationsforreliabilitystudies(77),50MRIswereconsidered

areasonablesamplesizeforthisstudy.Thepredefinedinclusioncriteriawere:availability

ofcervicalspineMRIandasatisfactorysignal‐to‐noiseratioasjudgedbyanexperienced

radiologist.Thisevaluationwasbasedontheradiologist's25yearsofclinicalMRI

experience.Asshownintheflowchart(Figure3),78wereexcluded,leaving90cervical

spineMRIs.Fromalistofthese90,alternateMRIswerechosen,toensureasrandoma

selectionaspossible.

Theimagesweresuppliedbyfivedifferenthospitals.SagittalT1‐weightedandT2‐

weightedsequenceswereavailableforallMRIsandaxialT2‐weightedsequencesfor94%

oftheimages.Inaddition,82%oftheMRIshadobliqueT2‐weightedsequencesavailable.

METHODS

Page|37

Figure3FlowchartofMRIsusedinStudies2and3

Readers

ThreereadersassessedtheMRIs.ReaderA(theauthorofthisdissertation)wasamedical

doctorwith9yearsofpostgraduateclinicalexperienceincludingexperienceinspinalMRI

assessmentforclinicalpurposes.ShehadnopreviousformaleducationinMRI.ReaderB

wasaradiologistwith25yearsofclinicalexperienceinassessingMRI.ReaderCwasa

METHODS

Page|38

chiropractorholding10yearsofclinicalandacademicexperienceinspinalMRI.In

addition,hehadcompleteda1‐yearfull‐timeinternshipinspinalMRI.

Beforeperformingthestudy,readerAreceived2hoursoftuitionfromreaderBcovering

cervicalspineMRIassessment.Afterthis,sheevaluated50cervicalspineMRIsfrom

patientswithneckpain+/‐radiculopathyandcompletedclinicalnarrativereportsofthese.

ReaderBassessedthesameimages,correctedthereportsifnecessaryandexplained

relevantissuestoreaderA.Theimagesforthesenarrativereportswerenotpartofthe

reliabilitystudy.

Whileblindedtodemographicandclinicaldataaswellaspreviousassessments,readersA,

BandCindependentlyassessedtheMRIsofthefullsample.Thistookplaceovera

timeframeof5‐8weeks.ReaderAevaluatedtheimagestwice,andtopreventrecollection

ofassessments,thesecondevaluationwasperformed6weeksafterthefirstone.

Evaluationmanual,pilotingandworkstations

Toassistthereaders,anevaluationmanual(Appendix4)withwrittenandvisual

classificationsofMRIfindingswasdraftedbyReaderA,followedbyadjustmentsand

approvalbyReadersBandC.Followingthis,10MRIsfromthestudysamplewere

independentlyassessedbythereaders.Twoconsensusmeetingswereheld,withboththe

classificationsintheevaluationmanualandthepracticeofthereadersadjustedduringthis

pilotingphase.Usingradiologicalworkstations,theMRIswereassessedusingVitreaCore

(version1.0.0.404,vitalImagesInc.).

Variables

Basedontheliterature(59‐61,78‐88)andthepiloting,classificationsforcommonand

degenerativeMRIfindingsweredeveloped.TheincludedMRIfindingsarepresentedin

Table3,andasshown,theintentionwastocreatesimpleandclinicallyapplicable

definitions(61).Fromtheseclassifications,categorical(notordinal)datawereobtained.

METHODS

Page|39

Table3.MRIfindingsandcorrespondingclassifications

MRIfinding Category Description

Kyphosis 0 Normalorstraightenedlordosis 1 Kyphosis Discheight 0 Normal 1 Reducedheight Disccontour 0 Normal 1 Bulgeorprotrusion 2 Extrusion Spinalcanalstenosis 0 Normal 1 >50%obliterationofCSF,nocorddeformity 2 >50%obliterationofCSFwithcorddeformitybut

nosignalchange. 3 >50%obliterationofCSFwithcorddeformityand

signalchange Vertebralendplatesignalchange 0 Normal 1 Type1 2 Type2 3 Type3 4 Mixedtype1and2 Uncovertebralosteoarthritis 0 Normal 1 Definiteosteoarthritis Zygapophysealosteoarthritis 0 Normal 1 Definiteosteoarthritis Neuralforaminalstenosis 0 Normalor<50%fatobliteration 1 ≥50%fatobliterationwithorwithout

morphologicalchangesofthenerveroot

CSF:cerebrospinalfluid

Analyses

Beforethereliabilitycalculations,theprevalenceofpositivefindingsforallthreereaders

wasassessed.Thistabulationofprevalenceallowedfor1)assessmentofsample

homogeneityand2)identificationofpossiblesystematicdifferencesbetweenthereaders.

Boththeseissuesmayaffectreliabilityestimates(77,89).Thetabulationalsosuppliedthe

METHODS

Page|40

estimatesforobservedagreement(OA)andagreementbychance(AC)thatwereusedin

thepairwiseanalysesofreliability.Forthethree‐readerreliabilityanalysis,OAwas

computedbycalculatingthenumberofobservationswithcompleteagreementand

dividingthisnumberwiththenumberofassessedanatomicalsites.Thethree‐readerAC

wascomputedbymultiplyingthemarginalfractions(77).

Forthereliabilityanalyses,un‐weightedkappa(Κ)statisticswereusedduetothe

categoricalnatureofthedata.Undertheconditionoftotalindependenceamongthe

readers,Κisdefinedas

Κ=

Theanalyseswereperformedpairwise(A1B1,A1C1,B1C1andA1A2)andforallthree

readers(A1B1C1).BearinginmindtheinfluenceofprevalenceonΚ(77,89),reliabilitywas

onlycomputedforvariableswherethereadersagreedonaprevalenceof≥10%.The

nomenclaturesuggestedbyLandis&Koch(90)wasusedforinterpretationoftheΚ

estimates:

Κvalue Strengthofagreement

<0.0 Poor

0.0‐0.2 Slight

0.21‐0.4 Fair

0.41‐0.6 Moderate

0.61‐0.8 Substantial

0.81‐1.0 Almostperfect

ResultswerereportedasΚestimateswith95%confidenceintervalsalongwithestimates

ofobservedagreementandagreementbychanceforallfindings.

METHODS

Page|41

Study 3, the prognostic study

Designandparticipants

Aretrospectivecohortdesignwasusedincludingallthesicknessabsenteeswhowere

enrolledinStudy1(n=168).

Outcomevariable:Workparticipationscore(WPS)

LikeinStudy1,dataonsourceofincomeweresuppliedbyDREAM.Inanattemptto

capturebothreintegrationandmaintenanceofwork(91),theworkparticipationscore

(WPS)waschosenastheoutcome.TheWPSyieldsscoresbetween0%and100%:the

numeratorcomprisesthenumberofworkingweeksandthedenominatorthenumberof

follow‐upweeks(8).Weeksofself‐supportinDREAMwereconsideredasworkingweeks

andforthisstudy,weekswithstateeducationfundgrantsandrelatedbenefitswerealso

includedsincethesetransferincomeswereinterpretedasbeingfitforwork.Forthree

individualswhoweregrantedearlyretirement,thedenominatorcomprisedonlythe

numberofweeksbeforethefirstweekofretirementbenefits.

Thefollow‐upperiodcoveredweeks30‐104sincethemediantimeuntilthefirstRTW(4

consecutiveweeksofself‐support)was29weeks.Asuccessfulworkoutcomewasdefined

asWPS≥75%fortworeasons:1)thisthresholdwaspreviouslyfoundmoderatelyto

substantiallyreliablecomparedwithothercommonRTWmeasures(8)and2)itdefinedan

improvementinWPScomparedtothemeanWPStheyearbeforeenrolment(74%).This

yieldedadichotomousoutcomeforwhichthefollowingnomenclaturewillbeused:

SuccessfulWPS(s‐WPS)denotesWPS≥75%

UnsuccessfulWPS(u‐WPS)denotesWPS<75%

Prognosticvariables

Dataondemographicandpatient‐reportedinformationwereattainedfromthe

questionnairesdescribedunderStudy1.Thesedatacoveredsex,age,education,pain

intensity(11‐pointNumericRatingScale)(69),numberofsickleaveepisodes,durationof

sickleavepriortoenrollment(weeks)(13,14,44,64,65),whiplashtrauma(92),

METHODS

Page|42

psychosocialdimensionsofpain(ÖMPSQscore)(72),andworker'scompensationclaim

(7).

Clinicalvariableswereavailablefromtherheumatologist'sclinicalexamination.Tender

pointswereexaminedbyapplyingapressureofthethumb~4kgon18definedlocations

onthebody1.Thereliabilityoftherheumatologist'stenderpointsexaminationhasbeen

describedasexcellent(93)andinaccordancewithcommonpracticeamongclinicians,a

thresholdof≥11TPswaschosen(94).Radiculopathywasconsideredpresentinsickness

absenteeswho1)experiencedradiatingpainfromthenecktotheupperextremityand2)

hadoneormoreofthefollowingsigns:positiveforaminalcompressiontest,decreased

musclestrength,diminisheddeeptendonreflexes,orsensorydeficitsfollowingthepattern

ofadermatome.

Exceptforvertebralendplatesignalchanges(VESC),theMRIvariableslistedinTable3

wereassessedondisclevelsC2/C3toC7/T1.Theauthorofthisdissertation,anMDwith9

yearsofclinicalexperience,performedtheassessmentsofthe97availableMRIsas

illustratedinFigure3.Theinter‐raterreliabilityoftheseassessmentswassubstantialfor

allvariablesexceptzygapophysealosteoarthritisforwhichmoderatereliabilitywasfound.

Theintra‐raterreliabilitywassubstantialtoalmostperfect(Study2).Sincethereliability

ofVESCwasnotestablishedinStudy2,thisfindingwasnotassessed.AnMRIfindingwas

definedaspresentifitoccurredat≥1disclevel,i.e.ifspinalcanalstenosiswaspresenton

twodisclevels,itwascountedonlyonce.

Analyses

ForallvariablesexpectedtoinfluenceWPSprognosis,crudelogisticregressionanalyses

wereperformedandvariablesyieldingp<0.2werecarriedforward.Inthemultivariable

analyses,variableswereincludedusinganapproachresemblingtheorderofobtained

1Bilaterallocationsof:suboccipitalmuscleinsertions,anterioraspectsofintertransversespacesC5‐C7,

medialaspectofthesupraspinatusmuscle,midupperborderoftrapeziusmuscle,secondcostochondral

junction,2cmdistaltothelateralepicondyleoftheelbow,upperouterquadrantofthebuttock,thespot

posteriortothetrochantericprominence,andmedialfatpadoftheknee.

METHODS

Page|43

informationinaclinicalsetting.Themultivariableanalyseswereperformedstepwise

whichresultedinfourmodelsincludingthefollowingdomainsofinformation:

Model1:demographics

Model2:demographicandpatient‐reportedinformation

Model3:demographic,patient‐reportedandclinicalinformation

Model4:demographic,patient‐reported,clinicalandMRIinformation

Foreachmodel,theprobabilityofs‐WPSwascalculatedanddichotomized,sothatthe

sicknessabsenteeswereclassifiedashaving≥50%chanceofs‐WPSor<50%chanceofs‐

WPS.BytabulationofestimatedchanceandactualWPSoutcomes,calculationof

sensitivity,specificity,positiveandnegativepredictivevalues(PPVandNPV)andarea

undercurve,AUC,wasperformed.

Sensitivitydenotestheproportionofsicknessabsenteeswithu‐WPSwhowereclassified

ashaving<50%chanceofs‐WPS.Specificityistheproportionofsicknessabsentees

achievings‐WPSwhowerecorrectlyclassifiedashaving≥50%chance.PPVisthe

proportionofthoseclassifiedashaving<50%chancewhoexperiencedu‐WPSandNPV

theproportionofsicknessabsenteesclassifiedashaving≥50%chancewhoachieveds‐

WPS.

Forthepurposeofbestpossiblecomparabilitybetweenthemodels,estimatesweregiven

bothforthenumberofindividualswithcompletedatainagivenmodelandforthenumber

ofindividualswithcompletedataintheadjacentmodel.Forinstance,inModel1,estimates

weregivenforall168individualswithavailabledemographicinformationandforthe161

individualswithavailablepatient‐reportedinformationintheadjacentModel2.

Inordertoappraisethepropertiesofthemodels,andwiththeparticularaimofassessing

whetheraddingdomainsofinformationimprovedthepredictionforthosewithhighest

andlowestprobabilitiesofs‐WPS,furthertabulationsweremade.Thesecomparedactual

WPSoutcomestoestimatedchance,andinthesetables,chanceofs‐WPSwasdividedinto

fourcategories:chance<30%,chance30‐49%,chance50‐69%,andchance≥70%.

METHODS

Page|44

Nomenclature

InStudy3,theterm'prognosticvariable'wasusedforbaselinevariableswhichwere

(positivelyornegatively)associatedwiththeoutcomes‐WPS.Noinferencesweremade

aboutcausality(2).Theterms'predict'/'prediction'relatetotheabilityofthemodelsto

forecasttheprognosisregardingworkparticipation.

Context: The Danish Labour Market and The Danish Sickness Benefit Act

TheconductofStudy1andinterpretationofStudies1and3shouldbeseeninthelightof

societalandlegislativefactors(17).TheDanishlabourmarketischaracterizedbytheso‐

called'flexicuritymodel'whichimpliesalargedegreeofflexibilitywithinthelabourforce:

employersareallowedtofireandhiredependingonthegivenproductionneeds.Atthe

sametime,unemployedworkersareentitledtofinancialsecuritybymeansof

unemploymentbenefits.Finally,anactiveemploymentpolicy(95)commitstheauthorities

toofferunemployedworkersactivitiesandguidancerelevanttotheirfutureemployment.

Thisresponsibilityisheldbythemunicipalities.

Ifsickleaveoccurs,theDanishSicknessBenefitActseekstorecognizeneedsand

challengesofboththesicknessabsenteesandtheemployers(96).Sicknessabsenteesare

entitledtosicknessbenefitsiffulfillingcertaincriteriaregardinglabourmarketaffiliation

(thesecriteriavaryamongwageearners,self‐employed,andunemployed)(96).Thisright

appliesirrespectiveofthereasonforsickleave,beingwork‐relatedornot.Duringthetime

oftheRCT,therightforsicknessbenefitslasted52weeks.Onlysixparticipants(MDI:n=3

andBI:n=3)wereaffectedbyalegislativechangewhichreducedtherightforsickness

benefitsfrom52to22weeks.Belowarelistedtherightsandobligationsofthesickness

absentees,theemployersandthemunicipalities.

Wherespecifictimeperiodsaregiven,theseappliedduringthetimeofStudy1running

from2009‐2014:

Asicknessabsenteehasnoobligationtoinformtheemployeraboutthespecificreason

forsickleave.Thesicknessabsenteeishoweverobligedtocooperateregardingany

relevantmedicaltreatmentaswellasinitiativesaimedatRTW.Dependingonlabour

marketaffiliation(unemployed,wageearnerorself‐employed)andcollective

METHODS

Page|45

agreements,differentrulesensurethesicknessabsenteefinancialcompensation(in

somecasessicknessbenefits,insomecasestheusualsalarydependingonthecollective

agreement)(96).

Theemployercoverstheexpenses(sicknessbenefitsorusualsalary)forthefirst21

(2009‐2012)or30days(2012‐2014)ofsickleave.Afternomorethan4weeks,the

employermustmeetwiththesicknessabsenteetoexploreopportunitiesofRTW,

possibleworkmodificationsetc.Before35dayshaveelapsed,theemployerisobliged

toinformthemunicipalityabouttheworkeronsickleave(96).Thisinformationserves

twopurposes:

o Theamountofsicknessbenefitsisreimbursedtotheemployerfromthe

municipalitytherebydecreasingtheemployer'sexpenses.

o Relevantauthoritiesatthemunicipalitycanthenorganizeappropriate

initiativesaimingtoassistthepersontoreturntowork,seebelow.

Whenreceivingtheinformationaboutsickleavefromtheemployer,socialworkersat

themunicipalitycollectrelevantinformation,i.e.self‐reportfromthesicknessabsentee

andmedicalinformationfromtheresponsiblephysician.Basedonthisinformation,the

sicknessabsenteeisallocatedintooneofthreecategories(96):

o Category1forwhichcompleteRTWisexpectedwithin8weeks.

o Category2forwhichRTWisexpectedlaterthan8weeksbutwithoutneedfora

multidisciplinaryintervention.

o Category3forwhichRTWisexpectedlaterthan8weekswithaneedfor

multidisciplinaryintervention.

Forcategories2and3,ameetinginvolvingamunicipalitysocialworkerandthe

sicknessabsenteemustbearrangedwithinthefirst8weeksofsickleaveandatregular

intervalsthereafter.

Acertainagreement(§56‐agreement)inTheDanishSicknessBenefitActneedsmention.

Thisagreementappliesincasesofrecurrentshort‐termsickleave.Ifagreedupon,the§56‐

agreemententitlestheemployertoreimbursementofthesicknessbenefitamountfromthe

firstdayofsickleave(notafter21/28daysastheusualpracticedescribedabove).Itthus

exemptstheemployerfromexpensesandmayexempttheworkerfromasenseof"guilt"

forbeingsick(96).

METHODS

Page|46

Ethics approval

Allparticipantssignedinformedconsent.TheRCTwasregisteredatCurrentControlled

Trials,ISRCTN51739408.AllstudieswereapprovedbyTheDanishDataProtectionAgency

(J.no.1‐16‐02‐86‐16).TheRCTandtheprognosticstudywereapprovedbyTheCentral

DenmarkRegionCommitteesonHealthResearchEthics(M‐20090027)fromwhichthe

reliabilitystudywasgrantedexemption(caseno.86,Appendix6).

RESULTS

Page|47

Results

Study 1

Study1evaluatedtheeffectofamultidisciplinaryinterventioncomparedwithabrief

interventioninsicknessabsenteeswithneckorshoulderpain(97).Baselinecharacteristics

ofthestudypopulationarepresentedinTable4.

RESULTS

Page|48

FortheprimaryoutcomeofRTW,registrydataprovided100%followup,whilefollow‐up

dataonpainanddisabilitywereonlyavailablefor79studyparticipants.Analysesof

respondersversusnon‐respondersshowedthatthedropoutwasun‐skewedwithrespect

toRTWandallbaselinevariables,exceptfortheallocationtoexercisegroupsinthenested

RCT(moreresponderswereinthegeneralexercisegroupcomparedwithnon‐

responders).TheMDIhadamediandurationof4.6months(inter‐quartilerange:3.3–7.4)

comparedto3monthsintheBIgroup(inter‐quartilerange3–3).

FourstudyparticipantshadalreadyachievedRTWatbaseline,andthereforewere

excludedfromtheanalyses.Intheadjustedanalyses,14additionalstudyparticipantswere

excludedduetomissingvaluesforsomeofthevariables.Hence,164studyparticipants

wereincludedinthecrudeanalysesand150intheadjustedanalyses.Figure4showsthe

proportionofparticipantsstillonsickleaveduring1‐yearfollowup.RTWwasachievedby

98studyparticipants(50intheMDIgroup(59%)and48intheBIgroup(58%)).When

usingtheBIasthereference,thecrudehazardratiowas0.94(95%CI:0.63;1.41)andthe

adjustedhazardratiowas0.84(95%CI:0.54;1.31).

Forthesecondaryoutcomeofpain,comparisonwasalsomadeusingtheBIasthe

reference.Thecrudeoddsratioforachievingaclinicallyrelevantpainreduction≥2points

was1.10(95%CI:0.54;2.26)andchangedto1.18(95%CI:0.56;2.48)afteradjustmentfor

sex,age‐groupsandbaselinepainintensity.

Figure4.Fractionstillonsickleave(Kaplan‐Meier)

RESULTS

Page|49

Forthesecondaryoutcomeofdisability,crudelinearregressionanalysisshowedthatthe

meanCNFDSscoreafter1yearwas1.37points(95%CI:‐1.91;4.64)higherintheMDI

groupcomparedwiththeBIgroup.Thisestimatechangedto1.09points(95%CI:‐2.26;

4.45)afteradjustmentforsex,age‐groupsandbaselineCNFDSscore.

Study 2

InStudy2,mostofthestudypopulationwascomprisedofwomen(n=31,62%),andthe

meanagewas43.7years.Appendix5presentstheprevalenceofpositiveMRIfindingsfor

allthereaders.Atalldisclevels,theprevalenceofvertebralendplatesignalchanges(Modic

changes)was<10%fortheapriori‐decidedreadercomparisons(A1B1,A1C1,B1C1,

A1A2,andA1B1C1).Therefore,kappastatisticswerenotcalculatedforthisfinding.Forthe

otherMRIfindings,theprevalenceestimatesallowedforkappastatisticsincluding

observationsfromonetosixanatomicalsites(e.g.sixanatomicalsitesincludedinK

statisticsforneuralforaminalstenosis).FromAppendix5,itseemedthatthelabel

"reduceddischeight"wasusedmorefrequentlybyReaderC,andbyReaderAinher

secondassessment,comparedwiththeirfellowreaders.Noothersystematicdifferences

weresuspectedbasedonAppendix5.

Theinter‐raterreliabilityestimatesforallreadercomparisons(A1B1,A1C1,B1C1,and

A1B1C1)arepresentedinTable5.Forthemajorityofthefindings,theoverallinter‐rater

reliability(A1B1C1)rangedfrommoderatetoalmostperfect(moderatetosubstantialfor

discheight,disccontour,uncovertebralandzygapophysealosteoarthritis;moderateto

almostperfectforspinalcanalstenosis;andsubstantialtoalmostperfectforkyphosisand

neuralforaminalstenosis).Forneuralforaminalstenosis,exploratoryanalyseswere

performedtoassesstheinter‐raterreliabilitywhenincludingonlyMRIswithoblique

images.AscanbeseenfromTable6,thisdidnotchangetheestimatesbuttheconfidence

intervalswereslightlybroadened.

Table7showstheintra‐raterreliabilityestimates(A1A2).Theseweremoderateto

substantialfordischeightandmoderatetoalmostperfectforspinalcanalstenosisand

zygapophysealosteoarthritis.Theintra‐raterreliabilitywassubstantialtoalmostperfect

fordisccontour,uncovertebralosteoarthritisandneuralforaminalstenosis,whilefor

kyphosistheintra‐raterreliabilitywasalmostperfect.

RESULTS

Page|50

Table5.Inter‐raterreliabilityestimates

MRIfinding nReaderpair

Observedagreement

(%)

Agreementbychance

(%)Kappa(95%CI)

Kyphosis* 50 A1B1A1C1B1C1

92.0 56.4 0.82(0.75;0.89) 49 89.8 53.6 0.78(0.71;0.85) 49 89.8 52.8 0.78(0.71;0.86) 49 A1B1C1 85.7 31.2 0.79(0.73;0.85) Discheight** 150 A1B1

A1C1B1C1

92.0 52.8 0.83(0.74;0.92) 200 80.0 52.8 0.58(0.46;0.69) 150 77.3 50.0 0.55(0.42;0.68) 150 A1B1C1 74.7 26.4 0.65(0.57;0.74) Disccontour** 177 A1B1

A1C1B1C1

76.8 43.4 0.59(0.49;0.70) 177 79.7 43.3 0.64(0.53;0.74) 200 80.0 47.6 0.62(0.52;0.72) 177 A1B1C1 68.4 21.7 0.61(0.54;0.69) Spinalcanalstenosis** 100 A1B1 97.0 76.0 0.88(0.68;1.00) 100 A1C1 91.0 73.5 0.66(0.47;0.83) 100 B1C1 92.0 74.3 0.69(0.48;0.86) 100 A1B1C1 90.0 63.0 0.74(0.57;0.86) Vertebralendplatesignalchange Toolowprevalences(i.e.≤10%)

Uncovertebralosteoarthritis*** 222 A1B1 90.1 68.0 0.69(0.57;0.81) 237 A1C1 89.0 68.6 0.65(0.53;0.77) 230 B1C1 87.4 70.9 0.57(0.43;0.71) 222 A1B1C1 83.3 53.0 0.65(0.51;0.76) Zygapophysealosteoarthritis*** 270 A1B1 94.8 74.2 0.80(0.70;0.90) 144 A1C1 87.5 74.9 0.50(0.31;0.70) 184 B1C1 85.9 78.9 0.33(0.13;0.53) 135 A1B1C1 83.0 61.0 0.56(0.43;0.70) Neuralforaminalstenosis*** 268 A1B1 90.7 64.1 0.74(0.65;0.84) 287 A1C1 90.2 64.2 0.73(0.63;0.82) 275 B1C1 87.6 65.8 0.64(0.53;0.75) 268 A1B1C1 84.0 46.0 0.73(0.63;0.82)*nreferstothenumberofMRIsassessed.**nreferstothenumberofdisclevelsassessed.***nreferstothenumberofanatomicalsitesassessed(bypoolingrightandlefthandside)

RESULTS

Page|51

Table6.Un‐weightedkappaestimatesforneuralforaminalstenosis.Assessmentofallimagesvs.assessmentofonlyimageswithobliqueslicesavailableReaderpair N Allimages N Onlyimageswithobliqueslices

A1B1 268 0.74(0.65;0.84) 232 0.74(0.65;0.84)A1C1 287 0.73(0.63;0.82) 254 0.72(0.62;0.82)B1C1 275 0.64(0.53;0.75) 233 0.64(0.53;0.76)A1B1C1 268 0.73(0.63;0.82) 232 0.70(0.62;0.78)

Table7.Intra‐raterreliabilityestimates

MRIfinding nReaderpair

Observedagreement

(%)

Agreementbychance

(%)Kappa(95%CI)

Kyphosis* 50 A1A2 96.0 59.6 0.90(0.85;0.96)Discheight** 200 A1A2 84.0 51.5 0.67(0.57;0.77)Disccontour** 174 A1A2 88.5 43.9 0.80(0.71;0.87)Spinalcanalstenosis** 50 A1A2 94.0 76.6 0.73(0.51;0.90)Vertebralendplatesignalchange Prevalencestoolow(i.e.≤10%)Uncovertebralosteoarthritis*** 281 A1A2 90.4 67.0 0.71(0.61;0.81)Zygapophysealosteoarthritis*** 240 A1A2 90.8 68.8 0.71(0.59;0.82)Neuralforaminalstenosis*** 287 A1A2 90.6 62.6 0.75(0.66;0.84)*nreferstothenumberofMRIsassessed.**nreferstothenumberofdisclevelsassessed.***nreferstothenumberofanatomicalsitesassessed(bypoolingrightandlefthandside)

Study 3

Table8showsthebaselinecharacteristicsofboththeentirestudysampleandtheMRI

sample.Ascanbeseen,demographic,patient‐reported,andclinicalvariableswere

similarlydistributedinthetwosamplesasweretheinterventiongroupsoftheRCT(97).

Achievementoftheoutcomes‐WPS(aworkparticipationscore≥75%inWeeks30‐104

afterenrollment)wasalsosimilarlydistributed;itwasachievedby47.6%intheentire

studypopulationandby47.4%intheMRIsample.

RESULTS

Page|52

ResultsfromthecrudelogisticregressionanalysesareseeninTable9.Forthosevariables

yieldingap<0.2,increasedoddsofs‐WPSwereseenformalesex,radiculopathyand

kyphosis,whiledecreasedoddswereseenforincreasedsickleaveduration(months),

increasedÖMPSQscore,≥11tenderpoints,andspinalcanalstenosis.Theremaining

variablesdidnotmeetthecriterionofp<0.2andforthisreasonwerenotcarriedforward

inthemultivariableanalyses(age,education,worker'scompensationclaim,whiplash,

multidisciplinaryintervention,discheightreduction,disccontourchange,neuralforaminal

stenosis,zygapophysealanduncovertebralosteoarthritis(Table9).


Entirestudysample MRIsample

Demographics N NMRI Malesex,n(%) 168 53(31.5) 97 30(30.9)Age,median(IQR) 168 40.5(34.5;48.9) 97 41.4(35.3;49.9)Patient‐reportedinformation Painintensity(0‐10)lastweek,median(IQR) 158 7(5;8) 91 7(5;8)Whiplashtrauma,n(%) 168 24(14.3) 97 13(13.4)Sickleaveduration(months),median(IQR) 168 2.3(1.2;3.0) 97 2.3(1.4;3.0)ÖMPSQscore,mean(SD) 161 122(23.7) 93 121.8(22.1)Education≥3years,n(%) 155 29(18.7) 88 21(23.9)Ongoingworker'scompensationclaim,n(%) 139 36(25.9) 82 19(23.2)Clinicalinformation Radiculopathy,n(%) 168 41(24.4) 97 28(28.9)≥11tenderpoints,n(%) 168 57(33.9) 97 38(39.2)Interventiongroup Briefintervention,n(%)Multidisciplinaryintervention,n(%)

16883(49.4)85(50.6)

9749(50.5)48(49.5)

MRIinformation Kyphosis,n(%) 97 23(23.7)Discheightreduction,n(%) 97 71(73.2)Bulge,protrusionorextrusion,n(%) 83 69(83.1)Spinalcanalstenosis,n(%) 97 14(14.4)Neuralforaminalstenosis,n(%) 84 46(54.8)Zygapophysealosteoarthritis,n(%) 83 39(47.0)Uncovertebralosteoarthritis,n(%) 78 37(47.4)

RESULTS

Page|53

Theestimatesofpredictivevalues,specificity,sensitivityandAUCarepresentedinTable

10.Model1wasbasedonsexonly,i.e.equivalenttoassumingthatallmenwouldhave≥

50%chanceofs‐WPSandallwomenwouldhave<50%chance.Thediscriminatoryability

ofModel1measuredbyAUCwas0.58,whilesensitivitywas0.76andspecificity0.40.A

sensitivityof0.76reflectsthatamongallthosewithu‐WPS,76%werecorrectlyidentified

byModel1.Aspecificityof0.40reflectsthatamongthoseachievings‐WPS,only40%were

correctlyidentified,leaving60%asfalsepositives(i.e.withunnecessaryworriesabout

thosepeople’sworkparticipation).

Table9.Crudeoddsratiosforworkparticipationscore≥75%(s‐WPS).Demographicinformation N Oddsratio(95%CI) p‐valueMalesex 168 2.13(1.10;4.13)* 0.03Age(years) 168 0.99(0.96;1.03) 0.75Patient‐reportedinformation Painintensity(numericratingscale) 161 0.99(0.96,1.02) 0.38Whiplashtrauma(yesvs.no) 168 0.76(0.31;1.81) 0.53Sickleaveduration(months) 168 0.60(0.46;0.78)* <0.001ÖMPSQscore(numericalvariable) 161 0.97(0.95;0.98)* <0.001Education≥3years(yesvs.no) 155 1.22(0.54;2.73) 0.63Ongoingworker'scompensationclaim 139 0.65(0.30;1.40) 0.27Clinicalinformation Radiculopathy(yesvs.no) 168 2.54(1.22;5.33)* 0.01≥11tenderpoints(yesvs.no) 168 0.51(0.27;0.99)* 0.05Multidisciplinaryintervention(MDIvs.BI) 168 0.79(0.43;1.45) 0.44MRIinformation Kyphosis(yesvs.no) 97 2.6(0.98;6.89)* 0.06Discheightreduction(yesvs.no) 97 1.07(0.44;2.64) 0.88Disccontourchange(bulge,protrusionorextrusion)(yesvs.no) 83 1.30(0.40;4.13) 0.66Spinalcanalstenosis(yesvs.no) 97 0.39(0.11;1.35)* 0.14Neuralforaminalstenosis(yesvs.no) 84 1.13(0.48;2.68) 0.78Zygapophysealosteoarthritis(yesvs.no) 83 0.76(0.32;1.82) 0.54Uncovertebralosteoarthritis(yesvs.no) 78 1.21(0.50;2.95) 0.68s‐WPS:successfulworkparticipationscoreÖMPSQ:ÖrebroMusculoskeletalPainScreeningQuestionnaireMRI:MagneticResonanceImaging*:Variablescarriedforwardinthemultivariableanalyses

RESULTS

Page|54

RESULTS

Page|55

Whenpatient‐reportedinformationwasaddedinModel2,increasedestimateswereseenfor

negativepredictivevalue,meaningthatamongthoseclassifiedashaving≥50%chanceofs‐

WPS,thiswasactuallyachievedby70%.Theestimateofpositivepredictivevaluealso

increasedinModel2,meaningthat72%hadu‐WPSamongthosewhowereclassifiedas

having<50%chance.InModel2,specificityandAUCestimatesalsoincreasedcomparedwith

Model1.InModel3,clinicalinformationwasaddedwhichcausedminorchangesinnegative

predictivevalueandsensitivity.NoneofthevalueswereincreasedbyaddingMRIinformation

inModel4.

Finally,Table11showsmoredetailregardingthediscriminatoryabilitiesofthemodels.

BasedonModel1,nobodywasclassifiedashaving<30%or≥70%chanceofs‐WPS.The

additionofpatient‐reportedinformationinModel2increasedthisdiscrimination,i.e.someof

thesicknessabsenteeswereclassifiedashaving<30%or≥70%chanceofs‐WPS.Thisability

todiscriminatebetweenthosehavingthelowestandhighestchancewasnotchangedby

addingclinicalandMRIinformationinModels3and4(judgedbycomparingthedistribution

ofchancebetweenthemodels).

RESULTS

Page|56

Table11.Cross‐tabulationsshowingthedistributionbetweentheclassificationofchanceandactualWPSoutcomes

Informationincluded

N

Outcome

Model1:Demographics

168

Chanceofs‐WPS s‐WPS u‐WPS<30% 0 0

30%to49% 48 6750%to69% 32 21

≥70% 0 0

Model2:Demographic+patient‐reported

161

OutcomeChanceofs‐WPS s‐WPS u‐WPS

<30% 11 3330%to49% 13 3050%to69% 28 19

≥70% 24 3

Model3:Demographic+

patient‐reported+clinical

161


<30% 11 3330%to49% 14 3450%to69% 27 14

≥70% 24 4

Model4:Demographic+

patient‐reported+clinical+MRI

93


<30% 5 2330%to49% 8 1250%to69% 17 12

≥70% 13 3WPS:workparticipationscores‐WPS:successfulworkparticipationscoreu‐WPS:unsuccessfulworkparticipationscore

DISCUSSION

Page|57

Discussion

Comparison with the literature, Studies 1 and 3

ThekeyfindingofStudy1wasthatRTW,painanddisabilityoutcomesweresimilarinthe

MDIandBIgroups(97).ThekeycontributionofStudy3wastheassessmentofdifferent

prognosticmodels:knowledgeaboutsex,sickleavedurationandÖMPSQscore(Model2)

providedimportantinformationinthepredictionofworkparticipation,andthisprediction

didnotimprovewhenaddingclinicalandMRIinformation(Models3and4).Itisalso

noteworthythatworkoutcomesweremodestinbothstudies.InStudy1,only59%inthe

MDIgroupand58%intheBIgroupachievedRTWwithin1‐yearfollowup(97).Followup

waslonger(itreached2years)inStudy3,butamodestworkoutcomewasstillseen:48%

achieveds‐WPS.

Rememberingthatsickleaveisinfluencedbyanumberofsystemsandstakeholders

(Figure1),thefindingsofStudies1and3willbecomparedwiththeliteratureusingthis

'workdisabilityarena'asaframework.Morespecifically,populationcharacteristics

(representingthepersonalsystem)maycontributetounderstandingthemodestwork

outcomesofStudies1and3whichwillbediscussedfirst,followedbyadiscussionofthe

prognosticmodelsinStudy3,sincethesewerealsobasedonpopulationcharacteristics.

Stillrememberingthe'workdisabilityarena',propertiesofinterventions(careprovidedby

thehealthcaresystem),work‐relatedandlegislative/societalfactorswillalsobediscussed.

Populationcharacteristics(discussingworkoutcomes)

Lessthan60%achievedRTWinStudy1(97)andonly48%achieveds‐WPSinStudy3,

outcomeswhichareprobablyattributedtopopulationcharacteristics.Firstandforemost,

thecurrentstudypopulationwasfromasecondarycaresetting,andsuchstudy

populationsgenerallyhavepoorerworkoutcomescomparedwithstudypopulationsfrom,

forexample,workplaces(98).TheGPs'decisionregardingreferraltosecondarycaremay

reflectaseverityofworkdisabilitythatexplainsthemodestworkoutcomes.

Otherpopulationcharacteristicswhichmayhaveaffectedworkoutcomesincludepain

intensity,educationallevel,sickleavedurationpriortoenrollmentandpsychosocial

DISCUSSION

Page|58

aspectsofpain.Themedianpainintensitywas7(Tables4and8),i.e.halfthestudy

participantsscoredtheirpainbetween7and10onanumericratingscale.Increasingpain

intensitywaspreviouslyfoundtobeassociatedwithpoorerworkoutcomes

(44,64,65,99,100)andhence,thehighpainscoresseemprobablecontributorsto

understandingthemodestworkoutcomes.

Supportedbypreviousfindings(64,65),educationallevelmayalsocontributeto

understandingtheworkoutcomes.Itmaybeharderforworkerswithminimalorno

educationtomaintainappropriateworkwhichaccommodatestheirconditioncompared

withworkerswithmoreeducation,duetoagreaterlikelihoodofmanuallabour.Since

approximately80%ofthestudypopulationhadaneducationoflessthan3yearsorno

education(Tables4and8),thismaycontributefurthertounderstandingthework

outcomesofStudies1and3.

Thedurationofsickleaveisprobablyalsoapartoftheexplanation,andwhileallstudy

participantsfulfilledthecriterionof4‐16weeksofsickleave,onequarterofthemhadbeen

onsickleavefor≥12weeks(Table4).Post‐hocexploratoryanalyseswereperformedto

assessRTWinStudy1forthosewithshortersickleaveduration.Theseanalysesshowed

thatforthosewithsickleave≤12weeks,65%achievedRTW.Usingathresholdof≤8

weeks,thecorrespondingmeasurewas72%.Therewerenostatisticallysignificant

differencesbetweenMDIandBIinanyoftheexploratoryanalyses.Exploratoryanalyses

werealsoperformedforStudy3.Amongthosewithsickleaveduration≤12weeks,55%

achieveds‐WPS,whilethiswasachievedby68%amongthosewithsickleaveduration≤8

weeks,i.e.betterthanintheentiresample(48%).Theseestimateswereinlinewith

extensiveliteraturedocumentingthenegativeinfluenceofprolongedsickleaveonwork

outcomes(13,14,23,44,64,65,101,102).Hadtheinclusioncriterionregardingsickleave

durationbeennarrower(e.g.only4‐8weeks),theMDIwouldprobablystillnothave

outperformedtheBI,butbetterworkoutcomeswouldlikelyhavebeenseeninStudies1

and3.

Likewise,thepresenceofpsychosocial‘yellowflags’measuredbytheÖrebro

MusculoskeletalPainScreeningQuestionnaire(ÖMPSQ)wasalsoassociatedwithpoor

workoutcomesinpreviousstudies/reviews(72,73,103).InStudy1,>90%hadanÖMPSQ

DISCUSSION

Page|59

score>90(Table4).Thisthresholdhaspreviouslybeenreportedtoidentifytheriskofsick

leavewithasensitivity>0.8(104,105).Thetoolwasoriginallydevelopedforidentification

ofindividualsatriskofprolongedsickleave(73),andthehighÖMPSQscoresreflectthe

presenceofpsychosocial‘yellowflags’(copingstrategies,fearavoidancebeliefs,

expectationsofRTW,andemotionalstate)whicharelikelytoexplainthemodestwork

outcomesofStudies1and3.

Populationcharacteristics(discussingtheprognosticmodels)

InStudy3,clinicalandMRIinformation(Models3and4)didnotimprovepredictionofs‐

WPScomparedtosex,sickleavedurationandÖMPSQscorealone(Model2).Forthis

reason,Model2willprimarilybeusedforcomparisonwithotherstudies.Despitea

comprehensiveliteraturesearchadaptedfroma2017review(106),noprognosticstudies

wereidentifiedwhichclassifiedprognosticvariablesandperformedstatisticalanalyses

similartothoseusedinStudy3.Therefore,theresultscouldnotbedirectlycomparedwith

otherstudies.Somestudiesdidhoweverreportmeasureswhichareadequatefor

comparison.

InaNorwegianstudy(107),sicknessabsenteeswereaskedtopredictsickleaveexceeding

26weeks,andthisself‐predictionyieldedapositivepredictivevalue(PPV)of0.78(vs.0.72

inModel2).Thismayreflectthatsicknessabsenteeshavedetailedandelaborateinsight

intoallrelevantbiopsychosocialaspectsoftheirsituationandfurthermore,that

achievement/non‐achievementofgivenworkoutcomesmayalsobeinfluencedbythevery

samesicknessabsentees.IncontrasttothehighPPVwerethemeasuresofsensitivityin

thesamestudy(107).Self‐predictionyieldedasensitivityof0.28,whilesensitivitybased

ontheassessmentofdedicatedmedicalconsultantsonlyreached0.07.Iftheintenthad

beentoidentifyindividualsneedingacertainRTWintervention,boththesickness

absenteesandthemedicalconsultantsinthisstudy(107)wouldhaveidentified

considerablyfewerindividualsthanModel2(sensitivity=0.74).Whencomparingthe

informationthatwasavailabletothemedicalconsultants(107)withtheinformationof

Model2,themostimportantdifferencewasthatsickleavedurationpriortoenrollment

wasonlyknowninhalfofthecasesinFletenetal.'sstudy.Ashasalreadybeenargued,this

isanimportantprognosticfactorforworkoutcomes.

DISCUSSION

Page|60

Inanotherstudy(108),rehabilitationprofessionalspredictedthechanceofRTW,andtheir

estimateswereconcordantwithactualRTWin73%ofthecases,i.e.similartothe

specificityofModel2(0.68).Therehabilitationprofessionalshadknowledgeaboutsexand

sickleavedurationlikeinModel2,buttheiradditionalknowledgeaboutthereasonforsick

leave,unemployment,age>45yearsand‘gutfeeling’(108)didnotelicitbetterspecificity

thaninModel2.

Withouttheuseofanystatisticalmodels,butinsteadbasedonclinicalexperience,

physicianswereaskedtodescribewhy195sickleavecasereportswereconsidered

particularlyproblematicinaqualitativestudy(109).Ofthesecases,63%hadsickleave

lasting>12monthsand29%>3months.Onlyinaminorityofcasesweretheresultsofthe

clinicalexaminationdescribed(itmayhavebeenperformedinmorecasesbutwasnot

described,hencepresumablynotconsideredimportantbythephysicians).Theresultsof

Study3suggestthattheycorrectlyconsideredlongsickleavedurationasproblematicand

correctlyofferedattentiontotheclinicalexaminationinonlyaminorityofcases.

InModel2,thevariablescontributingtopredictionofworkparticipationweresex,sick

leavedurationandÖMPSQscore.Thefactthatsickleavedurationandpsychosocialaspects

ofpainmeasuredbyÖMPSQwerepredictiveofs‐WPScorroboratedthewell‐established

influenceofthesefactorsonworkoutcomesasdescribedabove

(13,14,44,64,65,72,73,103,110).Regardingtheimpactofsex,resultshavebeenconflicting

(44,64),andgiventhedichotomousnatureofsex,themodestpredictiveabilitiesofModel1

(basedonlyonsex)werenotsurprising.

ComparedwithModel2,clinicalinformation(radiculopathy,≥11tenderpoints)andMRI

information(kyphosis,spinalcanalstenosis)wereaddedinModels3and4,andneitherof

thesemodelselicitednotableimprovementsinthepredictionofsuccessfulwork

participation.Acrudeassociationbetweentenderpoints(numericalvariable)andwork

outcomeswasfoundinacohortstudyoflowbackpainpatients,buttheassociationwas

notmaintainedinthemultivariablemodel(111).Thesamecohortstudydidnotfind

radiculopathyassociatedwithworkoutcomes(111)whichwascorroboratedbyareview

onsubacuteandchroniclowbackpain(106).Thesepreviousfindingsareinlinewiththe

limitedimprovementofpredictioninModel3.RegardingMRIinformation,onlyonestudy

DISCUSSION

Page|61

wasidentifiedwhichexaminedtheassociationwithworkoutcomes.Thisstudy(112)

suggestedanassociationbetweenModictype1changesandunsuccessfulRTW.

Unfortunately,Modictype1changescouldnotbeexploredinStudy3sincetheintra‐rater

andinter‐raterreliabilityofthisfindinghadnotbeenestablished(Study2).

Somepopulationcharacteristicsthatareknowntoinfluenceworkoutcomeswerenot

includedinthemultivariablemodelsofStudy3.Examplesincludeeducationalleveland

workers'compensationclaims.Basedontheoddsratios(Table9),theinfluenceofthese

variableswasinlinewithpreviousfindings(64,65,113),butsincetheydidnotfulfillthe

criterionofp<0.2,theywerenotcarriedforwardinthemultivariableanalyses.The

limitedsamplesizemightexplainwhythiscriterioncouldnotbefulfilledinthecrude

analysesanditcannotberefutedthatalargersampleincludingmoreinformationinthe

multivariablemodelsmighthaveimprovedpredictionofsuccessfulworkparticipation.

Componentsofinterventions

Understandingofthecomponentsoftheinterventionsmaycontributetoknowledgeasto

whytheMDIdidnotimproveRTW,painanddisabilitycomparedwiththeBIinStudy1.

Similarresultswereseeninastudyonlowbackpain(114).Commontothisstudy(114)

andStudy1wasthatboththeBIandtheMDIgroupswereofferedthesameclinical

interventionbydedicatedspineclinicians.Theimpactofthisshouldprobablynotbe

underestimated.Secondly,theMDIhadamediandurationof4.6monthswhichwaslonger

thanafewotherinterventionsshowingpositiveeffectonworkoutcomes(115‐117).Ifthe

durationoftheMDIkepttheparticipantsina‘pendingstate’whileallalongtheir

probabilityreducedofreturningtothelabourmarket,thisdurationmighthavebeen

counterproductive.Thirdly,andaboveall,akeyfeatureinunderstandingthefailureofthe

MDIseemstobethelimitedinvolvementofworkplaces,afindingwhichgovernedthe

designofTablesAandB(Appendix7)whichoffercomparisonswithotherRCTs.TableA

presentsstudieswithworkplaceinvolvementandTableB,studieswithscarceorno

workplaceinvolvement.Beyondthat,thecontentofthenewinterventionsintheseRCTs

differedgreatly:fromclinicalassessmentandguidanceresemblingtheBIofStudy1(38‐

40),tocognitivebehaviouraltherapy(116,118‐122),self‐managementprogrammes

(123,124)ordietarysupplements(122).Additionalinformationcoversthesurnameofthe

DISCUSSION

Page|62

firstauthor,publicationyear,painlocation,sickleavedurationpriortoenrollment,and

finally,thestudyresultsbychoiceofcolour:Greenforstatisticallysignificantimprovement

ofworkoutcomesfollowingthenewinterventioncomparedwiththereference

intervention,andredforstudieswherenosuchdifferencewasfound.

Despitetheinterventiondifferences,itisreadilyseenthatthemajorityofRCTswith

improvedworkoutcomesactivelyinvolvedtheworkplacesintheRTWprocess(115‐

117,125,126)whichisinlinewithreviewsassessingtheeffectofworkplaceinterventions

onworkoutcomes(41,42).InStudy1,suchinvolvementwasindeedtheintentforallstudy

participantsattendingtheMDI,butsincemanyofthemwerenotinclinedtoinvolvetheir

workplaces,itwasonlydonein22%ofthecases.

AnotherinterestingcharacteristicemergedfromtheseRCTs,namelythatwiththe

exceptionofonestudy(116),RCTsenrollingsicknessabsenteeswithsickleaveduration≤

3monthsimprovedworkoutcomesinmorecases(38‐40,115,117,125,126)thandidRCTs

enrollingsicknessabsenteeswithasickleavedurationexceeding3months(114,118,120‐

122,127‐130).Whethertheinterventioncontentorthesickleavedurationexplainwhy

workoutcomeswerenotimprovedintheselatterstudies(114,118,120‐122,127‐130)

cannotbedetermined.Thissimplydemonstratesthattheinfluenceofsickleaveduration

wasfurthersupportedwhenexaminingtheseRCTs.

Work‐relatedandsocietal/legislativefactors

Theinfluenceofwork‐relatedfactors(e.g.physicallystrenuousworkorpoorperceived

support)ontheprocessofsickleaveiswell‐established(7,44,64)andsoistheimpactof

societalandlegislativefactors(34,62,131).Sinceworkplaceswereonlyinvolvedina

minorityofcases,itcannotberefutedthatwork‐relatedfactorsmightpartlyexplainthe

modestworkoutcomesofStudies1and3.Itmayalsobethatsocietal/legislativefactors

haveaffectedtheseworkoutcomes.

Forinstance,thelimitedinvolvementofworkplacesinStudy1wasexplainedbythefact

thatmanyparticipantswerenotinterestedinsuchinvolvement.Thismayattributetothe

DanishActonHealthInformation(132),anactwiththepurposeofpreventinghealth‐

relateddiscriminationinworkplaces.Theactensuresthatemployersonlyunderspecial

DISCUSSION

Page|63

circumstanceshavetherighttoknowthehealthconditionsoftheiremployees.InStudy1,

participantsmaynothavewantedtheiremployerstoknowthespecificreasonforsick

leaveandthereforedeclinedworkplaceinvolvement.Withtheavailabledata,thiscannot

beunraveled,butitcouldrepresentacounterproductiveinterplaybetweenlegislativeand

workplace‐relatedinfluences(theDanishActonHealthInformationismeanttoprotectthe

workers,butiftheemployersdonotknowthereasonforsickleaveitmaybedifficultto

providethebesthelpandaccommodationattheworkplaces).

Norcanitbeelucidatedorrefutedwhetherwork‐relatedorsocietalfactorshaveaffected

theprognosisexploredinStudy3.Owingtothelimitedsamplesizeandtheavailabledata,

thecontributionofthesecouldnotbeexplored.

Comparison with the literature, Study 2

InStudy2,primarilysubstantialreliabilityestimateswerefoundsuggestingthatboththe

inter‐raterandtheintra‐rateragreementwasnotablybetterthanwhatcouldbeexpected

bychance.Fourkeyissuesmayexplainthedifferencesinreliabilityestimateswhen

comparingtheresultsofStudy2withotherreliabilitystudiesonMRIofthecervicalspine:

1)theeducationalbackgroundoftheMRIreaders,2)thequalityoftheimages,3)theuseof

anevaluationmanual,and4)theclassificationofMRIfindings.

DifferencesintheeducationalbackgroundandlevelofexperienceoftheMRIreadersmay

explainwhythereliabilityestimateswerepoorerinStudy2comparedwiththoseofother

studieswhichusedreaderswithsimilareducationandexperience(58,59).

Butsimilareducationalbackgrounddoesnotseemtoensurehighreliabilityestimates.

AlthoughtheMRIreadersinanotherstudyhadsimilareducationandexperience,the

reliabilityestimatesofthisstudy(133)werepoorercomparedwithStudy2.Apotential

explanationisabetterimagequalityinStudy2whichhaslikelyreducedrandomvariation

intheMRIassessments.Anotherpossibleexplanationistheuseoftheevaluationmanual

whichreducedbothrandomandsystematicvariationintheMRIassessments.

Finally,theconsequentuseoftheevaluationmanualalongwithsimple,comprehensible

classificationsofMRIfindingsareprobableexplanationswhythereliabilityestimatesof

Study2exceededthoseofotherstudies(60,82,134,135)

DISCUSSION

Page|64

Strengths

Thestudiesofthisdissertationhavesomestrengthsbothintermsofthechosenmethods

andtheattempttocoverresearchquestionsthat,todate,havescarcelybeencoveredinthe

literature.

First,therandomiseddesignofStudy1waschosenbecausethisdesignisthestandardof

excellenceforstudiesoftreatmenteffect,duetoitspreventionagainstbothknownand

unknownconfounders.

Second,biaswasreducedindifferentways:

Bywayofallocationconcealment,selectionbiaswasavoidedinStudy1;i.e.the

allocationtointerventiongroupscouldnotbeaffectedbytheclinicians,

investigatorsorstudyparticipants.

Thestudyparticipantsandtherheumatologisthadnoknowledgeaboutthe

subsequentrandomallocationtointerventiongroups,nordidtheyknowthe

outcomesofthestudywhenprovidingtheirassessmentsofexplanatoryvariables

(studyparticipantsfilledinthebaselinequestionnairesandtherheumatologist

performedtheclinicalexaminationpriortorandomisation).Thisprotectedagainst

informationbiasinStudies1and3.

AssessmentofthechosenworkoutcomesinStudies1and3wasperformedwithout

knowledgeoftheinterventiongroupsorotherbaselinecharacteristics,thusalso

preventinginformationbias.

TheuseofregistrydatainStudies1and3ensured100%followuponwork

outcomes,therebyreducingtheriskofattritionbias.

BlindingofthethreeMRIreadersinStudy2servedasprotectionagainstthe

informationbiasthatcouldhavethreatenedtheestimatesifthereadershad

possessedknowledgeabouttheageofthestudyparticipants,theirclinicaldataor

theMRIassessmentsoftheirfellowreaders.

BlindingoftheMRIreaderinStudy3protectedagainsttheinformationbiasthat

couldhaveoccurredifshehadknowntheworkoutcomesofthestudyparticipants

whenassessingtheMRIs.

DISCUSSION

Page|65

Third,therigoroususeoftheevaluationmanualinStudy2wastheprerequisitefor

minimisingbothsystematicandrandomvariationintheMRIassessmentsandenhancing

reliability.

AfourthstrengthwasthestatisticalapproachinStudy3whichresembledthetypicalorder

ofinformationsoughtinaclinicalsetting.Theriskofdata‐drivenresultswasreducedby

theaprioridecisiononthisstatisticalapproach.

Fifth,exploringtheimpactofMRIfindingsonworkoutcomesisconsideredastrengthsince

therehasbeenapaucityofstudiesinvestigatingthis.Indeed,ampleliteraturehascovered

theinfluenceofpsychosocial,work‐related,andsocietalfactorsonthecourseofsickleave

(7,17,62).Withthisinmind,therationaleforexploringapossibleimpactofclinicaland

MRIfindingsonWPSprognosiscouldbereasonablyquestioned.Nevertheless,theaimof

Study3wasconsideredrelevantduetothefollowingconsiderations:

ThenumberofcervicalspineMRIshasincreaseddisproportionatelycomparedwith

theprevalenceofneckandshoulderpaininDenmark(47‐49).IncreaseduseofMRI

isalsoseenelsewhere(50,51).

PatientsbelievethatMRIhasthecapacitytorevealthetruthabouttheirsituation

(52).

Sincemedicaldiagnosesarestillrequiredbythesysteminordertolegitimisesick

leave,aninappropriatefocusonthemedicalconditionratherthanonpsychosocial,

work‐relatedandsocietalissuesmaybeobservedamongbothsicknessabsentees

andtheirhealthcareproviders.

Subsequenttotheabove,theexplorativeapproachofStudy3wasconsideredrelevant,

withthepotentialofprovidingimportantknowledgetoallstakeholdersinvolvedinthe

processofsickleave.

Finally,thesicknessbenefitlegislationinDenmarkunderwentonlyaminorchangeinthe

studyperiod.Alegislativechangewhichreducedtherightforsicknessbenefitsfrom52to

22weeksaffectedonlysixparticipants(MDI:n=3andBI:n=3).Ifmorestudy

participantshadbeenaffected,biastowardthenullhypothesiscouldhaveoccurredin

DISCUSSION

Page|66

Studies1and3.Butasthiswasnotthecase,thelegislativechangeisnotsuspectedtohave

affectedtheresults.

Limitations

WhenassessingtheresultsofthecurrentPhDdissertation,methodologicallimitations

mustbeconsidered.Suchconsiderationsarecoveredinthefollowingsection:Notallofthe

themesareconsideredtohaveaffectedtheestimatesbutareneverthelessdescribed,

becausethiscarefulconsiderationisnecessarytoassessthecredibilityoftheresults.

Samplesizeconsiderations

InStudy1,theprecedingpowercalculationrequired170studyparticipants(85ineach

group)todetecta15%differenceinRTWgivenapowerof70%andasignificancelevelof

0.05.Overall,168participantswereenrolled(85inMDIand83inBIgroup),buttheissue

ofinteresthereisnotsomuchthelackoftwoparticipantsbutratherthechosenpowerin

theoriginalsamplesizecalculation.Apowerof70%yieldsa30%riskthatadifferencein

RTWbetweeninterventiongroupswouldnotbedetectedasstatisticallysignificant.Such

generalconsiderationsconcerningpowerandtheriskoftype2errorsshouldalwaysbe

made.InthespecificcaseofStudy1however,andasoutlinedincomparisonwiththe

literature,thefailureoftheMDIisnotsuspectedtorepresentatype2error.

InStudy2,apowercalculationwasnotperformed,butsamplesizerecommendationsfor

reliabilitystudieswerefollowed(n=50)(77).Ifalargersamplesizehadbeenused,the

confidenceintervalswouldhavebeennarrowedfurther,butsubstantialchangesinthe

reliabilityestimateswouldnotbeexpected.

Finally,thestudypopulationofStudy3comprisedallofthestudyparticipantsfromStudy

1(n=168),andowingtothissamplesizeonlyalimitednumberofprognosticvariables

couldbeincludedinthemultivariableanalyses.Hadthesamplesizebeenlarger,itis

possiblethatinclusionofmorevariables(e.g.work‐relatedfactors)mighthaveyielded

betterdiscriminatoryabilitiesofthemodels.

DISCUSSION

Page|67

Bias

Afterrandomisation(Table2),blindingofthestudyparticipantswithrespectto

treatmentallocationwasnotperformedinStudy1.Theoretically,knowledgeabout

thetreatmentallocationmightintroduceinformationbias(knowledgeaboutthe

expandedtreatmentratherthantheactualMDIcontentitselfcouldhavecaused

betterRTWintheMDIgroup–andviceversafortheBIgroup).Owingtothenature

oftheinterventionsinStudy1,blindingofthestudyparticipantswasnotpossible.

AndsinceequalproportionsofthetwointerventiongroupsachievedRTW,thisrisk

ofinformationbiasisnotsuspectedtohavedistortedtheestimates.

Afterrandomisation,blindingofthecliniciansinStudy1wasnotpossibleeither

(Table2).Theoretically,knowledgeabouttheinterventiongroupcouldhave

affectedthepracticeandcommunicationofthecliniciansinwaysthataffectedthe

studyoutcomesmorethanthecontentoftheinterventions.Butasoutlinedabove,

RTWratesweresimilarinthetwogroups.Forthisreason,thelackofclinician

blindingisnotsuspectedtohavebiasedtheestimatesofStudy1.

InStudy2,readerAreceivedtuitionfromreaderBpriortoperformingthe

reliabilitystudy.Thistuitioncouldrepresentapotentialsourceofbias,whichwas

addressedbymeticuloususeoftheevaluationmanual.Byfollowingthismanual,all

threereadershadtoadjusttotheresearchsettingofStudy2comparedwiththeir

usualclinicalpractice.Furthermore,basedontheprevalencetableofStudy2,

readersAandBevaluatedseveralMRIvariableswithsomenotabledifferences.In

conclusion,anypotentialbiasfollowingfromthetuitionseemstohavebeen

satisfactorilyminimised.

Confounding

InStudy1,adjustmentsweremadeforfivepotentialconfounders.Thiswasdoneoutof

fidelitytodecisionsmadepriortodataappraisal.However,onlythreeofthesevariables

(age≤40years,physician'sdiagnosesandpart‐timesickleave)weresuspectedtobe

unevenlydistributedbetweentheinterventiongroupsbasedonTable4.Ingeneral,the

attempttoaddressconfoundingintheanalysesshouldberenderedirrelevantgiventhe

inherentprotectionagainstknownandunknownconfoundersintherandomiseddesign.

However–especiallyinsmallstudypopulations–thereisariskthatconfoundingfactors

DISCUSSION

Page|68

willbeunevenlydistributeddespiterandomisation.Inhindsight,itwouldhavebeen

preferabletorecommendthatnoadjustmentsbemadeoratleastonlyadjustmentsfor

variablesthatwerebothunevenlydistributedbetweeninterventiongroupsand

constitutedpotentialconfoundersbasedonknowledgefromtheliterature.

Self‐reportedsickleaveandregisterdata

WhiletheuseofDREAMhassomeattractiveadvantagesregardingtheoutcome

assessment(coveredintheStrengthssection),potentiallimitationsoccurintwocases:

1. Theregistrationatbaseline

AtthetimeofenrolmentinStudy1,DREAMdatawerenotaccessible.Therefore,

fulfillmentofthesickleavecriterion(4‐16weeks)hadtorelyonself‐reportbasedon

whichall168participantsweresick‐listedatbaseline.However,whenappraising

DREAMdata,15oftheseparticipantsdidnotfulfillthecriterionof4weekssickleave.It

couldbetemptingtoexcludethese15studyparticipantsfromtheanalysestohaveas

‘clean’registrydataaspossible.Thiswasdoneinapreviousstudy(136)butnotin

Study1owingtothefollowingconsiderations:

Exclusionwouldimplyathreattotherandomiseddesignanditsinherent

protectionagainstconfounding.

Exclusionwouldintroduceariskofselectionbiasifthedistributionofexposure

(allocatedinterventiongroup)andoutcome(RTW)amongthe15study

participantsdifferedfromtherestofthesample.Thisdistributionwasactually

skewed:fiveofthe15studyparticipantsdidnotachieveRTW,fourofwhom

wereintheBIgroup,whiletheremaining10studyparticipantsachievedRTW,

sevenofwhomwereintheMDIgroup.Hence,exclusionofthe15study

participantswouldhaveintroducedattritionbiasandledtounderestimationof

theeffectoftheMDI.

SinceDREAMdatawerenotavailableonthedayofinclusion,theenrolmenthad

torelyonself‐report.Accordingtoclinicalexperience,patientsinsecondary

caredoknowwhethertheyareonsickleaveornot.Tofurthersupportthis,the

self‐reportedmeasureofsickleaveduringthepastmonthhasbeenfound

reliableandvalid(137).

DISCUSSION

Page|69

Finally,whilethese15werenotregisteredashavingsickleavebeforebaselinein

DREAM,11ofthemwereregisteredashavingsickleaveimmediatelyafter

baseline.Thatleavesonlyfourparticipantswhoexperiencedtheoutcome(4

consecutiveweeksofself‐support)immediatelyafterbaseline.Bydefault,these

fourwereexcludedfromtheCoxproportionalhazardsanalysis.

2. Outcomeassessment–theregistrationofshort‐termsickleaves

AnoverviewofDREAMregistrationmechanismsthatmayaffecttheappraisalofwork

outcomesarepresentedinTable12below.Duetotheseregistrationmechanisms,

absencesduetosickleaveshorterthantheemployerperiod(21daysfrom2009‐2012

and30daysfrom2012‐2014)areusuallynotregistered.Thisimpliesthatonlysick

leaveexceeding21/30consecutivedayswillappearassickleaveintheregistry.This

introducesariskofmisclassification,namelyariskthatsickleaveshorterthan21/30

daysdoesnotappearintheregistry.Thiscanbeillustratedbycomparingtwo

imaginaryworkers:Xwhoisonsickleavefor15continuousweeksandYwhohas15

episodesof1‐weeksickleave.InDREAM,workerXwillappearashavinghad15weeks

ofsickleavewhileworkerYwillappearasnothavingbeenonsickleaveatall.InTable

12,thispotentialmisclassificationisdenotedmisclassificationtypeA.

Theoppositeriskofmisclassificationappliestoworkerswhohavea§56‐agreement.

Fortheseworkers,theregistrationproceduremakessickleaveappearintheregistry

alreadyfromDay1.Ifaworkerwitha§56‐agreementhas15episodesof1‐weeksick

leave,itwillappearas15weeksofsickleave(i.e.nomisclassificationcontrarytothe

above‐mentionedexampleofworkerY).Butaworkerwitha§56‐agreementwhohas

15individualdaysofsickleavewillappearashaving15weeksofsickleaveinDREAM.

Thisisexplainedbythehierarchyofcodes(misclassificationtypeB).

DISCUSSION

Page|70

Table 12. Overview of registration mechanisms in DREAM and potential types ofmisclassificationLegalarrangementoftheworker

Patternofsickleave

Numberofsickleaveweeksinregistry

Mechanisminregistrationprocedure

Typeofmisclassi‐fication

Withorwithouta§56‐agreement

15continuousweeksofsickleave

15 Allsickleaveexceedingtheemployerperiodwillberegistered1:1.Thatis,thelengthofsickleaveappears1:1inDREAM.

None

Withouta§56‐agreement

15individualepisodesofsickleave,eachlasting1week

0 Onlysickleaveexceedingtheemployerperiodwillberegistered.Thatis,periodsshorterthantheemployerperioddonotappearassickleaveinDREAM.

A

Witha§56agreement

15individualepisodesofsickleave,eachlasting1week

15 The§56‐agreement"overrules"theusualregistrationprocedureandsickleavesappearfromday1inDREAM.

None

Witha§56‐agreement

15daysofsickleave,eachofwhichappearingindifferent,notnecessarilycontinuousweeks

15 Owingtothehierarchyofcodes,aweekwith4self‐supportdaysand1sickleavedaywillappearas1weekofsickleave.

B

TheextenttowhichmisclassificationtypeAhasoccurredcanunfortunatelynotbe

unraveled.Iftheproblemiscommon,theworkoutcomesinStudies1and3mighthave

beenoverestimated.Sincethepotentialmisclassificationisexpectedtobeundifferentiated

(notassociatedwithinterventiongroupsinStudy1andnotassociatedwithprognostic

variablesinStudy3),itcouldpotentiallycausebiastowardsthenullhypothesis.

DISCUSSION

Page|71

RegardingmisclassificationtypeB,dataonthedistributionof§56‐agreementsinthestudy

populationwerenotavailable,sotheimpactofthispotentialmisclassificationcan

unfortunatelynotbecompletelyelucidated.Itwas,however,possibletoidentifystudy

participantswithabsencesduetosickleaveshorterthantheemployerperiod.Underthe

assumptionthattheseweeksrepresented§56‐agreementsandhenceoverestimatedthe

durationofsickleave,theyweretentativelyregardedasworkingweeks.Thesesensitivity

analysesdidnotchangetheestimatesofStudies1and3,soitdoesnotseemthat

misclassificationtypeBhasbiasedtheresultsofthesestudies.

Externalvalidity

Cautioniswarrantedwhenmakinginferencesbasedontheresultsofthisdissertation.The

startingpointofthedissertationisthatofStudy1,theRCT,which,owingtothedesign,

datacollection,andanalyses,isconsideredtohavestronginternalvalidity.TheRCThas

limitations,however,withrespecttoexternalvalidity.Thisisexplainedbytwomajor

issues:

Theenrollmentofstudyparticipantsdependedonfulfillmentofinclusionand

exclusioncriteria.Sinceapproximatelyhalfofthereferredindividualscouldnotbe

included(Figure2),considerationmustbemadeastowhetherthosenotincluded

differfromthosewhowereincludedandwhetheranysuchdifferencescouldhave

influencedtheoutcomes.Forethicalreasons(lackofexplicitconsent),thiscannot

beelucidatedbycomparisonofparticipantswithnon‐participants.

Theenrollmentofparticipantsdependedonthereferralofeligibleindividualsfrom

theirGPs.WhetherornottheGPshavereferredonlysicknessabsenteesforwhom

theyexpectedaparticularlypoor(orgood)prognosiscanunfortunatelynotbe

assessedwiththeavailabledata.

Followingonfromthesetwoconsiderations,theresultsofStudy1areconsidered

generalisabletopatientswithneckorshoulderpainwhoareseeninsecondarycareand

whofulfilltheinclusionandexclusioncriteriathatwereappliedintheRCT.Sincethe

cohortinStudy3comprisedallRCTparticipants,thesameconsiderationsregarding

externalvalidityapplytoStudy3.

DISCUSSION

Page|72

ThefindingsofStudy2suggestthathealthcareprofessionalscanachieveagreementbetter

thanwhatisexpectedbychance,iflessexperiencedreadersreceivepropertraining,an

experiencedradiologistisinvolvedandsimpleclassificationsoffindingsareapplied.

However,theconditionsofStudy2werethoseofacontrolledresearchsetting,nota

clinicalsetting.Inclinicalsettings,theremaybedifferencesfromhospitaltohospitalasto

howcloselyradiologistsandclinicianscollaborateandthusdifferencesmayariseintheir

agreementonMRI.Therefore,theresultsofStudy2arenotnecessarilygeneralisableto

otherhealthcareprofessionalsinothersettings.Toassessthis,independentstudieswould

beneeded.

CONCLUSIONS

Page|73

Conclusions

Firstly,themultidisciplinaryinterventiondidnotimprove1‐yearRTW,painordisabilityin

sicknessabsenteeswithneckorshoulderpainwhencomparedwiththebriefintervention.

Secondly,inter‐raterestimatesshowedprimarilysubstantialreliability(K≥0.61)among

threehealthcareprofessionalsassessingdegenerativefindingsoncervicalspineMRI.For

allfindings,theintra‐raterreliabilityestimateswerehigher.

Thirdly,thepredictionofworkparticipationwasnotimprovedwhenaddingclinicaland

MRIinformationascomparedwiththepredictionbasedondemographicsandpatient‐

reportedinformationalone.

IMPLICATIONS

Page|74

Implications

Thisdissertationcontributednewknowledgeregardinginterventions,MRIassessmentand

prognosisinsicknessabsenteeswithneckorshoulderpain.Giventheresultsofthis

dissertationandinthelightofthescientificliterature,someimplicationsdeserve

consideration.

For clinical practice

First,theMDIdidnotimprovereturntowork,painordisabilitywhencomparedwiththe

BI(97).Overall,theseresultsareinlinewithlowbackpainstudiescomparingthesame

twointerventions(114,130,138)anddonotsupportbroadimplementationoftheMDIin

itsgivenform.

Further,clinicalandMRIinformationprovidednoadditionalvalueinthepredictionof

workparticipationcomparedwithdemographicandpatient‐reportedinformationalone.

Discriminatingbetweenthosewiththelowestandhighestprobabilityofsuccessfulwork

participationhadsomelimitationsandhence,cautioniswarrantedregardinguseofthe

modelsinclinicalpractice.Moreover,Study3wasexploratoryinnature,highlightinga

needforindependentstudies,ideallyinbothprimaryandsecondarycaresettings.Yet,the

findingsareusefulinthecommunicationwithbothsicknessabsenteesandother

stakeholders:whenassessingworkparticipationprognosis,focusshouldbeon

demographics(sex)andpatient‐reportedinformation(sickleavedurationand

psychosocialdimensionsofpain(ÖMPSQ))ratherthanonclinicalandMRIinformation.

For research

Overall,thereisstillaneedforresearchexploringthetimingandintensitywithwhich

differentstakeholdersshouldbeinvolvedinRTWinterventions(139).Basedonthe

literature,thereisreasontobelievethatamultidisciplinaryinterventioninvolvingthe

workplacewillimproveworkoutcomes(41,42).Uniforminterventionsarenotneeded,but

ratherinterventionsthattargettheindividualneedsofthesicknessabsenteeswithinthe

boundariesofthegivenlegislativeandsocietalcontexts.Theinterventionmapping

approach(previouslyusedfordevelopingaRTWprogrammeinoccupationalLBP(140))

IMPLICATIONS

Page|75

couldoffertheframeworkfordevelopingsuchinterventionsforsicknessabsenteeswith

neckorshoulderpain.

Butallsicknessabsenteesdonotnecessarilyrequireamultidisciplinaryintervention

(139);somemaybenefitfrommoremodestinterventionsandthereisanobvioussocietal

interestinachievingthebestpossibleRTWoutcomeswiththeleastcostlyinterventions.In

Denmark,thereisaneedforresearchcomparingtheBItousualtreatment(i.e.whatis

offeredbyprimaryhealthcare).InNorway,aninterventionlargelysimilartoourbriefone

causedsubstantiallyreducedsickleaverateswhencomparedwithusualcare,bothinthe

shortandthelongterm(38‐40).

ThereliabilityestimatesofStudy2weremostlysubstantial(K≥0.61),i.e.agreement

notablyexceededtheagreementthatcouldbeexpectedbychance.SinceStudy2was

performedinacontrolledresearchsetting,confirmatorystudiesinclinicalsettingsare

needed.AndalthoughitisindeedimportanttoagreeonMRIinterpretation,such

agreementisnot‘enough’.WhenusingMRI,onestepistoensurereliabilityinthe

assessmentofimages.Butitisalsoimportanttofurtherexploretheassociationsbetween

imagingandworkoutcomes.Furtherstudies,preferablyfrombothprimaryandsecondary

caresettings,areneededtoassesswhetherclinicalandMRIinformationimprovethe

predictionofworkparticipation.

PUBLICHEALTHPERSPECTIVES

Page|76

Public health perspectives

Intheabsenceofamagicbulletregardingthecontentofinterventions,theburdenofsick

leaveduetoneckorshoulderpainremainssubstantial.Akeyfeatureofaddressingthis

publichealthproblemistogetallplayersonside(34,141).Thatis,althoughtheymayhave

differentvaluesandinterests,thesystemsandstakeholdersin'thearenainworkdisability

prevention'needtoworktogethertowards1)achangeinmaladaptivepainbeliefs,2)a

behaviourchange,and3)mutualrecognitioninallwingsofthearena,acknowledgingthe

importanceandinfluenceoftheotherstakeholders.

Inthegeneralpopulation,thereisstillaneedforknowledgeaboutthebenigncharacterof

neckandshoulderpain,abouttheimportanceofstayingactiveandstayingatworkif

possible.Changingmaladaptivepainbeliefsmaynotpreventneckorshoulderpain

occurring,nordoeschangingbeliefsnecessarilyelicitbehaviourchange(30).Butitwould

constituteaprerequisiteforchangingcounter‐productiveillnessbehaviourandthereby

establishaprerequisiteforreducingsickleaveifreinforcedbytheotherstakeholdersof

thearena.

Inthehealthcaresystem,healthcareprofessionalsfromdifferentdisciplinesneedto

provideconsistentinformationandadviceaboutneckandshoulderpain,aboutcauses,

treatmentandprognosis.Aswithlowbackpain,thereisalsoaneedforputtinganendto

ineffectivetreatmentswhichdonotreinforcehealthycopingstrategiesandmaykeep

patientsinacounterproductivesickrole(141).

Workplacesandlegislative/societalstructuresareequallyimportantingettingallplayers

onside.Bothshouldpromoteandsupportachangeofpainbeliefsandconcurrentlyoffer

theopportunityandmotivationforbehaviourchange,i.e.forstayingat,orreturningto,

workunderreasonablecircumstances.Ashaspreviouslybeenemphasised,lawsand

regulationsshouldmirroracknowledgementofthecomplexinterplayandinfluences

amongstakeholders(139).

REFERENCES

Page|77

References

(1)SchildHH.MRIMadeEasy....wellalmost.1sted.Bonn:BayerPharmaAG;2016.

(2)HerbertRD.Cohortstudiesofaetiologyandprognosis:they'redifferent.JPhysiother2014

Dec;60(4):241‐244.

(3)MokkinkLB,TerweeCB,PatrickDL,AlonsoJ,StratfordPW,KnolDL,etal.TheCOSMINstudy

reachedinternationalconsensusontaxonomy,terminology,anddefinitionsofmeasurement

propertiesforhealth‐relatedpatient‐reportedoutcomes.JClinEpidemiol2010Jul;63(7):737‐745.

(4)MurphyRJ,CarrAJ.Shoulderpain.BMJClinEvid2010Jul22;2010:1107.

(5)HochbergMC,SilmanAJ,SmolnenJS,WeinblattME,WeismanMH.Chapter70:TheShoulder.

Rheumatology.5thed.:MosbyElsevier;2011.p.684.

(6)ModicMT,SteinbergPM,RossJS,MasarykTJ,CarterJR.Degenerativediskdisease:assessment

ofchangesinvertebralbodymarrowwithMRimaging.Radiology1988Jan;166(1Pt1):193‐199.

(7)LoiselP,CoteP.HandbookofWorkDisability,chapter5.In:LoiselP,AnemaJR,editors..1sted.

NewYork:SpringerScience+BusinessMedia;2013.p.59.

(8)BieringK,HjollundNH,LundT.Methodsinmeasuringreturntowork:acomparisonof

measuresofreturntoworkfollowingtreatmentofcoronaryheartdisease.JOccupRehabil2013

Sep;23(3):400‐405.

(9)Hogg‐JohnsonS,vanderVeldeG,CarrollLJ,HolmLW,CassidyJD,GuzmanJ,etal.Theburden

anddeterminantsofneckpaininthegeneralpopulation:resultsoftheBoneandJointDecade

2000‐2010TaskForceonNeckPainandItsAssociatedDisorders.Spine(PhilaPa1976)2008Feb

15;33(4Suppl):S39‐51.

(10)LuimeJJ,KoesBW,HendriksenIJ,BurdorfA,VerhagenAP,MiedemaHS,etal.Prevalenceand

incidenceofshoulderpaininthegeneralpopulation;asystematicreview.ScandJRheumatol

2004;33(2):73‐81.

REFERENCES

Page|78

(11)InternationalAssociationofPain.IASPTaskForceonTaxonomy.Definitionofpain.1994;

Availableat:http://www.iasp‐pain.org/Education/Content.aspx?ItemNumber=1698.Accessed

09/10,2018.

(12)MisailidouV,MalliouP,BenekaA,KaragiannidisA,GodoliasG.Assessmentofpatientswith

neckpain:areviewofdefinitions,selectioncriteria,andmeasurementtools.JChiroprMed2010

Jun;9(2):49‐59.

(13)CarrollLJ,Hogg‐JohnsonS,vanderVeldeG,HaldemanS,HolmLW,CarrageeEJ,etal.Course

andprognosticfactorsforneckpaininthegeneralpopulation:resultsoftheBoneandJointDecade

2000‐2010TaskForceonNeckPainandItsAssociatedDisorders.Spine(PhilaPa1976)2008Feb

15;33(4Suppl):S75‐82.

(14)CarrollLJ,Hogg‐JohnsonS,CoteP,vanderVeldeG,HolmLW,CarrageeEJ,etal.Courseand

prognosticfactorsforneckpaininworkers:resultsoftheBoneandJointDecade2000‐2010Task

ForceonNeckPainandItsAssociatedDisorders.Spine(PhilaPa1976)2008Feb15;33(4

Suppl):S93‐100.

(15)vanderWindtDA,KoesBW,BoekeAJ,DevilleW,DeJongBA,BouterLM.Shoulderdisordersin

generalpractice:prognosticindicatorsofoutcome.BrJGenPract1996Sep;46(410):519‐523.

(16)WintersJC,SobelJS,GroenierKH,ArendzenJH,Meyboom‐deJongB.Thelong‐termcourseof

shouldercomplaints:aprospectivestudyingeneralpractice.Rheumatology(Oxford)1999

Feb;38(2):160‐163.

(17)LoiselP,DurandM‐,BertheletteD,VézinaN,BarilR,GagnonD,etal.DisabilityPrevention.

NewParadigmfortheManagementofOccupationalBackPain.DisManageHealthOutcomes

2001;9(7):351‐360.

(18)PrinsR.HandbookofWorkDisability,chapter1.In:LoiselP,AnemaJR,editors.Handbookof

WorkDisability.1sted.NewYork:SpringerScience+BusinessMedia;2013.p.3.

(19)WaddellG,BurtonAK.Isworkgoodforyourhealthandwell‐being?2006.

(20)CoteP,vanderVeldeG,CassidyJD,CarrollLJ,Hogg‐JohnsonS,HolmLW,etal.Theburdenand

determinantsofneckpaininworkers:resultsoftheBoneandJointDecade2000‐2010TaskForce

onNeckPainandItsAssociatedDisorders.Spine(PhilaPa1976)2008Feb15;33(4Suppl):S60‐74.

REFERENCES

Page|79

(21)CoteP,KristmanV,VidmarM,VanEerdD,Hogg‐JohnsonS,BeatonD,etal.Theprevalenceand

incidenceofworkabsenteeisminvolvingneckpain:acohortofOntariolost‐timeclaimants.J

ManipulativePhysiolTher2009Feb;32(2Suppl):S219‐26.

(22)FlachsEM,EriksenL,Koch,M,B,RydJT,DibbaE,Skov‐EttrupL,etal.NationalInstituteof

PublicHealth,UniversityofSouthernDenmark.TheBurdenofDiseasesinDenmark–diseases.

Copenhagen:DanishHealthAuthority;2015.2015;2.0(01/09).

(23)KuijpersT,vanderWindtDA,vanderHeijdenGJ,TwiskJW,VergouweY,BouterLM.A

predictionruleforshoulderpainrelatedsickleave:aprospectivecohortstudy.BMCMusculoskelet

Disord2006Dec6;7:97‐2474‐7‐97.

(24)RudbeckM,JensenSL,FonagerK.Arthroscopicsubacromialdecompressionandpredictorsof

long‐termsickleavebenefitandpermanentbenefits.JShoulderElbowSurg2013Sep;22(9):1167‐

1172.

(25)LundT,KivimakiM,LabriolaM,VilladsenE,ChristensenKB.Usingadministrativesickness

absencedataasamarkeroffuturedisabilitypension:theprospectiveDREAMstudyofDanish

privatesectoremployees.OccupEnvironMed2008Jan;65(1):28‐31.

(26)LabriolaM,LundT.Self‐reportedsicknessabsenceasariskmarkeroffuturedisability

pension.ProspectivefindingsfromtheDWECS/DREAMstudy1990‐2004.IntJMedSci2007May

17;4(3):153‐158.

(27)KivimakiM,FormaP,WikstromJ,HalmeenmakiT,PenttiJ,ElovainioM,etal.Sicknessabsence

asariskmarkeroffuturedisabilitypension:the10‐townstudy.JEpidemiolCommunityHealth

2004Aug;58(8):710‐711.

(28)SvendsenSW,FrostP,JensenLD.Timetrendsinsurgeryfornon‐traumaticshoulderdisorders

andpostoperativeriskofpermanentworkdisability:anationwidecohortstudy.ScandJRheumatol

2012Feb;41(1):59‐65.

(29)TompaE.HandbookofWorkDisability,chapter4.HandbookofWorkDisability.1sted.New

York:SpringerScience+BusinessMedia;2013.p.43.

REFERENCES

Page|80

(30)GrossDP,DeshpandeS,MiciakM,WernerEL,RenemanMF,BuchbinderR.HandbookofWork

Disability,chapter24.HandbookofWorkDisability.1sted.NewYork:SpringerScience+Business

Media;2013.p.389.

(31)WilkinsonR,MarmotM.Socialdeterminantsofhealth:thesolidfacts(WHO).2003(2):18.

(32)Costa‐BlackKM,FeuersteinM,LoiselP.HandbookofWorkDisability,chapter6.Handbookof

WorkDisability.1sted.NewYork:SpringerScience+BusinessMedia;2013.p.71.

(33)SaundersSL,NedelecB.Whatworkmeanstopeoplewithworkdisability:ascopingreview.J

OccupRehabil2014Mar;24(1):100‐110.

(34)LoiselP,BuchbinderR,HazardR,KellerR,ScheelI,vanTulderM,etal.Preventionofwork

disabilityduetomusculoskeletaldisorders:thechallengeofimplementingevidence.JOccup

Rehabil2005Dec;15(4):507‐524.

(35)ChoiBC,PakAW.Multidisciplinarity,interdisciplinarityandtransdisciplinarityinhealth

research,services,educationandpolicy:1.Definitions,objectives,andevidenceofeffectiveness.

ClinInvestMed2006Dec;29(6):351‐364.

(36)CoteP,WongJJ,SuttonD,ShearerHM,MiorS,RandhawaK,etal.Managementofneckpain

andassociateddisorders:AclinicalpracticeguidelinefromtheOntarioProtocolforTrafficInjury

Management(OPTIMa)Collaboration.EurSpineJ2016Jul;25(7):2000‐2022.

(37)WaddellG,BurtonAK,KendallNAS.VocationalRehabilitation.Whatworks,forwhom,and

when?2013(01/09).

(38)IndahlA,VelundL,ReikeraasO.Goodprognosisforlowbackpainwhenleftuntampered.A

randomizedclinicaltrial.Spine(PhilaPa1976)1995Feb15;20(4):473‐477.

(39)IndahlA,HaldorsenEH,HolmS,ReikerasO,UrsinH.Five‐yearfollow‐upstudyofacontrolled

clinicaltrialusinglightmobilizationandaninformativeapproachtolowbackpain.Spine(PhilaPa

1976)1998Dec1;23(23):2625‐2630.

(40)HagenEM,EriksenHR,UrsinH.Doesearlyinterventionwithalightmobilizationprogram

reducelong‐termsickleaveforlowbackpain?Spine(PhilaPa1976)2000Aug1;25(15):1973‐

1976.

REFERENCES

Page|81

(41)vanVilsterenM,vanOostromSH,deVetHC,FrancheRL,BootCR,AnemaJR.Workplace

interventionstopreventworkdisabilityinworkersonsickleave.CochraneDatabaseSystRev2015

Oct5;(10):CD006955.doi(10):CD006955.

(42)CullenKL,IrvinE,CollieA,ClayF,GensbyU,JenningsPA,etal.EffectivenessofWorkplace

InterventionsinReturn‐to‐WorkforMusculoskeletal,Pain‐RelatedandMentalHealthConditions:

AnUpdateoftheEvidenceandMessagesforPractitioners.JOccupRehabil2017Feb21.

(43)KamperSJ,ApeldoornAT,ChiarottoA,SmeetsRJ,OsteloRW,GuzmanJ,etal.Multidisciplinary

biopsychosocialrehabilitationforchroniclowbackpain:Cochranesystematicreviewandmeta‐

analysis.BMJ2015Feb18;350:h444.

(44)RinaldoU,SelanderJ.Returntoworkaftervocationalrehabilitationforsick‐listedworkers

withlong‐termback,neckandshoulderproblems:Afollow‐upstudyoffactorsinvolved.Work

2016Sep27;55(1):115‐131.

(45)BussieresAE,PetersonC,TaylorJA.Diagnosticimagingguidelineformusculoskeletal

complaintsinadults‐anevidence‐basedapproach‐part2:upperextremitydisorders.JManipulative

PhysiolTher2008Jan;31(1):2‐32.

(46)BussieresAE,TaylorJA,PetersonC.Diagnosticimagingpracticeguidelinesformusculoskeletal

complaintsinadults‐anevidence‐basedapproach‐part3:spinaldisorders.JManipulativePhysiol

Ther2008Jan;31(1):33‐88.

(47)NationalDanishPatientRegistry.Availableat:

http://www.esundhed.dk/sundhedsregistre/LPR/Sider/LPR04_Tabel.aspx.Accessed11/23,2017.

(48)ChristensenAI,DavidsenM,JuelK.TheNationalHealthProfile2013.2014March5th2014:37.

(49)JensenHAR,DavidsenM,ChristensenAI.TheNationalHealthProfile2017.2018March6th

2018:41.

(50)HealthSciencesCentre,Winnipeg:DiagnosticImagingAnnualReport2015‐2016.2016:22.

(51)CanadianInstituteforHealthInformation:MedicalImaginginCanada,2007.2008:19‐91‐92.

REFERENCES

Page|82

(52)PetersenL,BirkelundR,AmmentorpJ,Schiøttz‐ChristensenB."AnMRIrevealsthetruthabout

myback":aqualitativestudyaboutpatients'expectationsandattitudestowardthevalueofMRIin

theassessmentofbackpain.EurJPersCentHealthc2016;4(3):453‐458.

(53)LakerSR,ConcannonLG.Radiologicevaluationoftheneck:areviewofradiography,

ultrasonography,computedtomography,magneticresonanceimaging,andotherimaging

modalitiesforneckpain.PhysMedRehabilClinNAm2011Aug;22(3):411‐28,vii‐viii.

(54)SembranoJN,YsonSC,KanuOC,BramanJP,SantosER,HarrisonAK,etal.Neck‐shoulder

crossover:howoftendoneckandshoulderpathologymasqueradeaseachother?AmJOrthop

(BelleMeadNJ)2013Sep;42(9):E76‐80.

(55)ChoCH,LeonardDA.Theimportanceoffluencyofterminologyintheapplicabilityofimaging

findings.SpineJ2016Jan1;16(1):61‐62.

(56)KolstadF,MyhrG,KvistadKA,NygaardOP,LeivsethG.Degenerationandheightofcervical

discsclassifiedfromMRIcomparedwithpreciseheightmeasurementsfromradiographs.EurJ

Radiol2005Sep;55(3):415‐420.

(57)MannE,PetersonCK,HodlerJ.Degenerativemarrow(modic)changesoncervicalspine

magneticresonanceimagingscans:prevalence,inter‐andintra‐examinerreliabilityandlinktodisc

herniation.Spine(PhilaPa1976)2011Jun15;36(14):1081‐1085.

(58)MiyazakiM,HongSW,YoonSH,MorishitaY,WangJC.Reliabilityofamagneticresonance

imaging‐basedgradingsystemforcervicalintervertebraldiscdegeneration.JSpinalDisordTech

2008Jun;21(4):288‐292.

(59)ParkHJ,KimSS,LeeSY,ParkNH,ChungEC,RhoMH,etal.ApracticalMRIgradingsystemfor

cervicalforaminalstenosisbasedonobliquesagittalimages.BrJRadiol2013

May;86(1025):20120515.

(60)XuC,DingZH,XuYK.Comparisonofcomputedtomographyandmagneticresonanceimaging

intheevaluationoffacettropismandfacetarthrosisindegenerativecervicalspondylolisthesis.

GenetMolRes2014May30;13(2):4102‐4109.

REFERENCES

Page|83

(61)FuMC,WebbML,BuerbaRA,NewayWE,BrownJE,TrivediM,etal.Comparisonofagreement

ofcervicalspinedegenerativepathologyfindingsinmagneticresonanceimagingstudies.SpineJ

2016Jan1;16(1):42‐48.

(62)WernerEL,CoteP.Lowbackpainanddeterminantsofsicknessabsence.EurJGenPract

2009;15(2):74‐79.

(63)BültmannU,BrouwerS.HandbookofWorkDisability,chapter10.In:LoiselP,AnemaJ,

editors..1sted.NewYork:SpringerScience+BusinessMedia;2013.p.149.

(64)CancelliereC,DonovanJ,StochkendahlMJ,BiscardiM,AmmendoliaC,MyburghC,etal.Factors

affectingreturntoworkafterinjuryorillness:bestevidencesynthesisofsystematicreviews.

ChiroprManTherap2016Sep8;24(1):32‐016‐0113‐z.eCollection2016.

(65)DesmeulesF,BraenC,LamontagneM,DionneCE,RoyJS.Determinantsandpredictorsof

absenteeismandreturn‐to‐workinworkerswithshoulderdisorders.Work2016Sep

27;55(1):101‐113.

(66)BuchbinderR,GreenS,YoudJM.Corticosteroidinjectionsforshoulderpain.Cochrane

DatabaseSystRev2003;(1)(1):CD004016.

(67)DunsfordA,KumarS,ClarkeS.Integratingevidenceintopractice:useofMcKenzie‐based

treatmentformechanicallowbackpain.JMultidiscipHealthc2011;4:393‐402.

(68)RolvingN,ChristiansenDH,AndersenLL,SkotteJ,YlinenJ,JensenOK,etal.Effectofstrength

traininginadditiontogeneralexerciseintherehabilitationofpatientswithnon‐specificneckpain.

Arandomizedclinicaltrial.EurJPhysRehabilMed2014Dec;50(6):617‐626.

(69)BoltonJE,HumphreysBK,vanHedelHJ.Validityofweeklyrecallratingsofaveragepain

intensityinneckpainpatients.JManipulativePhysiolTher2010Oct;33(8):612‐617.

(70)JordanA,MannicheC,MosdalC,HindsbergerC.TheCopenhagenNeckFunctionalDisability

Scale:astudyofreliabilityandvalidity.JManipulativePhysiolTher1998Oct;21(8):520‐527.

(71)Armijo‐OlivoS,WoodhouseLJ,SteenstraIA,GrossDP.PredictivevalueoftheDASHtoolfor

predictingreturntoworkofinjuredworkerswithmusculoskeletaldisordersoftheupper

extremity.OccupEnvironMed2016Dec;73(12):807‐815.

REFERENCES

Page|84

(72)KarranEL,McAuleyJH,TraegerAC,HillierSL,GrabherrL,RussekLN,etal.Canscreening

instrumentsaccuratelydeterminepooroutcomeriskinadultswithrecentonsetlowbackpain?A

systematicreviewandmeta‐analysis.BMCMed2017Jan19;15(1):13‐016‐0774‐4.

(73)LintonSJ,HalldenK.Canwescreenforproblematicbackpain?Ascreeningquestionnairefor

predictingoutcomeinacuteandsubacutebackpain.ClinJPain1998Sep;14(3):209‐215.

(74)StapelfeldtCM,JensenC,AndersenNT,FletenN,NielsenCV.Validationofsickleavemeasures:

self‐reportedsickleaveandsicknessbenefitdatafromaDanishnationalregistercomparedto

multipleworkplace‐registeredsickleavespellsinaDanishmunicipality.BMCPublicHealth2012

Aug15;12:661‐2458‐12‐661.

(75)KovacsFM,AbrairaV,RoyuelaA,CorcollJ,AlegreL,TomasM,etal.Minimumdetectableand

minimalclinicallyimportantchangesforpaininpatientswithnonspecificneckpain.BMC

MusculoskeletDisord2008Apr10;9:43‐2474‐9‐43.

(76)PoolJJ,OsteloRW,HovingJL,BouterLM,deVetHC.Minimalclinicallyimportantchangeofthe

NeckDisabilityIndexandtheNumericalRatingScaleforpatientswithneckpain.Spine(PhilaPa

1976)2007Dec15;32(26):3047‐3051.

(77)deWetHCW,TerweeCBea.Chapter5:Reliability.MeasurementinMedicineCambridge:

CambridgeUniversityPress;2011.p.96‐126.

(78)Bojsen‐MoellerF.Chapter8:Hvirvelsoejlen(TheSpine).BevaegeapparatetsAnatomi.12thed.

Copenhagen:MunksgaardDanmark;2001.p.89.

(79)FardonDF,WilliamsAL,DohringEJ,MurtaghFR,GabrielRothmanSL,SzeGK.Lumbardisc

nomenclature:version2.0:RecommendationsofthecombinedtaskforcesoftheNorthAmerican

SpineSociety,theAmericanSocietyofSpineRadiologyandtheAmericanSocietyof

Neuroradiology.SpineJ2014Nov1;14(11):2525‐2545.

(80)JacobsLJ,ChenAF,KangJD,LeeJY.ReliableMagneticResonanceImagingBasedGrading

SystemforCervicalIntervertebralDiscDegeneration.AsianSpineJ2016Feb;10(1):70‐74.

(81)KalichmanL,SuriP,GuermaziA,LiL,HunterDJ.Facetorientationandtropism:associations

withfacetjointosteoarthritisanddegeneratives.Spine(PhilaPa1976)2009Jul15;34(16):E579‐

85.

REFERENCES

Page|85

(82)KangY,LeeJW,KohYH,HurS,KimSJ,ChaiJW,etal.NewMRIgradingsystemforthecervical

canalstenosis.AJRAmJRoentgenol2011Jul;197(1):W134‐40.

(83)KimS,LeeJW,ChaiJW,YooHJ,KangY,SeoJ,etal.ANewMRIGradingSystemforCervical

ForaminalStenosisBasedonAxialT2‐WeightedImages.KoreanJRadiol2015Nov‐

Dec;16(6):1294‐1302.

(84)MaattaJH,KarppinenJ,PaananenM,BowC,LukKD,CheungKM,etal.RefinedPhenotypingof

ModicChanges:ImagingBiomarkersofProlongedSevereLowBackPainandDisability.Medicine

(Baltimore)2016May;95(22):e3495.

(85)ShimJH,ParkCK,LeeJH,ChoiJW,LeeDC,KimDH,etal.AcomparisonofangledsagittalMRI

andconventionalMRIinthediagnosisofherniateddiscandstenosisinthecervicalforamen.Eur

SpineJ2009Aug;18(8):1109‐1116.

(86)NouriA,MartinAR,MikulisD,FehlingsMG.Magneticresonanceimagingassessmentof

degenerativecervicalmyelopathy:areviewofstructuralchangesandmeasurementtechniques.

NeurosurgFocus2016Jun;40(6):E5.

(87)WiltseLL,BergerPE,McCullochJA.Asystemforreportingthesizeandlocationoflesionsin

thespine.Spine(PhilaPa1976)1997Jul1;22(13):1534‐1537.

(88)YochumTR,RoweLJ.Chapter10:ArthriticDisorders.EssentialsofSkeletalRadiology

Baltimore:LippincottWilliams&Wilkins;2004.p.951‐1134.

(89)Guggenmoos‐HolzmannI.Howreliablearechance‐correctedmeasuresofagreement?Stat

Med1993Dec15;12(23):2191‐2205.

(90)LandisJR,KochGG.Themeasurementofobserveragreementforcategoricaldata.Biometrics

1977Mar;33(1):159‐174.

(91)WasiakR,YoungAE,RoesslerRT,McPhersonKM,vanPoppelMN,AnemaJR.Measuringreturn

towork.JOccupRehabil2007Dec;17(4):766‐781.

(92)Biering‐SorensenS,MollerA,StoltenbergCD,HolmJW,SkovPG.Thereturn‐to‐workprocess

ofindividualssick‐listedbecauseofwhiplash‐associateddisorder:athree‐yearfollow‐upstudyina

REFERENCES

Page|86

Danishcohortoflong‐termsicknessabsentees.BMCPublicHealth2014Feb4;14:113‐2458‐14‐

113.

(93)JensenOK,CallesenJ,NielsenMG,EllingsenT.Reproducibilityoftenderpointexaminationin

chroniclowbackpainpatientsasmeasuredbyintraraterandinter‐raterreliabilityandagreement:

avalidationstudy.BMJOpen2013Feb26;3(2):10.1136/bmjopen‐2012‐002532.Print2013.

(94)WolfeF,SmytheHA,YunusMB,BennettRM,BombardierC,GoldenbergDL,etal.The

AmericanCollegeofRheumatology1990CriteriafortheClassificationofFibromyalgia.Reportof

theMulticenterCriteriaCommittee.ArthritisRheum1990Feb;33(2):160‐172.

(95)DanishMinistryofEmployment:DanishActonActiveEmploymentPolicy.2016;Availableat:

https://www.retsinformation.dk/forms/r0710.aspx?id=184891.Accessed27/09,2018.

(96)EmploymentDMO.DanishSicknessBenefitAct.2014;Availableat:

https://www.retsinformation.dk/Forms/R0710.aspx?id=161736#Kap8.Accessed06/01,2017.

(97)MollLT,JensenOK,Schiottz‐ChristensenB,StapelfeldtCM,ChristiansenDH,NielsenCV,etal.

ReturntoWorkinEmployeesonSickLeaveduetoNeckorShoulderPain:ARandomizedClinical

TrialComparingMultidisciplinaryandBriefInterventionwithOne‐YearRegister‐BasedFollow‐Up.

JOccupRehabil2017Aug23.

(98)Wynne‐JonesG,CowenJ,JordanJL,UthmanO,MainCJ,GlozierN,etal.Absencefromworkand

returntoworkinpeoplewithbackpain:asystematicreviewandmeta‐analysis.OccupEnviron

Med2014Jun;71(6):448‐456.

(99)ValentinGH,PilegaardMS,VaegterHB,RosendalM,OrtenbladL,VaeggemoseU,etal.

Prognosticfactorsfordisabilityandsickleaveinpatientswithsubacutenon‐malignantpain:a

systematicreviewofcohortstudies.BMJOpen2016Jan6;6(1):e007616‐2015‐007616.

(100)SteenstraIA,BusseJW,Hogg‐JohnsonS.HandbookofWorkDisability,chapter16.In:Loisel

P,AnemaJR,editors.HandbookofWorkDisability.1sted.NewYork:SpringerScience+Business

Media;2013.p.255.

(101)LindellO,JohanssonSE,StrenderLE.Predictorsofstablereturn‐to‐workinnon‐acute,non‐

specificspinalpain:lowtotalpriorsick‐listing,highselfpredictionandyoungage.Atwo‐year

prospectivecohortstudy.BMCFamPract2010Jul20;11:53‐2296‐11‐53.

REFERENCES

Page|87

(102)CarstensenTB,FinkP,OernboelE,KaschH,JensenTS,FrostholmL.SickLeavewithin5Years

ofWhiplashTraumaPredictsRecovery:AProspectiveCohortandRegister‐BasedStudy.PLoSOne

2015Jun22;10(6):e0130298.

(103)HockingsRL,McAuleyJH,MaherCG.Asystematicreviewofthepredictiveabilityofthe

OrebroMusculoskeletalPainQuestionnaire.Spine(PhilaPa1976)2008Jul1;33(15):E494‐500.

(104)LintonSJ,BoersmaK.Earlyidentificationofpatientsatriskofdevelopingapersistentback

problem:thepredictivevalidityoftheOrebroMusculoskeletalPainQuestionnaire.ClinJPain2003

Mar‐Apr;19(2):80‐86.

(105)BergstromG,HagbergJ,BuschH,JensenI,BjorklundC.Predictionofsicknessabsenteeism,

disabilitypensionandsicknesspresenteeismamongemployeeswithbackpain.JOccupRehabil

2014Jun;24(2):278‐286.

(106)SteenstraIA,MunhallC,IrvinE,OranyeN,PassmoreS,VanEerdD,etal.SystematicReview

ofPrognosticFactorsforReturntoWorkinWorkerswithSubAcuteandChronicLowBackPain.J

OccupRehabil2017Sep;27(3):369‐381.

(107)FletenN,JohnsenR,FordeOH.Lengthofsickleave‐whynotaskthesick‐listed?Sick‐listed

individualspredicttheirlengthofsickleavemoreaccuratelythanprofessionals.BMCPublicHealth

2004Oct12;4:46‐2458‐4‐46.

(108)vonCelsingAS,SvardsuddK,WallmanT.Predictingreturntoworkamongsickness‐certified

patientsingeneralpractice:propertiesoftwoassessmenttools.UpsJMedSci2014

Aug;119(3):268‐277.

(109)EngblomM,AlexandersonK,RudebeckCE.Characteristicsofsick‐listingcasesthat

physiciansconsiderproblematic‐‐analysesofwrittencasereports.ScandJPrimHealthCare

2009;27(4):250‐255.

(110)LintonSJ,BoersmaK.Earlyidentificationofpatientsatriskofdevelopingapersistentback

problem:thepredictivevalidityoftheOrebroMusculoskeletalPainQuestionnaire.ClinJPain2003

Mar‐Apr;19(2):80‐86.

REFERENCES

Page|88

(111)JensenOK,Stengaard‐PedersenK,JensenC,NielsenCV.Predictionmodelforunsuccessful

returntoworkafterhospital‐basedinterventioninlowbackpainpatients.BMCMusculoskelet

Disord2013Apr19;14:140‐2474‐14‐140.

(112)JensenOK,NielsenCV,SorensenJS,Stengaard‐PedersenK.Type1Modicchangeswasa

significantriskfactorfor1‐yearoutcomeinsick‐listedlowbackpainpatients:anestedcohort

studyusingmagneticresonanceimagingofthelumbarspine.SpineJ2014Nov1;14(11):2568‐

2581.

(113)HestbaekL,RasmussenC,Leboeuf‐YdeC.Financialcompensationandvocationalrecovery:a

prospectivestudyofsecondarycareneckandbackpatients.ScandJRheumatol2009Nov‐

Dec;38(6):481‐487.

(114)JensenC,JensenOK,ChristiansenDH,NielsenCV.One‐yearfollow‐upinemployeessick‐

listedbecauseoflowbackpain:randomizedclinicaltrialcomparingmultidisciplinaryandbrief

intervention.Spine(PhilaPa1976)2011Jul1;36(15):1180‐1189.

(115)AnemaJR,SteenstraIA,BongersPM,deVetHC,KnolDL,LoiselP,etal.Multidisciplinary

rehabilitationforsubacutelowbackpain:gradedactivityorworkplaceinterventionorboth?A

randomizedcontrolledtrial.Spine(PhilaPa1976)2007Feb1;32(3):291‐8;discussion299‐300.

(116)LambeekLC,vanMechelenW,KnolDL,LoiselP,AnemaJR.Randomisedcontrolledtrialof

integratedcaretoreducedisabilityfromchroniclowbackpaininworkingandprivatelife.BMJ

2010Mar16;340:c1035.

(117)LintonSJ,BoersmaK,TraczykM,ShawW,NicholasM.EarlyWorkplaceCommunicationand

ProblemSolvingtoPreventBackDisability:ResultsofaRandomizedControlledTrialAmongHigh‐

RiskWorkersandTheirSupervisors.JOccupRehabil2016Jun;26(2):150‐159.

(118)HaldorsenEM,KronholmK,SkouenJS,UrsinH.Multimodalcognitivebehavioraltreatmentof

patientssicklistedformusculoskeletalpain:arandomizedcontrolledstudy.ScandJRheumatol

1998;27(1):16‐25.

(119)AasdahlL,PapeK,VasseljenO,JohnsenR,GismervikS,JensenC,etal.EffectsofInpatient

MulticomponentOccupationalRehabilitationversusLessComprehensiveOutpatientRehabilitation

onSomaticandMentalHealth:SecondaryOutcomesofaRandomizedClinicalTrial.JOccupRehabil

2017Sep;27(3):456‐466.

REFERENCES

Page|89

(120)LytsyP,CarlssonL,AnderzenI.Effectivenessoftwovocationalrehabilitationprogrammesin

womenwithlong‐termsickleaveduetopainsyndromeormentalillness:1‐yearfollow‐upofa

randomizedcontrolledtrial.JRehabilMed2017Jan31;49(2):170‐177.

(121)MarholdC,LintonSJ,MelinL.Acognitive‐behavioralreturn‐to‐workprogram:effectsonpain

patientswithahistoryoflong‐termversusshort‐termsickleave.Pain2001Mar;91(1‐2):155‐163.

(122)RemeSE,TveitoTH,HarrisA,LieSA,GrasdalA,IndahlA,etal.CognitiveInterventionsand

NutritionalSupplements(TheCINSTrial):ARandomizedControlled,MulticenterTrialComparinga

BriefInterventionWithAdditionalCognitiveBehavioralTherapy,SealOil,andSoyOilforSick‐

ListedLowBackPainPatients.Spine(PhilaPa1976)2016Oct15;41(20):1557‐1564.

(123)AndersenLN,Juul‐KristensenB,SorensenTL,HerborgLG,RoesslerKK,SogaardK.Efficacyof

TailoredPhysicalActivityorChronicPainSelf‐ManagementProgrammeonreturntoworkforsick‐

listedcitizens:A3‐monthrandomisedcontrolledtrial.ScandJPublicHealth2015Nov;43(7):694‐

703.

(124)AndersenLN,Juul‐KristensenB,SorensenTL,HerborgLG,RoesslerKK,SogaardK.Longer

termfollow‐uponeffectsofTailoredPhysicalActivityorChronicPainSelf‐Management

Programmeonreturn‐to‐work:Arandomizedcontrolledtrial.JRehabilMed2016Nov

11;48(10):887‐892.

(125)LoiselP,AbenhaimL,DurandP,EsdaileJM,SuissaS,GosselinL,etal.Apopulation‐based,

randomizedclinicaltrialonbackpainmanagement.Spine(PhilaPa1976)1997Dec

15;22(24):2911‐2918.

(126)BultmannU,ShersonD,OlsenJ,HansenCL,LundT,KilsgaardJ.Coordinatedandtailored

workrehabilitation:arandomizedcontrolledtrialwitheconomicevaluationundertakenwith

workersonsickleaveduetomusculoskeletaldisorders.JOccupRehabil2009Mar;19(1):81‐93.

(127)BrendbekkenR,EriksenHR,GrasdalA,HarrisA,HagenEM,TangenT.ReturntoWorkin

PatientswithChronicMusculoskeletalPain:MultidisciplinaryInterventionVersusBrief

Intervention:ARandomizedClinicalTrial.JOccupRehabil2017Mar;27(1):82‐91.

(128)MyhreK,MarchandGH,LeivsethG,KellerA,Bautz‐HolterE,SandvikL,etal.Theeffectof

work‐focusedrehabilitationamongpatientswithneckandbackpain:arandomizedcontrolledtrial.

Spine(PhilaPa1976)2014Nov15;39(24):1999‐2006.

REFERENCES

Page|90

(129)AasdahlL,PapeK,VasseljenO,JohnsenR,GismervikS,HalsteinliV,etal.EffectofInpatient

MulticomponentOccupationalRehabilitationVersusLessComprehensiveOutpatientRehabilitation

onSicknessAbsenceinPersonswithMusculoskeletal‐orMentalHealthDisorders:ARandomized

ClinicalTrial.JOccupRehabil2018Mar;28(1):170‐179.

(130)JensenC,JensenOK,NielsenCV.Sustainabilityofreturntoworkinsick‐listedemployeeswith

low‐backpain.Two‐yearfollow‐upinarandomizedclinicaltrialcomparingmultidisciplinaryand

briefintervention.BMCMusculoskeletDisord2012Aug25;13:156‐2474‐13‐156.

(131)BartysS,FrederiksenP,BendixT,BurtonK.Systeminfluencesonworkdisabilityduetolow

backpain:Aninternationalevidencesynthesis.HealthPolicy2017Aug;121(8):903‐912.

(132)MinistryofEmployment:TheDanishHealthInformationLaw.1996April1996;286:1‐3.

(133)MatsumotoM,FujimuraY,SuzukiN,NishiY,NakamuraM,YabeY,etal.MRIofcervical

intervertebraldiscsinasymptomaticsubjects.JBoneJointSurgBr1998Jan;80(1):19‐24.

(134)KuijperB,BeelenA,vanderKallenBF,NolletF,LycklamaANijeholtGJ,deVisserM,etal.

InterobserveragreementonMRIevaluationofpatientswithcervicalradiculopathy.ClinRadiol

2011Jan;66(1):25‐29.

(135)KoS,ChoiW,ChaeS.Comparisonofinter‐andintra‐observerreliabilityamongthethree

classificationsystemsforcervicalspinalcanalstenosis.EurSpineJ2017Jun13.

(136)LindholdtL,LabriolaM,NielsenCV,HorsbolTA,LundT.Sequenceanalysistoassesslabour

marketparticipationfollowingvocationalrehabilitation:anobservationalstudyamongpatients

sick‐listedwithlowbackpainfromarandomisedclinicaltrialinDenmark.BMJOpen2017Jul

20;7(7):e015661‐2016‐015661.

(137)LintonSJ,HalldenK,HellsingA‐.TheReliabilityofSelf‐reportedSickAbsenteeism:APilot

Study.ScandJBehavTher1995;24:145‐150.

(138)PedersenP,NielsenCV,JensenOK,JensenC,LabriolaM.Employmentstatusfiveyearsaftera

randomisedcontrolledtrialcomparingmultidisciplinaryandbriefinterventioninemployeeson

sickleaveduetolowbackpain.ScandJPublicHealth2018May;46(3):383‐388.

REFERENCES

Page|91

(139)FrancheRL,BarilR,ShawW,NicholasM,LoiselP.Workplace‐basedreturn‐to‐work

interventions:optimizingtheroleofstakeholdersinimplementationandresearch.JOccupRehabil

2005Dec;15(4):525‐542.

(140)AmmendoliaC,CassidyD,SteenstaI,SoklaridisS,BoyleE,EngS,etal.Designingaworkplace

return‐to‐workprogramforoccupationallowbackpain:aninterventionmappingapproach.BMC

MusculoskeletDisord2009Jun9;10:65‐2474‐10‐65.

(141)BuchbinderR,vanTulderM,ObergB,CostaLM,WoolfA,SchoeneM,etal.Lowbackpain:a

callforaction.Lancet2018Jun9;391(10137):2384‐2388.

(142)VermeulenSJ,AnemaJR,SchellartAJ,KnolDL,vanMechelenW,vanderBeekAJ.A

participatoryreturn‐to‐workinterventionfortemporaryagencyworkersandunemployedworkers

sick‐listedduetomusculoskeletaldisorders:resultsofarandomizedcontrolledtrial.JOccup

Rehabil2011Sep;21(3):313‐324.

(143)HagenEM,OdelienKH,LieSA,EriksenHR.Addingaphysicalexerciseprogrammetobrief

interventionforlowbackpainpatientsdidnotincreasereturntowork.ScandJPublicHealth2010

Nov;38(7):731‐738.

APPENDICES

Page|92

Appendices

Appendix1Study1

Appendix2Study2

Appendix3Study3

Appendix4TheevaluationmanualusedinStudy2

Appendix5PrevalencetableofpositiveMRIfindingsinStudy2

Appendix6ExemptionfromtheEthicalCommittee,Study2

Appendix7OverviewofRCTsaimedatRTW

Study 11

Vol.:(0123456789)1 3

J Occup Rehabil DOI 10.1007/s10926-017-9727-9

Return to Work in Employees on Sick Leave due to Neck or Shoulder Pain: A Randomized Clinical Trial Comparing Multidisciplinary and Brief Intervention with One-Year Register-Based Follow-Up

Line Thorndal Moll1,2,3 · Ole Kudsk Jensen3 · Berit Schiøttz-Christensen4 · Christina Malmose Stapelfeldt1,2 · David Høyrup Christiansen5 · Claus Vinther Nielsen1,2 · Merete Labriola1,2

© The Author(s) 2017. This article is an open access publication

the BI group during the 1 year of follow-up. Results showed a statistically non significant tendency towards a lower rate of RTW in the MDI group than in the BI group (adjusted HR = 0.84, 95% CI 0.54, 1.31). There were no statistically significant differences in secondary outcomes between the MDI and BI groups. Conclusion The brief and the multidis-ciplinary interventions performed equally with respect to both primary and secondary outcomes. The added focus on RTW in the multidisciplinary group did not improve RTW rates in this group.

Keywords Return to work · Sick leave · Neck pain · Shoulder pain · Rehabilitation

AbbreviationsBI Brief interventionCNFDS Copenhagen Neck Functional Disability ScaleDASH Disabilities of the arm, shoulder and handGP General practitionerHR Hazard rateIQR Inter quartile rangeITT Intention to treatMDI Multidisciplinary interventionMCIC Minimally clinically important changeOR Odds ratioÖMPQ Örebro Musculoskeletal Pain QuestionnaireRCT Randomized controlled trialRTW Return to work

Background

Musculoskeletal disorders are widely recognized as common causes of disability and sick leave [1–3]. Among muscu-loskeletal disorders, neck and shoulder pain are common,

Abstract Purpose The aim of this study was to evaluate the effect of a multidisciplinary intervention (MDI) com-pared to a brief intervention (BI) with respect to return to work (RTW), pain and disability in workers on sick leave because of neck or shoulder pain. Methods 168 study par-ticipants with sickness absence for 4–16 weeks due to neck or shoulder pain were enrolled in a hospital-based clinical study and randomized to either MDI or BI. The primary outcome was RTW obtained by a national registry on public transfer payments. Secondary outcomes were self-reported pain and disability levels. One-year follow-up RTW rates were estimated by Cox proportional hazard regression adjusted for gender, age, sick leave prior to inclusion, part-time sick leave and clinical diagnosis. Secondary outcomes were analysed using logistic and linear regression analysis for pain and disability, respectively. Results In the MDI group, 50 participants (59%) experienced four or more con-tinuous weeks of RTW while 48 (58%) returned to work in

* Line Thorndal Moll [email protected]

1 DEFACTUM, Central Denmark Region, P.P. Oerums Gade 11, bygn. 1B, 8000 Aarhus C, Denmark

2 Section of Clinical Social Medicine and Rehabilitation, Department of Public Health, Aarhus University, P.P. Oerums Gade 9-11, bygn. 1B, 8000 Aarhus C, Denmark

3 Spine Centre, Diagnostic Centre, Silkeborg Regional Hospital, Falkevej 1-3, 8600 Silkeborg, Denmark

4 Spine Centre of Southern Denmark, Hospital Lillebaelt Middelfart and Institute of Regional Health Research, University of Southern Denmark, Oestre Hougvej 55, 5500 Middelfart, Denmark

5 Department of Occupational Medicine, University Research Clinic, Regional Hospital West Jutland, Gl. Landevej 61, 7400 Herning, Denmark

http://orcid.org/0000-0003-4132-0310

http://crossmark.crossref.org/dialog/?doi=10.1007/s10926-017-9727-9&domain=pdf

J Occup Rehabil

1 3

though prevalence estimates tend to differ across studies, primarily due to differences in case definitions. In the gen-eral population, estimates of the 12-month prevalence are 2–11% for activity-limiting neck pain [3] and 5–47% for shoulder pain [4]. Among workers, 11–14% report activity limitation due to neck pain [5]. Worldwide, neck pain is the fourth most common reason for years lived with disability [1] and in Denmark, 16% of days on sick leave in 2015 were caused by neck pain [6]. Not only does sickness absence imply costs for society [7]; the potentially detrimental impli-cations to the individual are also well described [8] as are the association between long-term sick leave and the increased risk of premature withdrawal from the labour market [9–11]. In accordance with the above, sickness absence as a focus of political concern is well established [7].

Over the past decades, the challenge of rehabilitating sickness absentees with musculoskeletal disorders has been addressed [12, 13]. Populations suffering from low back pain (LBP) are well represented in the body of literature; studies on sub-acute LBP offer moderate evidence on the positive effect of multidisciplinary rehabilitation in terms of improving disability and reducing sickness absence [14]. For chronic LBP, it is suggested based on moderate evidence that multidisciplinary rehabilitation is superior to physiotherapy with respect to return to work (RTW), pain and disability and superior to usual care with respect to pain and disabil-ity [12]. A recent review on back, neck and shoulder pain found positive RTW outcomes in studies using a multidis-ciplinary approach and the assignment of case managers [15]. The involvement of workplaces has also been proven beneficial [13–17]. In Denmark, the work outcomes of dif-ferent studies have not been unanimous. Thus, a study from 2009 suggested positive outcomes on RTW and duration of sick leave when applying coordinated, tailored work reha-bilitation in workers with musculoskeletal disorders [18]. In this study [18] however, only 19% of the participants had neck pain. More recent Danish studies evaluating work outcomes found positive effect of tailored physical activ-ity after 3 months [19], an effect which was however not maintained at 11 months of follow-up [20]. Like in the study by Bültmann et al. [18], these studies included participants with both back, neck and shoulder pain [19, 20]. So while studies investigating pain and disability in neck and shoulder participants are common, participants with these pain loca-tions often constitute only a minority in studies investigat-ing work outcomes. Regarding shoulder disorders, the work outcomes of a Danish study evaluating physiotherapy exer-cises and occupational medical assistance are awaited [21]. In a review on the effect of different treatments for impinge-ment syndrome1 [24] only few studies reported RTW as an

outcome; neither of these fulfilled the authors’ criteria for “high quality study” and neither of these evaluated the effect of multidisciplinary interventions. Accordingly, how to reha-bilitate workers on sick leave with neck and shoulder pain is a question yet to be addressed [23, 24].

Aims

The aim of this study was to evaluate the effect of a multidis-ciplinary intervention (MDI) compared to a brief interven-tion (BI) with respect to RTW, pain and disability in workers on sick leave due to neck or shoulder pain.

Methods

Design and Participants

The study was conducted as a randomized clinical trial at The Spine Centre, Silkeborg Regional Hospital, Denmark. General practitioners (GPs), physiotherapists and chiro-practors in the primary sector from seven municipalities received written information about the study to display in their waiting rooms. GPs were encouraged to refer patients that fulfilled the inclusion criteria. The flow of participants is presented in Fig. 1. From May 2009 through January 2014, 328 people were screened for eligibility. Inclusion criteria were: Age 18–60 years, the primary reason for sick leave being pain in the neck, shoulders or upper thoracic region, fluency in Danish and self-reported full- or part-time sick leave for 4–12 weeks. The duration of sick leave was a prag-matic choice: patients with sick leave shorter than 4 weeks were considered to have a fairly good chance of returning to work spontaneously whereas an upper limit was chosen because longer sick leaves are associated with lower RTW chances [15]. The criterion was however changed to 4–16 weeks shortly after starting the project due to low number of referrals from GPs. Exclusion criteria were: Continuing or progressive signs of nerve root impingement implying plans for operation, known substance abuse or pregnancy, neck-, back- or shoulder-surgery within the last year, other specific or serious musculoskeletal disease and primary psychiatric disorder. Participants with comorbid psychiatric disorder considered to be in clinical remission were not excluded. 168 participants were included and completed the 1-year follow-up (Fig. 1).

1 The most common shoulder disorder is subacromial impingement syndrome (pain arising from the mechanical impingement of soft tis-sue structures (tendons and bursa) between the humeral head and the

coracoacromial joint). It covers a range of pathologies from inflam-mation of the tendon and bursa to degeneration and ultimately rupture of the tendons [22].

Footnote 1 (continued)

J Occup Rehabil

1 3

Randomization

An overview of the interventions is presented in Table 1. At the first visit to the Spine Centre, all participants were offered participation in the study and their written informed consent was provided. At this baseline visit, all participants were examined by a rheumatologist and a physiotherapist. Two weeks later, the first follow-up visit with the physi-otherapist took place (primarily with the aim of ensuring adherence to the given exercises and making adjustments if needed) and simple randomization was carried out. A

secretary made a telephone call to an externally placed computer and thereby allocated the participants to brief or multidisciplinary intervention.

Multidisciplinary Intervention Group (MDI)

In addition to the clinical examination at baseline, the par-ticipants in the MDI group had a case manager assigned who primarily had the responsibility of coordinating communica-tion among stakeholders. Individual meetings between par-ticipants and their respective case managers were scheduled

Fig. 1 Participant flow diagram

Table 1 Contacts with the Spine Centre in the two intervention groups

Baseline: clinical examination and advice

2 weeks: follow-up at physiotherapist randomization

3–4 weeks: 1st meeting with case manager

3–6 weeks: infor-mation on MRI findings

12 weeks: follow-up at the physio-therapist

MDI: RTW plan and meetings with case manager

MDI group + + + + + +BI group + + − + + −

J Occup Rehabil

1 3

within 1–2 weeks after the randomization visit (Table 1). At this first meeting, they went through a standardized interview on work history, private life, pain and disability. With the aim of full or partial RTW a rehabilitation plan was made. The participant met with the case manager once or repeatedly depending on need and progress. If relevant, consultations with a psychologist were arranged (n = 12). The role as case manager was held by a social worker, a spe-cialist of clinical social medicine or an occupational thera-pist. The case manager discussed relevant matters at regular team conferences not attended by the participant. Present at these team conferences were the rheumatologist, the three case managers, the physiotherapists and in relevant cases the psychologist. At the time of the study, the idea of drawing upon the expertise of the multidisciplinary team along with the access to psychologist appointments when needed was an attempt to encompass all relevant biopsychosocial con-siderations regarding the RTW process of the MDI group.

In 19 cases, roundtable discussions were arranged at the workplace and in three additional cases the case manager phoned the employer of the participant. The workplace involvement was optional and decided by the participants who in many cases wished to keep their health problems secret to their employers. This can be ascribed to the Dan-ish Health Information Law [25]. In context of the Danish flexicurity model where employers have wide opportunities to “fire and hire”, the purpose of this law is to prevent dis-crimination of workers due to health issues. The law ensures that employers only under special circumstances are enti-tled to know about the health conditions of their employees. If RTW was considered impossible, an alternative plan to remain in work was made, for instance by jobs supported by the social system. To ensure a standardized multidisciplinary intervention, the entire team received 1–2 h of supervision every 2 months from a general practitioner specialized in cognitive therapy. Cases were closed when the participants returned to work and the MDI support could not proceed after this was achieved. If RTW was deemed impossible, a meeting was arranged with the municipality’s social service centre.

All Participants

Regardless of intervention group, all participants were examined by a rheumatologist and a physiotherapist at their first visit to The Spine Centre (Table 1). These two health care providers were both blinded to the subsequent random allocation to intervention groups. The rheuma-tologist recorded the medical history and performed a thorough clinical examination. This was followed by information about the usually limited correlation between pain and imaging of the cervical spine [26] and about

aerobic exercise being beneficial for pain. Furthermore, the participants were reassured that normal daily activities, work and exercise would not be harmful. This approach was based on the findings by Indahl et al., suggesting the beneficence of reducing fear and maintaining physical activity [27]. Magnetic resonance imaging (MRI) of the cervical spine was performed except when shoulder prob-lems were the obvious cause of pain. Participants with clinical signs of radiculopathy were informed about the good spontaneous prognosis and about the possibility of surgery in case of no improvement. If necessary, lab tests were done, and analgesic treatment was adjusted. The diagnostic accuracy of musculoskeletal ultrasound imag-ing has been reported moderate to high [28] for which rea-son participants suspected for primary shoulder disorders had ultrasound imaging of the shoulder performed. In case of ultrasonographic inflammation, a steroid injection was offered (n = 2; one in each group) [29]. The physiotherapist examined all participants in a standardised manner includ-ing neuromuscular testing and measuring isometric neck strength, except in those with radiculopathy. The latter were tested by the McKenzie method. This method is sup-ported by moderate evidence for LBP [30] and widely used in NP though less well documented. It was none-the-less used to help participants control their pain.

At a follow-up visit approximately 3–6 weeks after enrolment (Table 1), the rheumatologist explained the MRI findings in a reassuring way and all participants had their last follow-up visit with the physiotherapist 12 weeks after their first visit.

To ensure coordination between stakeholders, cop-ies of the medical records were sent to the participant, the GP and the municipal social services responsible for reimbursement of sick leave compensation. Except for the described follow-up visits with the rheumatologist and the physiotherapist (Table 1), those allocated to the brief inter-vention group were offered no further intervention. They were advised to resume work when possible. If in need for advice or additional treatment, they were recommended to consult their GP.

Nested in this randomized controlled trial (RCT) was a smaller RCT testing the effect of two different exer-cise programs, which has been reported previously [31]. Enrolled in the nested RCT were 83 of the participants with nonspecific neck pain who were randomly allocated to one of two home-based exercise groups. Some were allocated to a general physical activity group (GPA) (n = 40) and the remaining participants (n = 43) were allo-cated to a group doing both general physical exercise AND specific strength training (SST). The primary outcome of this trial was pain intensity, and no difference was found between the groups.

J Occup Rehabil

1 3

Context

In Denmark in the years from 2009 to 2014, when the study was conducted, a worker falling ill had the right for sick leave benefits for 52 weeks. If criteria for extending the 52 weeks were not fulfilled, only some citizens could receive other social transfer benefits from their municipality [32] since the right to other transfer benefits depended—among other things—on the spouse’s income.

Variables and Outcomes

Baseline data were collected from a questionnaire completed by the participants prior to the clinical examination. This questionnaire covered socio-demographic factors, health issues, disability and work-related factors. Pain intensity was measured on an 11-point numeric ranking scale from 0 (no pain) to 10 (worst imaginable pain) [33], and psychosocial dimensions of pain were measured by the Örebro Musculo-skeletal Pain Questionnaire (ÖMPQ) [34, 35]. For partici-pants with primary shoulder disorder, disability was meas-ured by disabilities of the arm, shoulder and hand (DASH) [36] and for the rest of the study population by the Copenha-gen Neck Functional Disability Scale (CNFDS) [37]. Mental health was measured by the SF-36 mental health subscale [38]. The duration of sick leave was dichotomized at a cutoff value of 12 weeks [14, 39].

The primary outcome RTW was defined as the first period of four consecutive weeks of self-support for individuals who were self-supporting before their sick leave. For those individuals who held jobs supported by the social system prior to their sick leave, four consecutive weeks of return to this job was defined as RTW. The choice of 4 weeks was explained by the wish to ensure comparability with the previously conducted LBP study [40] at The Spine Cen-tre. RTW and sick leave compensation data were attainable from the Danish Register for Evaluation of Marginalisation (DREAM)—a national registry on public social and health-related benefits registered on a weekly basis and adminis-tered by The Danish Ministry of Employment. Since July 1991, all Danish citizens having received any type of social or health-related benefits are registered in DREAM. The source of income is registered by means of a 3-digit code and ordered hierarchically [41].

One year after inclusion, postal questionnaires were sent to the participants. These questionnaires provided data on the secondary outcomes: changes in pain level (numeric ranking scale) [33] and disability level as measured by the CNFDS [37] (participants with primary shoulder disorder excluded from the analysis). Changes in pain levels were calculated by subtracting 1-year follow-up pain levels from baseline levels. Due to a large proportion of non-respond-ers leaving only nine participants with primary shoulder

disorder with follow-up disability measures (DASH) (MDI n = 1, BI n = 8), this outcome measure was omitted.

Analyses

Prior to the study, a power calculation was carried out based on the assumption that there would be a 15% difference in RTW between the groups. Given a power (1-β) of 70%, a sample size of 85 in each group was required (two-sided α = 0.05).

The distribution of baseline characteristics was presented after excluding missing values. For those variables not ful-filling the assumption of normality, median values and inter quartile ranges (IQR) were reported.

The time to RTW during 1 year of follow-up was esti-mated using survival analysis (Kaplan–Meier). RTW rates in the two groups were compared using Cox proportional hazard regression. Competing risks were defined as death and emigration. The assumption of proportional hazards was assessed and confirmed using log-minus-log plots (not shown). Crude and adjusted hazard ratios (HR) were cal-culated according to the intention to treat (ITT) principle with adjustment for known prognostic variables for RTW: sex, age (≤40/>40 years) and duration of sick leave (≤/>12 weeks) [39, 42] as well as part-time sick leave (yes/no) and clinical diagnoses (non-specific neck pain, radiculopathy, primary shoulder disorder).

For the secondary outcome pain; two-way scatter plots (not shown) could not justify the assumption of linearity between follow-up and baseline scores. Furthermore, a minimally clinically important change (MCIC) defined as ≥2 points (yes/no) [43, 44] was considered relevant and hence, data on pain intensity changes were dichotomized according to this. Logistic regression analysis estimating crude odds ratio (OR) and adjusted OR (gender, age groups (≤/>40 years) and baseline pain intensity) was performed. To our knowledge there is no consensus on a cutoff value for a MCIC for the secondary outcome disability as measured by CNFDS. And as the model for linear regression adjust-ing for gender, age groups and baseline CNFDS values was checked and accepted by diagnostic plots of the residuals, this outcome measure was calculated by linear regression analysis. Positive values of β0 reflect increased disability levels. Due to the risk of over-fitted models in the secondary outcome analyses, the number of potential confounders was reduced to three variables compared to five in the analyses of time to RTW.

For those individuals lost to follow-up on the secondary outcomes (n = 89), a non-response analysis of responders versus non-responders was performed comparing the allo-cation to intervention groups, achievement of the primary outcome and all baseline characteristics (data not shown). These analyses were performed using an unpaired T test,

J Occup Rehabil

1 3

Fisher’s exact test, Chi squared test (χ2) or the Wilcoxon rank-sum test, depending on type and distribution of the var-iable. The statistical software package STATA 13.1 was used for analysis and p values <0.05 were regarded as statistically significant. Statistical analyses were performed by research-ers outside the hospital and independently from those who gave the interventions.

Ethical Approval

All participants signed informed consent. The study is reg-istered at Current Controlled Trials, ISRCTN51739408. It was approved by The Danish Data Protection Agency (J. No. 2007-58-0010) and by the regional ethical committee (M-20090027).

Results

Study Population

After inclusion of 168 participants, the study was closed in January 2014 primarily due to changes in the data manage-ment unit making it impossible to continue the same method of randomization, secondarily due to recruitment difficulties.

Table 2 shows baseline characteristics of the study par-ticipants. The access to register data on the primary outcome allowed for 100% follow-up, whereas a considerable dropout rate (n = 89) was seen on the secondary outcomes gathered by questionnaires. A non-response analysis revealed no dif-ferences between responders and non-responders regard-ing allocation to intervention group, achievement of the primary outcome or any of the baseline variables except for allocation to exercise groups (among responders, more participants were in the general exercise group compared to non-responders).

Primary Outcome: RTW

For the primary outcome RTW the total number of events was 98 and the total follow-up time was 5492 weeks. At baseline, four individuals had already experienced the event RTW and were therefore excluded from the analysis as were an additional number of 14 individuals due to missing val-ues in one or more of the variables that we adjusted for (Fig. 1). Thus, 164 and 150 individuals were included in the crude and adjusted analyses, respectively. None of the participants were excluded due to competing risks (death and emigration).

The proportion of participants in the two groups still on sick leave is illustrated in Fig. 2. In the MDI group, 50 par-ticipants (59%) returned to work during the 1-year follow-up while 48 participants (58%) in the BI group experienced

the event. The crude HR was 0.94 (95% CI 0.63; 1.41) and the adjusted HR was 0.84 (95% CI 0.54; 1.31). The median time to RTW was 44 weeks (IQR 18–52) in the MDI group and 32 weeks (IQR 12–52) in the BI group (p = 0.83). The median duration of the MDI intervention was 4.6 months (IQR 3.3–7.4) and 3 months (IQR 3–3) in the BI group.

Secondary Outcomes: Pain Intensity and Disability

The median pain score reduction was 2 units in both groups (MDI group IQR 0; 3. BI group IQR 0; 5). However, when comparing the MDI to the BI group, the crude OR for a clinically important pain reduction ≥2 points was 1.10 (95% CI 0.54; 2.26). Adjustment for gender, age-groups and base-line pain intensity yielded an OR of 1.18 (95% CI 0.56; 2.48). For disability, linear regression analysis yielded crude estimates of a non-significant CNFDS beta coefficient of 1.37 (95% CI −1.91; 4.64) points higher in the MDI group compared to the BI group at 1-year follow-up. After adjust-ment for gender, age-groups and baseline level of disability the coefficient changed, however still non-significantly, to 1.09 points (95% CI −2.26; 4.45) at follow-up.

Discussion

Two main findings from this study warrant exploration. One is the lack of difference between a multidisciplinary intervention compared to a brief intervention with respect to RTW, pain and disability in sick-listed workers with neck or shoulder pain. The other is the discouraging fact that less than 60% of the study population returned to work during the first year.

As for the lack of difference between the MDI and the BI; the study conducted by Bültmann et al. [18] reported a significant improvement in RTW status at 1-year follow-up in a Danish study on sick-listed workers with musculoskel-etal disorders. Some notable differences in interventions and study populations may explain why we did not find similar results. The involvement of workplaces was a key element as 45% of participants in the intervention group had round-table discussions arranged at the workplace in Bültmann’s study. Also, a maximum duration of the intervention equiva-lent to 3 months was settled on. The mean duration of sick leave prior to the intervention was approximately 6 weeks [18]. In the present study, only 19 (22%) in the MDI group had roundtable discussions arranged, sick leave was longer and the median duration of the MDI was 4.6 months (IQR 3.3–7.4).

Another possible explanation for the lack of differ-ence between the MDI and the BI groups could be the similarities of the clinical services provided by the rheu-matologist and the physiotherapist. The approach to the

J Occup Rehabil

1 3

participants in both groups was based on a non-injury model as inspired by Indahl et al. [27] and Hagen et al. [45]. Both Myhre et al. [46] and Brendbekken et al. [47] had the same similarities between control and intervention groups. They did not find differences in RTW outcomes either. The reassurance provided by thorough examinations and explanations from two clinicians dedicated to spine

disorders should probably not be underestimated—a point which has also previously been stated [40, 45].

Less than 60% of the participants returned to work dur-ing follow-up which is inferior to the results from similar studies describing RTW for more than 70% of their par-ticipants [18, 40, 46], and the modest RTW results warrant exploration.

Table 2 Baseline characteristics of the two intervention groups

CNFDS Copenhagen Neck Functional Disability Scale, SF-36 short-form 36, ÖMPQ Örebro Musculoskel-etal Pain Questionnaire, DASH disabilities of the arm, shoulder & hand, IQR inter quartile range

n Brief intervention (n = 83)

Multidisciplinary intervention (n = 85)

Age, mean (SD) in years 168 42.2 (10.39) 40.0 (9.17)≤40 years, n (%) 168 38 (45.8) 45 (52.9)Female gender, n (%) 168 56 (67.5) 59 (69.4)Marital status, single n (%) 160 10 (13.0) 11 (13.3)Education, n (%) None, brief courses, other 155 25 (32.9) 21 (26.6) Skilled workers, education <3 years 36 (47.4) 44 (55.7) Education ≥3 years 15 (19.7) 14 (17.7)

Current smoker, n (%) 161 39 (50.0) 39 (47.0)Pain intensity (0–10) last week, median (IQR) 158 7 (6 ; 8) 7 (5 ; 8)CNFDS score, mean (SD) 132 19.0 (5.53) 19.0 (5.51)DASH score, mean (SD) 20 64.2 (51.2) 53.5 (38.3)ÖMPQ score, n (%) <90 161 7 (9.0) 5 (6.0) 90–105 9 (11.5) 14 (16.9) >105 62 (79.5) 64 (77.1)

SF-36 mental health subscale, mean (SD) 161 60.3 (18.9) 58.0 (21.0)Musculoskeletal comorbidity n (%) Low back pain 148 21 (29.6) 25 (32.5) Leg pain 147 9 (12.7) 17 (22.4)

Physician’s diagnoses, n (%) Non-specific neck pain 168 50 (60.2) 57 (67.1) Radiculopathy 19 (22.9) 21 (24.7) Primary shoulder disorder 14 (16.9) 7 (8.2)

Sick leave duration, n (%) ≤12 weeks 168 60 (72.3) 66 (77.7)

Previous sick leaves due to neck/shoulder pain 0 previous sick leaves 158 38 (48.7) 30 (37.5) 1–2 previous sick leaves 12 (15.4) 25 (31.2) 3–4 previous sick leaves 13 (16.7) 14 (17.5) >4 previous sick leaves 15 (19.2) 11 (13.8)

Is your pain caused by your work, n (%) Answer “no” 146 32 (45.7) 35 (46.0)

Current part-time sick leave, n (%) Answer “yes” 154 13 (17.6) 27 (33.8)

Exercise group, n (%) General exercise 168 29 (34.9) 28 (32.9) Specific exercises 31 (37.4) 31 (36.5) Exercises for radiculopathy 23 (27.7) 26 (30.6)

J Occup Rehabil

1 3

In the randomized trials by Jensen et al. [40] and Myhre et al. [46], a multidisciplinary intervention much similar to the one used in the present study was offered; both reported successful RTW for approximately 70% of their participants. Differences in pain location might be an explanation, as only LBP patients were included in the former [40] whereas in the latter [46], both neck and back pain patients were included; however, the distribution of pain locations is not presented. In the above mentioned study by Bültmann [18], only 12% of the study population had neck pain. Recent studies by Andersen et al. [19, 20] found promising RTW results of tai-lored physical activity at 3 month follow-up but these were not maintained at 11 month follow-up; neither the tailored physical activity program nor the pain self management program improved RTW compared to the reference group. The outcome measure in these studies was RTW status (yes/no) and although different from the present four consecutive weeks of RTW [48], the proportion of participants return-ing to work was closer to our results than in the studies by Jensen et al. [40] and Myhre et al. [46]. A possible explana-tion could be a larger proportion of the study population suffering from neck and upper extremity pain. However, this information was not provided by Andersen et al.

While involvement of workplaces should be a key element in the process of RTW [15–17, 39, 42], our RTW results were notably poorer compared to the previously published LBP study by Jensen et al., although the rehabilitation pro-grams were very similar [40]. In contrast to the previously mentioned studies [18, 40, 46], the present study included only participants with neck and shoulder pain. This may lead to considerations of the possibility of a poorer RTW prognosis for people with neck and shoulder pain in general compared to people with LBP.

Apart from the pain location, the present study population also had baseline characteristics that might have influenced

the process of returning to work. At inclusion, the partici-pants were troubled by severe pain intensity and consider-able psychosocial impact of their pain (ÖMPQ) (Table 2). Both high pain intensity scores and ÖMPQ scores >90 have been shown to predict future sick leave [15, 34, 49] and thus may have affected RTW outcomes. At baseline, almost half of the study population had musculoskeletal comorbidity and approximately one-third had ≥3 previous sick leaves. Both factors are known to have negative prognostic value with respect to RTW [15, 42].

In studies with RTW outcomes similar to ours, explana-tions may also in part be found in baseline characteristics. Thus, in Andersen et al.’s studies [19, 20] where approxi-mately 60% returned to work, more than half of the study population had previous sick leave episodes. In the study by Brendbekken et al. [47], the mean duration of sick leave prior to inclusion was 147 days. Both number of previous sick leaves and current sick leave duration are negative prog-nostic factors for RTW [15].

The study had several strengths. One was the randomized design which ensured comparability between the two groups with the exception of a larger proportion of part-time sick-listed participants in the MDI group compared to the BI group. However, this variable was adjusted for. Second, we had 100% follow up on the primary outcome thus eliminat-ing the risk of attrition bias. A third strength of the study was the ITT analysis. The fact that baseline clinical examinations were carried out blindedly before randomization was con-sidered a further strength.

The study also had some limitations. First, given the nature of the interventions, it was not possible to perform all interventions in a blinded manner. A second potential weak-ness was the recruitment of participants. The GPs received written information about the study with encouragement to refer patients on sick leave due to neck and shoulder pain. They may have referred only high-risk patients because they would consider it more cost-effective to treat low-risk patients in primary care. Whether GPs have had such con-siderations is unknown. Although the referral pattern was similar to the LBP study [40] this aspect needs to be taken into account when considering generalizability of the study.

Third, participants with sickness absence lasting 4–16 weeks were included although longer sickness spells constitute an independent risk factor of not returning to work [15, 39]. An exploratory analysis to test if a more rigid inclusion criterion on sick leave (4–8 weeks) would have yielded different results was performed; this was not the case (data not shown). Fourth, the number of non-responders on the secondary outcomes was substantial (n = 89) introducing a potential risk of selection bias in the assessment of secondary outcomes. Non-response analysis (data not shown) did not show any statistically significant differences between responders and non-responders with

0.00

0.25

0.50

0.75

1.00

0 10 20 30 40 50Weeks

Intervention = brief Intervention = multidisciplinary

Fig. 2 Reduction in proportion of participants on sick leave during follow-up (Kaplan–Meier)

J Occup Rehabil

1 3

respect to intervention groups, RTW or any of the base-line variables. Only the allocation to exercise groups dif-fered between responders and non-responders. This was a difference not suspected to have biased the estimates of the secondary outcomes. Nor do we, to the best of our knowledge, consider the nested RCT [31] to threaten the estimation of the results in the present study. We base this on the equal distribution of exercise groups between the BI and the MDI groups (Table 2), and the fact that the participants had equal pain improvements following the exercise programmes in the nested RCT [31].

The access to register data on RTW allowed for 100% fol-low-up on the primary outcome and the validity of DREAM has previously been demonstrated [41]. A fifth limitation was that appraisal of register data revealed minor inconsist-encies at baseline between self-reported and register-based sick leave status. According to register data, 15 participants did not fulfil the inclusion criteria of sick leave ≥4 weeks. These participants were equally distributed between inter-vention groups and tentative per protocol analysis exclud-ing these participants did not alter the results (adjusted HR = 0.70. 95% CI 0.44–1.12). It cannot be ruled out that the ITT analysis might introduce a minor degree of non-differentiated information bias. But this does not change the overall estimates of RTW and apart from maintaining the strength of randomization, the ITT analysis also displays high external validity since self-reported sick leave status is the only accessible information on the day of inclusion.

Sixth, the time spent on the MDI warrants considera-tion. Due to the setup of the study, participants in the MDI group waited 1–2 weeks after randomization before receiv-ing the part of the intervention that differed from the BI group. Meanwhile, time at risk began at the day of randomi-zation for both groups. Remembering the poor prognosis associated with prolonged sick leave [9–11, 15, 39] this was inexpedient but unfortunately unavoidable. Seventh, due to the sample size, there is approximately 30% risk of type 2 errors, i.e. a risk of overlooking an actual difference between the MDI and the BI intervention. We do not, however, con-sider power problems to explain the lack of difference, but rather characteristics of the population and intervention as described above.

Finally, only a minority of participants in the MDI expe-rienced workplace involvement. In the latest review on workplace interventions, Cullen et al. present strong evi-dence on the positive work outcomes when applying multi-domain interventions orchestrated from the workplace [17] and it could be argued that workplace involvement should have been mandatory. As previously described, this was not possible, because the majority of participants preferred to keep their health problems secret to their employers. As described, this discretion regarding health issues is rooted the Danish Health Information Law [25]. Whether a stronger

focus on workplace involvement could have improved the results in the MDI group cannot be ruled out.

On the macro level, the “economic climate” is known to potentially affect sickness absence [7]. Our choice of out-come measure was constricted to four consecutive weeks of self-support, alternatively four consecutive weeks of holding a job supported by the social system. But since the study was performed during a period of economic recession in Den-mark, exploratory analyses were performed allowing for the outcome RTW to be also 4 weeks of unemployment benefits and State Education Fund Grants (both reflecting readiness to return to work). These analyses still did not show signifi-cant differences in RTW between the groups but increased the HR in favor of the MDI (data not shown). Rather than interpreting the increased HR as the results of a successful MDI intervention, this merely reflects the termination of employment for some of the MDI participants. The com-bination of general economic recession and an intervention lasting several weeks may have contributed to the loss of jobs for some of the MDI participants.

In conclusion, no difference was found in RTW rates between the BI and the MDI group. Nor were there any dif-ferences in follow-up pain and disability between the groups. We do however assume that the evidence on the effect of multidisciplinary interventions in LBP [12, 14] and other musculoskeletal disorders [15, 17] is transferable to neck and shoulder pain. For clinical practice, several studies over the years e.g., [27, 40, 45–47] have suggested efficacy of a brief clinical intervention based on a non-injury approach with a focus of diminishing fear and restoring/maintaining normal daily activities. Add-on of a multidisciplinary inter-vention including a case manager as in the current study does not seem to improve RTW outcomes. Rather, evidence suggests the necessary involvement of workplaces.

Another implication for clinical practice derives from the above recognition: There is not only a need for efficient RTW interventions but also for increased focus on prevent-ing sickness absence, i.e. how do clinicians identify patients at high risk of sickness absence? Feleus et al. recently pub-lished a study identifying three different trajectories for sick-ness absence (low, intermediate and high risk) in patients presenting in primary care with complaints of the arm, neck and shoulder [50]. They also identified bio-psycho-social variables associated with these trajectories. For whiplash-associated disorders, a tool predicting both chronic disability and full recovery has been developed [51, 52]. For neck pain however, current evidence does not support clinical use of neither prognostic nor prescriptive clinical prediction rules [53].

Better understanding of the prognostic factors and devel-opment of clinical prediction rules regarding RTW outcomes in neck and shoulder pain are suggested as future focus areas in research.

J Occup Rehabil

1 3

Acknowledgements The authors thank all participants and stake-holders for their cooperation.

Funding The Danish Rheumatism Foundation, Helga og Peter Korn-ing Foundation, Aase og Ejnar Danielsen Foundation, Aarhus Univer-sity, Tryg Foundation.

Author Contributions LTM: data management, analysis of data, primary responsible for the manuscript. CMS: assistance with analy-sis of data, critical appraisal of the manuscript. BS and ML: critical appraisal of the manuscript. DH: planning of the study, providing the clinical intervention (physiotherapist), critical appraisal of the manu-script. OKJ: planning of the study, providing the clinical intervention (physician), critical appraisal of the manuscript, power calculation. CVN: planning of the study, critical appraisal of the manuscript.

Compliance with Ethical Standards

Conflict of interest Line Thorndal Moll, Christina Malmose Sta-pelfeldt, Berit Schiöttz-Christensen, David Høyrup-Christiansen, Ole Kudsk Jensen, Claus Vinther Nielsen, Merete Labriola declare that they have no conflict of interests.

Ethical Approval All procedures followed were in accordance with the ethical standards of the regional ethical committee and with the Helsinki Declaration of 1975, as revised in 2000. Informed consent was obtained from all participants included in the study.

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea-tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appro-priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

References

1. Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2163–2196.

2. Waddell G. The back pain revolution. 1st ed. Edinburgh: Churchill Livingstone; 1999.

3. Hogg-Johnson S, van der Velde G, Carroll LJ, Holm LW, Cassidy JD, Guzman J, et al. The burden and determinants of neck pain in the general population: results of the Bone and Joint Decade 2000–2010 Task Force on Neck Pain and Its Associated Disor-ders. Spine (Phila Pa 1976). 2008;33(4 Suppl):S39–S51.

4. Luime JJ, Koes BW, Hendriksen IJ, Burdorf A, Verhagen AP, Miedema HS, et al. Prevalence and incidence of shoulder pain in the general population; a systematic review. Scand J Rheumatol. 2004;33(2):73–81.

5. Haldeman S, Carroll L, Cassidy JD. Findings from the bone and joint decade 2000 to 2010 task force on neck pain and its associ-ated disorders. J Occup Environ Med. 2010;52(4):424–427.

6. Flachs EM, Eriksen L, Koch M,B, Ryd JT, Dibba E, Skov-Ettrup L, et al. National Institute of Public Health, University of South-ern Denmark. The Burden of Diseases in Denmark—diseases. Copenhagen: Danish Health Authority; 2015.

7. Whitaker SC. The management of sickness absence. Occup Envi-ron Med. 2001;58(6):420–424 (quiz 424,410).

8. Waddell G, Burton AK. Is work good for your health and well-being? London: TSO; 2006.

9. Lund T, Kivimaki M, Labriola M, Villadsen E, Christensen KB. Using administrative sickness absence data as a marker of future disability pension: the prospective DREAM study of Danish pri-vate sector employees. Occup Environ Med. 2008;65(1):28–31.

10. Labriola M, Lund T. Self-reported sickness absence as a risk marker of future disability pension. Prospective findings from the DWECS/DREAM study 1990–2004. Int J Med Sci. 2007;4(3):153–158.

11. Kivimaki M, Forma P, Wikstrom J, Halmeenmaki T, Pentti J, Elovainio M, et al. Sickness absence as a risk marker of future disability pension: the 10-town study. J Epidemiol Community Health. 2004;58(8):710–711.

12. Kamper SJ, Apeldoorn AT, Chiarotto A, Smeets RJ, Ostelo RW, Guzman J, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350:h444. doi:10.1136/bmj.h444.

13. Waddell G, Burton AK, Kendall NAS. Vocational rehabilitation. What works, for whom, and when? London: TSO; 2013.

14. Karjalainen K, Malmivaara A, van Tulder M, Roine R, Jauhiainen M, Hurri H, et al. Multidisciplinary biopsychosocial rehabilitation for subacute low back pain among working age adults. Cochrane Database Syst Rev 2003;(2)(2):CD002193.

15. Rinaldo U, Selander J. Return to work after vocational reha-bilitation for sick-listed workers with long-term back, neck and shoulder problems: a follow-up study of factors involved. Work. 2016;55(1):115–131.

16. Loisel P, Abenhaim L, Durand P, Esdaile JM, Suissa S, Gosselin L, et al. A population-based, randomized clinical trial on back pain management. Spine (Phila Pa 1976). 1997;22(24):2911–2918.

17. Cullen KL, Irvin E, Collie A, Clay F, Gensby U, Jennings PA, et al. Effectiveness of workplace interventions in return-to-work for musculoskeletal, pain-related and mental health conditions: an update of the evidence and messages for practitioners. J Occup Rehabil. 2017. doi:10.1007/s10926-016-9690-x.

18. Bultmann U, Sherson D, Olsen J, Hansen CL, Lund T, Kilsgaard J. Coordinated and tailored work rehabilitation: a randomized con-trolled trial with economic evaluation undertaken with workers on sick leave due to musculoskeletal disorders. J Occup Rehabil. 2009;19(1):81–93.

19. Andersen LN, Juul-Kristensen B, Sorensen TL, Herborg LG, Roessler KK, Sogaard K. Efficacy of Tailored Physical Activity or Chronic Pain Self-Management Programme on return to work for sick-listed citizens: a 3-month randomised controlled trial. Scand J Public Health. 2015;43(7):694–703.

20. Andersen LN, Juul-Kristensen B, Sorensen TL, Herborg LG, Roessler KK, Sogaard K. Longer term follow-up on effects of Tailored Physical Activity or Chronic Pain Self-Management Programme on return-to-work: a randomized controlled trial. J Rehabil Med. 2016;48(10):887–892.

21. Svendsen SW, Christiansen DH, Haahr JP, Andrea LC, Frost P. Shoulder function and work disability after decompression surgery for subacromial impingement syndrome: a randomised controlled trial of physiotherapy exercises and occupational medical assistance. BMC Musculoskelet Disord. 2014;15:215. doi:10.1186/1471-2474-15-215.

22. Michener LA, McClure PW, Karduna AR. Anatomical and bio-mechanical mechanisms of subacromial impingement syndrome. Clin Biomech. 2003;18(5):369–379.

23. Karjalainen K, Malmivaara A, van Tulder M, Roine R, Jauhiainen M, Hurri H, et al. Multidisciplinary biopsychosocial rehabilitation for neck and shoulder pain among working age adults. Cochrane Database Syst Rev 2003;(2):CD002194.

24. Faber E, Kuiper JI, Burdorf A, Miedema HS, Verhaar JA. Treat-ment of impingement syndrome: a systematic review of the effects

http://creativecommons.org/licenses/by/4.0/


http://dx.doi.org/10.1136/bmj.h444

http://dx.doi.org/10.1007/s10926-016-9690-x

http://dx.doi.org/10.1186/1471-2474-15-215

J Occup Rehabil

1 3

on functional limitations and return to work. J Occup Rehabil. 2006;16(1):7–25.

25. Ministry of Employment: The Danish Health Information Law. 1996;286:1–3; https://www.retsinformation.dk/forms/r0710.aspx?id=81200.

26. Nordin M, Carragee EJ, Hogg-Johnson S, Weiner SS, Hurwitz EL, Peloso PM, et al. Assessment of neck pain and its associated disorders: results of the Bone and Joint Decade 2000–2010 Task Force on Neck Pain and Its Associated Disorders. J Manipulative Physiol Ther. 2009;32(2 Suppl):S117–S140.

27. Indahl A, Velund L, Reikeraas O. Good prognosis for low back pain when left untampered. A randomized clinical trial. Spine (Phila Pa 1976). 1995;20(4):473–477.

28. Henderson RE, Walker BF, Young KJ. The accuracy of diagnostic ultrasound imaging for musculoskeletal soft tissue pathology of the extremities: a comprehensive review of the literature. Chiropr Man Ther. 2015;23:31. doi:10.1186/s12998-015-0076-5.

29. Buchbinder R, Green S, Youd JM. Corticosteroid injections for shoulder pain. Cochrane Database Syst Rev 2003;(1):CD004016.

30. Dunsford A, Kumar S, Clarke S. Integrating evidence into prac-tice: use of McKenzie-based treatment for mechanical low back pain. J Multidiscip Healthc. 2011;4:393–402.

31. Rolving N, Christiansen DH, Andersen LL, Skotte J, Ylinen J, Jensen OK, et al. Effect of strength training in addition to gen-eral exercise in the rehabilitation of patients with non-specific neck pain. A randomized clinical trial. Eur J Phys Rehabil Med. 2014;50(6):617–626.

32. Employment DMO. Law on sick leave compensation ben-efits. 2014; https://www.retsinformation.dk/Forms/R0710.aspx?id=161736#Kap8. Accessed 06 Jan 2017.

33. Bolton JE, Humphreys BK, van Hedel HJ. Validity of weekly recall ratings of average pain intensity in neck pain patients. J Manipulative Physiol Ther. 2010;33(8):612–617.

34. Linton SJ, Boersma K. Early identification of patients at risk of developing a persistent back problem: the predictive validity of the Orebro Musculoskeletal Pain Questionnaire. Clin J Pain. 2003;19(2):80–86.

35. Hockings RL, McAuley JH, Maher CG. A systematic review of the predictive ability of the Orebro Musculoskeletal Pain Ques-tionnaire. Spine (Phila Pa 1976). 2008;33(15):E494–E500.

36. Beaton DE, Katz JN, Fossel AH, Wright JG, Tarasuk V, Bombar-dier C. Measuring the whole or the parts? Validity, reliability, and responsiveness of the disabilities of the arm, shoulder and hand outcome measure in different regions of the upper extremity. J Hand Ther. 2001;14(2):128–146.

37. Jordan A, Manniche C, Mosdal C, Hindsberger C. The Copenha-gen Neck Functional Disability Scale: a study of reliability and validity. J Manipulative Physiol Ther. 1998;21(8):520–527.

38. Bjorner JB, Damsgaard MT, Watt T, Groenvold M. Tests of data quality, scaling assumptions, and reliability of the Danish SF-36. J Clin Epidemiol. 1998;51(11):1001–1011.

39. Palmer KT, Harris EC, Linaker C, Barker M, Lawrence W, Cooper C, et al. Effectiveness of community- and workplace-based inter-ventions to manage musculoskeletal-related sickness absence and job loss: a systematic review. Rheumatology. 2012;51(2):230–242.

40. Jensen C, Jensen OK, Christiansen DH, Nielsen CV. One-year follow-up in employees sick-listed because of low back pain:

randomized clinical trial comparing multidisciplinary and brief intervention. Spine (Phila Pa 1976). 2011;36(15):1180–1189.

41. Stapelfeldt CM, Jensen C, Andersen NT, Fleten N, Nielsen CV. Validation of sick leave measures: self-reported sick leave and sickness benefit data from a Danish national register com-pared to multiple workplace-registered sick leave spells in a Danish municipality. BMC Public Health. 2012;12:661. doi:10.1186/1471-2458-12-661.

42. Cancelliere C, Donovan J, Stochkendahl MJ, Biscardi M, Ammen-dolia C, Myburgh C, et al. Factors affecting return to work after injury or illness: best evidence synthesis of systematic reviews. Chiropr Man Ther. 2016;24(1):32. doi:10.1186/s12998-016-0113-z. eCollection 2016.

43. Kovacs FM, Abraira V, Royuela A, Corcoll J, Alegre L, Tomas M, et al. Minimum detectable and minimal clinically important changes for pain in patients with nonspecific neck pain. BMC Musculoskelet Disord. 2008;9:43. doi:10.1186/1471-2474-9-43.

44. Pool JJ, Ostelo RW, Hoving JL, Bouter LM, de Vet HC. Minimal clinically important change of the Neck Disability Index and the Numerical Rating Scale for patients with neck pain. Spine (Phila Pa 1976). 2007;32(26):3047–3051.

45. Hagen EM, Eriksen HR, Ursin H. Does early intervention with a light mobilization program reduce long-term sick leave for low back pain? Spine (Phila Pa 1976). 2000;25(15):1973–1976.

46. Myhre K, Marchand GH, Leivseth G, Keller A, Bautz-Holter E, Sandvik L, et al. The effect of work-focused rehabilitation among patients with neck and back pain: a randomized controlled trial. Spine. 2014;39(24):1999–2006.

47. Brendbekken R, Eriksen HR, Grasdal A, Harris A, Hagen EM, Tangen T. Return to work in patients with chronic musculoskeletal pain: multidisciplinary intervention versus brief intervention: a randomized clinical trial. J Occup Rehabil. 2017;27(1):82–91.

48. Biering K, Hjollund NH, Lund T. Methods in measuring return to work: a comparison of measures of return to work follow-ing treatment of coronary heart disease. J Occup Rehabil. 2013;23(3):400–405.

49. Holtermann A, Hansen JV, Burr H, Sogaard K. Prognostic fac-tors for long-term sickness absence among employees with neck-shoulder and low-back pain. Scand J Work Environ Health. 2010;36(1):34–41.

50. Feleus A, Miedema HS, Bierma-Zeinstra SM, Hoekstra T, Koes BW, Burdorf A. Sick leave in workers with arm, neck and/or shoulder complaints; defining occurrence and discriminative trajectories over a 2-year time period. Occup Environ Med. 2017;74(2):114–122.

51. Ritchie C, Hendrikz J, Kenardy J, Sterling M. Derivation of a clinical prediction rule to identify both chronic moderate/severe disability and full recovery following whiplash injury. Pain. 2013;154(10):2198–2206.

52. Ritchie C, Hendrikz J, Jull G, Elliott J, Sterling M. External vali-dation of a clinical prediction rule to predict full recovery and ongoing moderate/severe disability following acute whiplash injury. J Orthop Sports Phys Ther. 2015;45(4):242–250.

53. Kelly J, Ritchie C, Sterling M. Clinical prediction rules for prognosis and treatment prescription in neck pain: a systematic review. Musculoskelet Sci Pract. 2017;27:155–164. doi:10.1016/j.math.2016.10.066.

https://www.retsinformation.dk/forms/r0710.aspx?id=81200

https://www.retsinformation.dk/forms/r0710.aspx?id=81200

http://dx.doi.org/10.1186/s12998-015-0076-5

https://www.retsinformation.dk/Forms/R0710.aspx?id=161736#Kap8


http://dx.doi.org/10.1186/1471-2458-12-661

http://dx.doi.org/10.1186/s12998-016-0113-z

http://dx.doi.org/10.1186/s12998-016-0113-z

http://dx.doi.org/10.1186/1471-2474-9-43

http://dx.doi.org/10.1016/j.math.2016.10.066

http://dx.doi.org/10.1016/j.math.2016.10.066

Study 22

METHODOLOGY Open Access

Degenerative findings on MRI of thecervical spine: an inter- and intra-raterreliability studyLine Thorndal Moll1,2,3* , Morten Wasmod Kindt4, Christina Malmose Stapelfeldt1,2 and Tue Secher Jensen4,5

Abstract

Background: Knowledge about the assessment reliability of common cervical spine changes is a prerequisite forprecise and consistent communication about Magnetic Resonance Imaging (MRI) findings. The purpose of this studywas to determine the inter- and intra-rater reliability of degenerative findings when assessing cervical spine MRI.

Methods: Fifty cervical spine MRIs from subjects with neck pain were used. A radiologist, a chiropractor and a second-year resident of rheumatology independently assessed kyphosis, disc height, disc contour, vertebral endplate signalchanges, spinal canal stenosis, neural foraminal stenosis, and osteoarthritis of the uncovertebral and zygapophysealjoints. An evaluation manual was composed containing classifications and illustrative examples, and ten of the MRIswere evaluated twice followed by consensus meetings to refine the classifications. Next, the three readersindependently assessed the full sample. Reliability measures were reported using prevalence estimates andunweighted kappa (Κ) statistics.Results: The overall inter-rater reliability was substantial (Κ≥ 0.61) for the majority of variables and moderate only forzygapophyseal osteoarthritis (Κ = 0.56). Intra-rater reliability estimates were higher for all findings.

Conclusions: The present classifications for some of the most common cervical degenerative findings yielded mainlysubstantial inter-rater reliability estimates and substantial to almost perfect intra-rater reliability estimates. .

Trial registration: Regional Data Protection Agency (J.no. 1–16–02-86-16). The letter of exemption from the RegionalEthical Committee is available from the author on request (case no. 86 / 2017).

Keywords: Magnetic resonance imaging, Reliability, Cervical spine, Degenerative, Classification, MRI, Agreement

BackgroundAlthough not recommended as routine imaging in neckpain [1, 2], the number of cervical MRIs has increasedby 18% compared to a 4.5% increase in neck pain preva-lence over recent years in Denmark [2–4]. While pa-tients believe in MRI to unveil the true cause of theirpain [5], health care professionals appreciate the advan-tages of MRI compared with other modalities of diagnos-tic imaging. The non-invasiveness, absence of radiationexposure and the capacity to discriminate soft tissue

changes are all highly valued in the field of musculoskel-etal imaging.When communicating MRI findings, the importance

of consistency and precision remains unaltered. Both foracademic and clinical purposes, a prerequisite for suchconsistency and precision is reliability in MRI assess-ments. Reliability is defined as “the extent to whichscores for patients who have not changed are the samefor repeated measurement under several conditions” [6].In the case of MRI, this means that while the images donot change, reliability reflects whether the image inter-pretation remains the same when assessed by differentraters (inter-rater reliability) or by the same rater at dif-ferent times (intra-rater reliability).Previous reliability studies on cervical spine MRI have

found moderate to almost perfect inter-rater reliability

* Correspondence: [email protected], Central Denmark Region, P.P. Oerums Gade 11, bygn. 1B,DK-8000 Aarhus C, Denmark2Section of Clinical Social Medicine and Rehabilitation, Department of PublicHealth, Aarhus University, P.P. Oerums Gade 9-11, bygn. 1B, DK-8000 AarhusC, DenmarkFull list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Moll et al. Chiropractic & Manual Therapies (2018) 26:43 https://doi.org/10.1186/s12998-018-0210-2

http://crossmark.crossref.org/dialog/?doi=10.1186/s12998-018-0210-2&domain=pdf

http://orcid.org/0000-0003-4132-0310

http://www.rm.dk/siteassets/sundhed/faginfo/forskning/datatilsynet/indberettet-til-datatilsynet/indberetning-tildatatilsynet-2016.pdf

mailto:[email protected]


http://creativecommons.org/publicdomain/zero/1.0/

in the assessments of disc-related parameters (kappa (Κ)0.44[7], Κ 0.43–0.65 [8] and Κ 0.73–0.83 [9]). Almostperfect reliability has been reported for assessments ofneural foraminal stenosis (Κ > 0.9 [10]), fair reliability forfacet joint arthrosis (Κ 0.23–0.38 [11]), and moderate tosubstantial reliability for spinal canal stenosis (Κ 0.55–0.72 [11]). Most studies have focused on only one or afew degenerative variables [7–13] and compared readerswith similar educational backgrounds and levels of ex-perience [7–10, 12–14].To our knowledge, only one reliability study on cer-

vical spine MRI has covered a broad range of commondegenerative findings [14] for which reason, furtherstudies are needed.

ObjectiveTo determine the inter- and intra-rater assessment reli-ability of degenerative findings (kyphosis, disc height,disc contour, vertebral endplate signal changes, spinalcanal stenosis, neural foraminal stenosis, uncovertebralosteoarthritis and zygapophyseal osteoarthritis) on MRIof the cervical spine.

MethodsSubjectsFifty MRIs of the cervical spine were chosen from amongsubjects previously enrolled in a randomized controlledtrial (RCT) [15]. Subjects for the RCT were recruited fromprimary health care professionals (physiotherapists, chiro-practors and general practitioners (GPs)). If subjects ful-filled the inclusion criteria (age 18–60 years, part-time orfull-time sick leave for 4–16 weeks owing to neck pain orshoulder pain, and fluency in Danish), their GPs referredthem to The Spine Centre, Silkeborg Regional Hospital,Denmark. For the current study, the predefined inclusioncriterion was the availability of a cervical spine MRI with asatisfactory signal-to-noise ratio. After assessment by themost experienced reader, 32 MRIs were excluded based onunsatisfactory signal-to-noise ratio. By choosing every sec-ond MRI among those remaining, 50 MRIs were selectedfor the current study. A study flow-chart is seen in Fig. 1.

Data collection - imagesThe MRIs were provided from five different hospitalscollaborating with The Spine Centre. The majority of theimages were obtained using a 1.5 T field strength. AllMRIs comprised sagittal T1-weighted and T2-weightedsequences, while an axial T2 sequence was available for94% and oblique T2 sequences were available for 82% ofthe images.

Data collection – readersThe three readers (Readers A, B and C) all assessed theimages independently over a time frame of 5–8 weeks.

Reader A was a second-year resident of rheumatologywith no previous formal education in MRI assessment.She had 9 years of postgraduate clinical experience includ-ing assessment of spinal MRI for clinical purposes. ReaderB was an experienced radiologist having worked withmusculoskeletal MRI for 25 years, mostly on a daily basis.Reader C was a chiropractor who had completed a 1-yearfulltime internship in spinal MRI in a radiology depart-ment. He had another 10 years of clinical and academicexperience with spinal MRI. Prior to the study, Reader Btaught Reader A assessment of cervical spine MRI for 2 h.Following this two-hour session, Reader A completed 50clinical narrative reports of cervical spine MRIs from pa-tients with neck pain with or without radiculopathy. Thesewere not part of the current study. The reports were cor-rected if necessary and approved by Reader B.For the intra-rater reliability assessment, Reader A

assessed all the images twice. The second assessmenttook place after 6 weeks to prevent recollection of thefirst assessments.

Evaluation manual, piloting and workstationsBased on the literature [10–14, 16–24], an evaluationmanual with written and visual classifications of thefindings was made by Reader A, adjusted and approvedby Readers B and C. Next, 10 MRIs from the study sam-ple were evaluated twice followed by consensus meet-ings. This piloting served the purpose of refining boththe classifications in the evaluation manual and the prac-tice of the readers. All images were de-identified, leaving

Fig. 1 Flowchart

Moll et al. Chiropractic & Manual Therapies (2018) 26:43 Page 2 of 8

the readers blinded to demographic and clinical data aswell as previous assessments. The images were assessedon radiological work stations using Vitrea Core (version1.0.0.404, Vital Images Inc.).

VariablesClassifications for common and degenerative MRI find-ings were developed based primarily on the existing lit-erature [10–13, 16–19, 23–26] and on experiences fromthe piloting. An effort was made to create definitionsthat were as simple as possible [14], assuming that sim-plicity is essential for clinical applicability. The mostcommon degenerative findings were chosen, includingkyphosis and vertebral endplate signal changes; all areroutinely considered by radiologists assessing cervicalspine MRIs at Silkeborg Regional Hospital. All the classi-fications yielded categorical (but not ordinal) data. Thecomplete list of variables is presented in Table 1. Exceptfor kyphosis, these findings were assessed for each of thesix cervical disc levels (level C2/C3 to C7/T1). Further-more, the neural foramina, uncovertebral and zygapo-physeal joints were assessed separately on the left andright hand side. The evaluation manual is available inAdditional file 1.

Data entry and statistical analysisAll three readers independently entered and stored datausing Epidata (Version 3.1., The EpiData Association,Odense, Denmark, 2003–2004). If assessment of a cer-tain finding was not possible due to the available se-quences, the particular finding was allotted the value ‘9’representing ‘missing’.In accordance with the recommendations for reliability

studies [27], 50 MRIs were included in the current study.Prior to the kappa (Κ) calculations, all readers’ preva-lence assessments were calculated, one variable at atime. This tabulation of data offered the opportunity of1) assessing the sample homogeneity and 2) identifyingany possible systematic differences between the readers;as both can affect the Κ estimates [27, 28]. Tabulationthus allowed for a clearer impression of agreement andpossible misclassification than offered by the Κ valuealone. Tabulation also provided estimates for observedagreement (OA) and agreement by chance (AC) for thepairwise analyses. For the overall three-reader analysis,OA was calculated by computing the number of obser-vations with complete agreement and dividing this num-ber with the total number of anatomical sites assessed.The three-reader AC was calculated by multiplication of

Table 1 MRI findings and corresponding classifications

MRI finding Category Description

Kyphosis 0 Normal or straightened lordosis

1 Kyphosis

Disc height 0 Normal

1 Reduced height

Disc contour 0 Normal

1 Bulge or protrusion

2 Extrusion

Spinal canal stenosis 0 Normal

1 > 50% obliteration of CSF, no cord deformity

2 > 50% obliteration of CSF with cord deformity but no signal change

3 > 50% obliteration of CSF with cord deformity and signal change

Vertebral endplate signal change 0 Normal

1 Type 1

2 Type 2

3 Type 3

4 Mixed type 1 and 2

Uncovertebral osteoarthritis 0 Normal

1 Definite osteoarthritis

Zygapophyseal osteoarthritis 0 Normal

1 Definite osteoarthritis

Neural foraminal stenosis 0 Normal or < 50% fat obliteration

1 ≥ 50% fat obliteration with or without morphological changes of the nerve root

CSF cerebrospinal fluid


the marginal fractions [27]. Reliability measures werecomputed using unweighted kappa statistics owing tothe categorical (as opposed to ordinal) nature of thedata. Given the condition of total independence amongthe readers, Κ is defined as

K ¼ OA−AC1−AC

where OA is observed agreement and AC agreement bychance [29]. Reliability measures were computed for thereaders in pairs (A1B1, A1C1, B1C1, A1A2) and over-all(A1B1C1). Acknowledging the influence of prevalenceon the Κ estimates [27, 28], these were only computedwhenever the readers in question agreed on prevalences≥10%. For each disc level, the left and right hand side as-sessments of neural foraminal stenosis, uncovertebraland zygapophyseal osteoarthritis were pooled beforecomputing reliability estimates. The interpretation of Κvalues followed the suggestions by Landis & Koch [29]:Κ value Strength of agreement

< 0.0: Poor

0.0–0.2 Slight

0.21–0.4 Fair

0.41–0.6 Moderate

0.61–0.8 Substantial

0.81–1.0 Almost perfect

Κ values were reported using 95% confidence intervalsand additional information on OA and AC were suppliedfor all findings. Analyses were performed using theSTATA (version 15.0; Stata Corporation, College Station,Texas, USA) software package.

EthicsAll subjects provided written informed consent. Thestudy was approved by the Regional Data ProtectionAgency (J.no. 1–16–02-86-16). Approval by the regionalethical committee was not needed due to the study’smethodological nature. The letter of exemption fromThe Central Denmark Region Committees on HealthResearch Ethics is available from the author on request(case no. 86 / 2017).

ResultsThe majority of the subjects were female (n = 31; 62%)with a mean age of 43.7 years (SD = 9.2). The prevalence ofpositive findings for all readers can be seen inAdditional file 2. For vertebral endplate signal changes,prevalence estimates were below 10% and thus too low forΚ statistics. For the remaining degenerative findings,prevalence estimates allowed for kappa statistics includingone to six anatomical sites (e.g. 2 disc levels ~ 100

observations included in Κ analysis for spinal canalstenosis). Further scrutiny of the prevalence table revealed aslight tendency for Reader C to assign the label “reduceddisc height” more frequently. Otherwise no systematicdifferences among the readers were identified.As shown in Table 2, the overall inter-rater reliability

(A1B1C1) ranged from moderate to almost perfect forthe majority of the findings (substantial to almost perfectfor kyphosis and neural foraminal stenosis; moderate toalmost perfect for spinal canal stenosis; and moderate tosubstantial for disc height, disc contour, uncovertebraland zygapophyseal osteoarthritis). Exploratory analyseswere made to assess the inter-rater reliability of neuralforaminal stenosis when including only MRIs with ob-lique images (Additional file 3). This did not change thereliability estimates but broadened the confidence inter-vals slightly.The intra-rater reliability estimates (Table 3) were slightly

better than those for inter-rater reliability. Almost perfectreliability was found for kyphosis and substantial to almostperfect reliability for disc contour, uncovertebral osteoarth-ritis and neural foraminal stenosis. For spinal canal stenosisand zygapophyseal osteoarthritis, moderate to almost per-fect intra-rater reliability was found while moderate to sub-stantial reliability was found for disc height.

DiscussionTo our knowledge, this is the first reliability studycovering eight common cervical MRI findings. Theoverall inter-rater reliability was substantial for all vari-ables except zygapophyseal osteoarthritis where moder-ate reliability was found. Intra-rater reliability wassubstantial for the majority of variables and almost per-fect for kyphosis. These reliability estimates reflect thatthe observed agreement notably exceeds the agreementthat can be expected by chance.For disc degeneration, other studies [9, 12] reported

higher reliability estimates than the disc height estimates inthe current study. Although the use of intraclasscorrelation coefficient in the study by Jacobs et al. [12] doesnot allow for direct comparison, possible explanations forthe reliability differences are the use of a ubiquitouslyaccessible reference image of a normal disc [12] and thenotable experience among readers with the sameeducational background [9].For disc contour, the reliability estimates were similar to

those of other studies despite the fact that we used athree-category classification compared to the previouslyreported dichotomous classifications [8, 30, 31] and com-parison of more experienced readers [30, 31].For spinal canal stenosis, the current study’s unweighted

reliability estimates exceeded those previously reported byuse of weighted kappa statistics [13, 32], although the useof weights are expected to yield higher estimates. A higher


http://www.rm.dk/siteassets/sundhed/faginfo/forskning/datatilsynet/indberettet-til-datatilsynet/indberetning-tildatatilsynet-2016.pdf

Table 2 Inter-rater reliability estimates

MRI finding n Reader pair Observed agreement (%) Agreement by chance (%) Kappa (95% CI)

Kyphosisa 50 A1B1 92.0 56.4 0.82 (0.75; 0.89)

49 A1C1 89.8 53.6 0.78 (0.71; 0.85)

49 B1C1 89.8 52.8 0.78 (0.71; 0.86)

49 A1B1C1 85.7 31.2 0.79 (0.73; 0.85)

Disc heightb 150 A1B1 92.0 52.8 0.83 (0.74; 0.92)

200 A1C1 80.0 52.8 0.58 (0.46; 0.69)

150 B1C1 77.3 50.0 0.55 (0.42; 0.68)

150 A1B1C1 74.7 26.4 0.65 (0.57; 0.74)

Disc contourb 177 A1B1 76.8 43.4 0.59 (0.49; 0.70)

177 A1C1 79.7 43.3 0.64 (0.53; 0.74)

200 B1C1 80.0 47.6 0.62 (0.52; 0.72)

177 A1B1C1 68.4 21.7 0.61 (0.54; 0.69)

Spinal canal stenosisb 100 A1B1 97.0 76.0 0.88 (0.68; 1.00)

100 A1C1 91.0 73.5 0.66 (0.47; 0.83)

100 B1C1 92.0 74.3 0.69 (0.48; 0.86)

100 A1B1C1 90.0 63.0 0.74 (0.57; 0.86)

Vertebral endplate signal change Too low prevalences (i.e. ≤ 10%)

Uncovertebral osteoarthritisc 222 A1B1 90.1 68.0 0.69 (0.57; 0.81)

237 A1C1 89.0 68.6 0.65 (0.53; 0.77)

230 B1C1 87.4 70.9 0.57 (0.43; 0.71)

222 A1B1C1 83.3 53.0 0.65 (0.51; 0.76)

Zygapophyseal osteoarthritisc 270 A1B1 94.8 74.2 0.80 (0.70; 0.90)

144 A1C1 87.5 74.9 0.50 (0.31; 0.70)

184 B1C1 85.9 78.9 0.33 (0.13; 0.53)

135 A1B1C1 83.0 61.0 0.56 (0.43; 0.70)

Neural foraminal stenosisc 268 A1B1 90.7 64.1 0.74 (0.65; 0.84)

287 A1C1 90.2 64.2 0.73 (0.63; 0.82)

275 B1C1 87.6 65.8 0.64 (0.53; 0.75)

268 A1B1C1 84.0 46.0 0.73 (0.63; 0.82)an refers to the number of MRIs assessedbn refers to the number of disc levels assessedcn refers to the number of anatomical sites assessed (by pooling right and left hand side)

Table 3 Intra-rater reliability estimates

MRI finding n Reader pair Observed agreement (%) Agreement by chance (%) Kappa (95% CI)

Kyphosisa 50 A1A2 96.0 59.6 0.90 (0.85; 0.96)

Disc heightb 200 A1A2 84.0 51.5 0.67 (0.57; 0.77)

Disc contourb 174 A1A2 88.5 43.9 0.80 (0.71; 0.87)

Spinal canal stenosisb 50 A1A2 94.0 76.6 0.73 (0.51; 0.90)

Vertebral endplate signal change Too low prevalences (i.e. ≤ 10%)

Uncovertebral osteoarthritisc 281 A1A2 90.4 67.0 0.71 (0.61; 0.81)

Zygapophyseal osteoarthritisc 240 A1A2 90.8 68.8 0.71 (0.59; 0.82)

Neural foraminal stenosisc 287 A1A2 90.6 62.6 0.75 (0.66; 0.84)an refers to the number of MRIs assessedbn refers to the number of disc levels assessedcn refers to the number of anatomical sites assessed (by pooling right and left hand side)


number of readers (six [13] and nine [32]) could explainthis difference, but even when compared to the threemost experienced readers in these studies, betterreliability estimates were still achieved in the currentstudy. The most probable reason appears to be the limitedintroduction of their classification [13, 32]. When usingboth written and visual descriptions, our moderate toalmost perfect reliability among readers with considerableexperience differences suggest good applicability of thisclassification of spinal canal stenosis.For zygapophyseal osteoarthritis, both the intra- and

inter-rater reliability estimates were better than previouslyreported [11], which is most likely explained by the use ofa dichotomous variable in the current study compared toa classification with four severity categories [11].For neural foraminal stenosis, this study still achieved

higher reliability estimates compared to studies withmore experienced readers [30, 31]. The inferior reliabilityestimates may be explained by unclear definitions [30]and by low prevalence estimates together with imagesobtained using a 0.5 T field strength [31]. Compared tothe study from which we modified the classification ofneural foraminal stenosis [10], the current study wasunable to reach the same almost perfect reliabilityestimates (Κ > 0.9). Nevertheless, we consider thesubstantial to almost perfect reliability to be satisfactory,bearing in mind differences in reader experience and theheterogeneous image material (i.e. images with differentfield strengths and available sequences). The modifiedclassification (dichotomous versus the original fourcategories) proved reliable and the association withclinical findings has previously been reported [33].

Methodological considerationsA limitation of the study is that it was not preceded by apower calculation. However; the confidence intervals forthe Κ estimates only comprised more than two levels(e.g. from moderate to almost perfect for spinal canalstenosis) in a minority of cases. A larger sample wouldhave narrowed the confidence intervals but wouldprobably not have caused substantial changes in thereliability estimates.Another limitation is the involvement of only reader A

in the intra-rater analysis. Two considerations explain this:1) previous reliability studies found higher [7–9, 12, 14, 21]or similar/higher [10, 11, 13] intra-rater reliability thaninter-rater reliability and 2) involvement of reader A wasnecessary since a future prognostic study will involve MRIassessments performed by reader A. As for the inter-raterreliability, the study included three readers, only one ofthese being a radiologist. However, the results suggest thatour method is applicable among other health care profes-sionals (i.e. rheumatologists and chiropractors) in a con-trolled research setting. Involvement of other relevant

healthcare professionals, e.g. spine surgeons, would havebeen desirable but was unfortunately not possible.Owing to the properties of Κ, the measure does not

disentangle systematic and random misclassification [28].Therefore, we provided the prevalence tables from whichwe find no suspicion of systematic misclassification.The prevalence table discloses a notable difference in

the number of disc levels assessed for disc contour onlevels C2/C3, C3/C4 and C7/T1: Reader A assessedfewer levels than Readers B and C owing to the lack ofaxial images of the selfsame disc levels. This discrepancysuggests a difference among the readers, and whetherthis partly explains why higher reliability estimates werenot achieved for disc contour cannot be refuted.Another potential limitation is that all MRIs were

derived only from individuals with neck pain. But sincecervical spine MRI is seldom performed in patientswithout neck pain and since the future use of theevaluation manual applies to patients with neck pain, weconsider the current sample appropriate for its purpose.Finally, a potential limitation of the study is the

heterogeneous image material (MRIs were performed atfive different hospitals. Different field strengths andsequences were available). Yet, as it resembles everydayclinical practice, this was an intended challenge and anattempt was made to manage this heterogeneity by using astandardized evaluation manual. The differences betweenOA and AC (Tables 2 and 3) reflect that both inter- andintra-rater agreement notably exceed the agreement thatcan be expected by chance. Furthermore, the high levels ofobserved agreement reflect only a minor degree of mis-classification. Based on these observations of OA, our in-terpretation is that the evaluation manual and thestandardized procedures explain the high levels of agree-ment rather than pure chance when assessing heteroge-neous images.Ultimately, the heterogeneous image material and the

use of three different health care professionals both addto the generalizability and thus constitute strengths ofthe study. The blinding of the readers, the use of simpleand easily comprehensible classifications along withregular encouragement to follow the evaluation manual,are other important strengths of the study.In contrast to the controlled settings of the current study,

a study comparing narrative MRI reports demonstratedconsiderable variability [34]. In this study [34], a patientwith low back pain and right L5 radicular symptoms hadlumbar spine MRI performed at 10 different MRI centerswithin 3 weeks. Comparison of the 10 narrative reportsrevealed considerable variability; none of the 49 describedfindings occurred in all 10 reports and only one findingoccurred in nine reports. Even if this amount of variabilityis unusually large [34], it supports our clinical experiencethat variability also prevails in the interpretation of cervical


spine MRIs. A possible way to overcome this is by usingclassifications sufficiently comprehensible to be applied 1)by different health care professionals and 2) when assessingheterogeneous images from different MRI scanners. Suchclassifications were presented in the current study.Confirmatory studies will be needed. If those studies wereto involve experienced radiologists, provide proper trainingfor lesser experienced MRI readers, and use an evaluationmanual, better reliability might be achieved in clinicalsettings. So far, the results suggest that the evaluation ofMRI findings can be used in controlled research settingsstudying individuals with neck pain. Suggestions for futureresearch include comparison of reliability with and withoutthe use of an evaluation manual. Also, including more thanone of each health care professional could allow forcomparison of experience levels both among and withindifferent types of health care professionals.

ConclusionsIn conclusion, the current study found substantial reliabilityfor the majority of included MRI findings. This suggests thatthe present classifications are sufficiently comprehensible tobe applied by different health care professionals whenassessing images from different MRI scanners. In our view,the proposed classifications are sufficiently reliable to beused for both quality assurance and further researchpurposes.

Additional files

Additional file 1: The evaluation manual used for assessment of theMRIs. (DOCX 2347 kb)

Additional file 2: A prevalence table reporting the frequency of positivefindings for all the readers. (DOCX 30 kb)

Additional file 3: A table of sensitivity analyses. For neural foraminalstenosis, kappa estimates are presented comparing the assessments of allimages vs. only images with available oblique slices. (DOCX 16 kb)

AbbreviationsAC: Agreement by chance; CSF: Cerebrospinal fluid; GP: General practitioner;MRI: Magnetic resonance imaging; OA: Observed agreement;RCT: Randomized controlled trial; SD: Standard deviation; Κ: Kappa

AcknowledgementsA special thanks to Brian Højgaard for readily providing technical supportwhenever needed.

FundingThis work was supported by the Tryg Foundation, Aarhus University Denmark,Danish Rheumatism Association, and Aase and Ejnar Danielsen Foundation.

Availability of data and materialsThe datasets used and/or analyzed during the current study are availablefrom the corresponding author on reasonable request.

Authors’ contributionsLTM, MWK and TSJ designed the study and collected the data. LTMperformed the statistical analyses and drafted the manuscript. All the authorscontributed to the interpretation of data. All the authors critically revised andapproved the final manuscript.

Ethics approval and consent to participateWritten informed consent was provided from the participants. The study wasapproved by the Regional Data Protection Agency (J.no. 1–16–02-86-16).Approval by the Regional Ethical Committee was not needed due to thestudy’s methodological nature. The letter of exemption from the RegionalEthical Committee is available from the author on request (case no. 86 / 2017).

Consent for publicationNot applicable

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1DEFACTUM, Central Denmark Region, P.P. Oerums Gade 11, bygn. 1B,DK-8000 Aarhus C, Denmark. 2Section of Clinical Social Medicine andRehabilitation, Department of Public Health, Aarhus University, P.P. OerumsGade 9-11, bygn. 1B, DK-8000 Aarhus C, Denmark. 3Spine Centre, DiagnosticCentre, University Research Clinic for Innovative Patient Pathways, SilkeborgRegional Hospital, Falkevej 1-3, DK-8600 Silkeborg, Denmark. 4Department forDiagnostic Imaging, Diagnostic Centre, University Research Clinic forInnovative Patient Pathways, Silkeborg Regional Hospital, Falkevej 1-3,DK-8600 Silkeborg, Denmark. 5Nordic Institute of Chiropractic and ClinicalBiomechanics, University of Southern Denmark, Campusvej 55, DK-5230Odense M, Denmark.

Received: 27 March 2018 Accepted: 13 August 2018

References1. Stochkendahl MJ, Kjaer P, Hartvigsen J, et al. National Clinical Guidelines for

non-surgical treatment of patients with recent onset low back pain orlumbar radiculopathy. Eur Spine J. 2018;27(1):60–75. https://doi.org/10.1007/s00586-017-5099-2.

2. Jensen HAR, Davidsen M, Christensen AI. The National Health Profile. 2017;2018:41.

3. National Danish Patient Registry. http://www.esundhed.dk/sundhedsregistre/LPR/Sider/LPR04_Tabel.aspx. Accessed 23 Nov 2017.

4. Christensen AI, Davidsen M, Juel K. The National Health Profile, vol. 2014;2013. p. 37.

5. Petersen L, Birkelund R, Ammentorp J, Schiøttz-Christensen B. "An MRIreveals the truth about my back": a qualitative study about patients’expectations and attitudes toward the value of MRI in the assessment ofback pain. Eur J Pers Cent Healthc. 2016;4(3):453–8.

6. Mokkink LB, Terwee CB, Patrick DL, et al. The COSMIN study reachedinternational consensus on taxonomy, terminology, and definitions ofmeasurement properties for health-related patient-reported outcomes. JClin Epidemiol. 2010;63(7):737–45.

7. Kolstad F, Myhr G, Kvistad KA, Nygaard OP, Leivseth G. Degeneration andheight of cervical discs classified from MRI compared with precise heightmeasurements from radiographs. Eur J Radiol. 2005;55(3):415–20.

8. Mann E, Peterson CK, Hodler J. Degenerative marrow (modic) changes oncervical spine magnetic resonance imaging scans: prevalence, inter- andintra-examiner reliability and link to disc herniation. Spine (Phila Pa 1976).2011;36(14):1081–5.

9. Miyazaki M, Hong SW, Yoon SH, Morishita Y, Wang JC. Reliability of amagnetic resonance imaging-based grading system for cervicalintervertebral disc degeneration. J Spinal Disord Tech. 2008;21(4):288–92.

10. Park HJ, Kim SS, Lee SY, et al. A practical MRI grading system for cervicalforaminal stenosis based on oblique sagittal images. Br J Radiol. 2013;86(1025):20120515.

11. Xu C, Ding ZH, Xu YK. Comparison of computed tomography and magneticresonance imaging in the evaluation of facet tropism and facet arthrosis indegenerative cervical spondylolisthesis. Genet Mol Res. 2014;13(2):4102–9.

12. Jacobs LJ, Chen AF, Kang JD, Lee JY. Reliable Magnetic Resonance ImagingBased Grading System for Cervical Intervertebral Disc Degeneration. AsianSpine J. 2016;10(1):70–4.


https://doi.org/10.1186/s12998-018-0210-2

https://doi.org/10.1186/s12998-018-0210-2

https://doi.org/10.1186/s12998-018-0210-2

https://doi.org/10.1007/s00586-017-5099-2

https://doi.org/10.1007/s00586-017-5099-2

http://www.esundhed.dk/sundhedsregistre/LPR/Sider/LPR04_Tabel.aspx


13. Kang Y, Lee JW, Koh YH, et al. New MRI grading system for the cervicalcanal stenosis. AJR Am J Roentgenol. 2011;197(1):W134–40.

14. Fu MC, Webb ML, Buerba RA, et al. Comparison of agreement of cervicalspine degenerative pathology findings in magnetic resonance imagingstudies. Spine J. 2016;16(1):42–8.

15. Moll LT, Jensen OK, Schiottz-Christensen B, et al. Return to Work inEmployees on Sick Leave due to Neck or Shoulder Pain: A RandomizedClinical Trial Comparing Multidisciplinary and Brief Intervention with One-Year Register-Based Follow-Up. J Occup Rehabil. 2017;28(2): 346–356.

16. Nouri A, Martin AR, Mikulis D, Fehlings MG. Magnetic resonance imagingassessment of degenerative cervical myelopathy: a review of structuralchanges and measurement techniques. Neurosurg Focus. 2016;40(6):E5.

17. Fardon DF, Williams AL, Dohring EJ, Murtagh FR, Gabriel Rothman SL, SzeGK. Lumbar disc nomenclature: version 2.0: Recommendations of thecombined task forces of the North American Spine Society, the AmericanSociety of Spine Radiology and the American Society of Neuroradiology.Spine J. 2014;14(11):2525–45.

18. Bojsen-Moeller F. Chapter 8: Hvirvelsoejlen (The Spine). In: BevaegeapparatetsAnatomi, vol. 89. Copenhagen: Munksgaard Danmark; 2001.

19. Wiltse LL, Berger PE, McCulloch JA. A system for reporting the size andlocation of lesions in the spine. Spine (Phila Pa 1976). 1997;22(13):1534–7.

20. Maatta JH, Karppinen J, Paananen M, et al. Refined Phenotyping of ModicChanges: Imaging Biomarkers of Prolonged Severe Low Back Pain andDisability. Medicine (Baltimore). 2016;95(22):e3495.

21. Kim S, Lee JW, Chai JW, et al. A New MRI Grading System for CervicalForaminal Stenosis Based on Axial T2-Weighted Images. Korean J Radiol.2015;16(6):1294–302.

22. Kalichman L, Suri P, Guermazi A, Li L, Hunter DJ. Facet orientation andtropism: associations with facet joint osteoarthritis and degeneratives. Spine(Phila Pa 1976). 2009;34(16):E579–85.

23. Shim JH, Park CK, Lee JH, et al. A comparison of angled sagittal MRI andconventional MRI in the diagnosis of herniated disc and stenosis in thecervical foramen. Eur Spine J. 2009;18(8):1109–16.

24. Yochum TR, Rowe LJ. Chapter 10: Arthritic Disorders. In: AnonymousEssentials of Skeletal Radiology. Baltimore: Lippincott Williams & Wilkins;2004. p. 951–1134.

25. Jensen TS, Bendix T, Sorensen JS, Manniche C, Korsholm L, Kjaer P.Characteristics and natural course of vertebral endplate signal (Modic)changes in the Danish general population. BMC Musculoskelet Disord. 2009;10:81. 2474-10-81

26. Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative diskdisease: assessment of changes in vertebral body marrow with MR imaging.Radiology. 1988;166(1 Pt 1):193–9.

27. de Wet HCW, Terwee CB, et al. Chapter 5: Reliability. In: de Wet HCW,Terwee CB, et al., editors. Anonymous Measurement in Medicine.Cambridge: Cambridge University Press; 2011. p. 96–126.

28. Guggenmoos-Holzmann I. How reliable are chance-corrected measures ofagreement? Stat Med. 1993;12(23):2191–205.

29. Landis JR, Koch GG. The measurement of observer agreement forcategorical data. Biometrics. 1977;33(1):159–74.

30. Kuijper B, Beelen A, van der Kallen BF, et al. Interobserver agreement on MRIevaluation of patients with cervical radiculopathy. Clin Radiol. 2011;66(1):25–9.

31. Matsumoto M, Fujimura Y, Suzuki N, et al. MRI of cervical intervertebral discsin asymptomatic subjects. J Bone Joint Surg Br. 1998;80(1):19–24.

32. Ko S, Choi W, Chae S. Comparison of inter- and intra-observer reliabilityamong the three classification systems for cervical spinal canal stenosis. EurSpine J. 2017;26(9):2290-2296.

33. Park HJ, Kim SS, Han CH, et al. The clinical correlation of a new practical MRImethod for grading cervical neural foraminal stenosis based on obliquesagittal images. AJR Am J Roentgenol. 2014;203(2):412–7.

34. Herzog R, Elgort DR, Flanders AE, Moley PJ. Variability in diagnostic errorrates of 10 MRI centers performing lumbar spine MRI examinations on thesame patient within a 3-week period. Spine J. 2017;17(4):554–61.


Study 33

1

Title 1

Prediction of work participation within 2 years in sickness absentees with neck or shoulder pain: the contribution of 2

demographic, patient-reported, clinical and imaging information 3

Journal 4

Journal of Occupational Rehabilitation 5

The names of the authors 6

Line Thorndal Moll1,2,3

, Anne Mette Schmidt1,2,4

, Christina Malmose Stapelfeldt1,2

, Merete Labriola1,2

, Ole Kudsk 7

Jensen3, Morten Wasmod Kindt

5, Tue Secher Jensen

5,6, Berit Schiøttz-Christensen

7 8

Corresponding author's affiliation and e-mail 9

1 DEFACTUM, Central Denmark Region, Aarhus, Denmark 10

P.P. Oerums Gade 11, bygn. 1B 11

DK-8000 Aarhus C. 12

2 Section of Clinical Social Medicine and Rehabilitation, Department of Public Health, Aarhus University, Denmark 13

P.P. Oerums Gade 9-11, bygn. 1B 14

DK-8000 Aarhus C. 15

3 Spine Centre, Diagnostic Centre, Silkeborg Regional Hospital, Denmark 16

Falkevej 1-3 17

DK-8600 Silkeborg 18

e-mail: [email protected] 19

20

Abstract 21

Purpose 22

This cohort study explored the contribution of demographic, patient-reported, clinical, and magnetic resonance 23

imaging (MRI) information to the prediction of work participation in sickness absentees with neck or shoulder pain. 24

2

Methods 25

From a secondary care setting, 168 sickness absentees with neck or shoulder pain were included. Based on registry 26

data, a successful work outcome was defined as ≥ 75% work participation score (WPS) from Weeks 30 to 104 after 27

enrolment. Prognostic variables were categorized into four domains (demographic, patient-reported, clinical, and 28

MRI) resembling the order of information obtained in a clinical setting. Crude logistic regression analyses were used 29

to identify prognostic variables for each domain (p<0.2). This was followed by multivariable analyses including the 30

identified variables in a domain-wise order. For each added domain, the probability of successful WPS was 31

dichotomized leaving two possible classifications: ≥ 50% chance of successful WPS or not. In cross-tabulations of 32

chance and the actual WPS outcome, positive and negative predictive values (PPV and NPV), sensitivity, specificity 33

and area under the curve (AUC) were calculated. 34

Results 35

The combination of demographic and patient-reported variables yielded an NPV of 70% and a PPV of 72%, while 36

specificity was 68%, sensitivity 74% and AUC 0.76. None of these values improved notably by adding clinical and 37

MRI variables as predictors of successful WPS. 38

Conclusions 39

These results suggest that - among sickness absentees with neck or shoulder pain – clinical and MRI variables 40

provide no additional information for the prediction of work participation compared with only demographic and 41

patient-reported information. 42

43

Keywords 44

Prognosis, sick leave, neck pain, shoulder pain, magnetic resonance imaging. 45

Background 46

During sick leave, estimating work prognosis is important as it helps guide expectations for all involved 47

stakeholders. The responsibility for this usually resides with healthcare professionals and it is often requested of 48

them by authorities involved in return-to-work (RTW) coordination and the disbursement of social benefits, i.e. 49

3

sickness benefits. When accommodating this request of work prognosis assessment for patients with neck or 50

shoulder pain, healthcare professionals have access to different domains of information. In a clinical setting, such 51

information is typically obtained in the following order: 1) demographics are known already when scheduling the 52

clinical encounter, and are followed by 2) patient-reported information which is a necessary prerequisite for 53

performing 3) a focused clinical examination, and deciding whether 4) imaging is needed (e.g. magnetic resonance 54

imaging (MRI)). 55

From this range of information, the literature offers insight into known prognostic factors. From the demographic 56

domain, older age is negatively associated with RTW [1, 2], while for the impact of sex, results are conflicting [1-3]. 57

Patient-reported information that is negatively associated with work outcomes includes symptom intensity, sick 58

leave duration [1, 3] and higher physical work demands [1]. From the clinical domain, low back pain studies have 59

found a negative association between radiating pain and RTW in the acute phase (< 6 weeks) [4] and moderate 60

evidence for no association with RTW in the sub-acute and chronic phases [2]. Also from the clinical domain, 61

fibromyalgia is associated with poorer work participation [5]. From the MRI domain, a single study involving 62

sickness absentees with low back pain found Modic type 1 changes on MRI to predict unsuccessful RTW [6]. 63

However, the prognostic value of imaging has been described as "a gap in the literature" [7]. This gap still deserves 64

investigation as the number of MRIs has increased more than the prevalence of people suffering from neck and 65

shoulder pain [8-10]. Moreover, the referral patterns have changed over recent years; e.g. in Denmark, not only 66

hospital physicians but also primary care clinicians (physicians and chiropractors) can request MRI. For clinicians in 67

both primary and secondary care, this change in MRI referral patterns has changed the range of information 68

available when estimating work prognosis. 69

The literature thus offers insight into factors associated with work prognosis. But to our knowledge, no previous 70

studies have examined to what extent the different domains of information contribute to the prediction of work 71

prognosis. Thus healthcare professionals’ assessment of work prognosis may be based on gut-feeling rather than 72

evidence. Therefore, the aim of this study was to explore the degree to which demographic, patient-reported, 73

clinical, and MRI information contribute to the prediction of work participation in sickness absentees with neck or 74

shoulder pain. 75

4

Methods 76

Design and participants 77

Using a cohort design, 168 sickness absentees referred to secondary care for assessment of neck or shoulder pain 78

were included in this study. They were all part of a randomized controlled trial (RCT) [11] that took place from 79

2009-2014 and compared the work outcomes after two different hospital-based interventions (multidisciplinary vs. 80

brief intervention). The inclusion criteria for the RCT were: sick leave 1-4 months owing to neck or shoulder pain, 81

age 18-60 years, and fluency in Danish. Exclusion criteria were: alcohol or drug addiction, pregnancy, specific 82

musculoskeletal/neurological disease, surgery within the past year, plans for surgery in the future, and primary 83

psychiatric disorder. Further details regarding recruitment and interventions have been described previously [11]. 84

Context (Danish legislation on sickness benefits) 85

According to The Danish Sickness Benefit Act, sickness absentees are entitled to financial compensation during sick 86

leave. This compensation can amount to sickness benefits or full wages depending on the union award conditions 87

the worker is entitled to. At the time of the RCT (2009-2014), the employer covered these expenses for the first 21-88

30 days of a sick leave spell (called the employer paid period) after which the municipality reimbursed some of the 89

employer's expenses (the sickness benefit amount) until RTW was achieved. In Denmark, a special agreement (§56) 90

can be settled upon in cases of recurrent short-term sick leave; it entitles the employer to be reimbursed from the 91

first day of sick leave, thereby reducing the employer's expenses [12]. 92

Outcome: Work participation score (WPS) 93

Data on work outcomes were supplied by the Danish Register for Evaluation of Marginalization (DREAM) which 94

covers the source of income (financial self-support or public transfer benefits) on a weekly basis since July 1991. 95

The registrations are ordered hierarchically; if different transfer benefit codes are registered in the same week, the 96

highest-ranking code will overwrite the others. Sickness benefits have a high priority and are only overwritten in 97

cases of emigration, age-related pensions or death [13]. 98

The work participation score (WPS) is a fraction yielding scores between 0 and 100% [14]. The numerator 99

comprises the number of weeks with financial self-support (interpreted as working weeks) as well as weeks with 100

state education fund grants and related benefits (benefits granted in cases of re-education). The denominator 101

5

comprises the total number of follow-up weeks [14]. In this study, three individuals were granted early retirement. 102

For these three, the denominator comprised the number of follow-up weeks prior to the first appearance of 103

retirement benefits in DREAM. Since the median time until the first RTW (4 consecutive weeks of financial self-104

support) was 29 weeks, follow-up time comprised the Weeks 30-104 and a successful outcome was defined as a 105

WPS ≥ 75%. The 75% threshold was chosen because it has shown moderate to substantial agreement compared with 106

other RTW measures [14] and because it defined an improvement compared with the mean WPS the year before 107

enrolment (74%). The following denotations were used: successful WPS (s-WPS) ≥ 75% and unsuccessful WPS (u-108

WPS) < 75%. 109

Prognostic variables: Demographic and patient-reported domain 110

These data were collected from questionnaires filled in by the participants prior to their first meeting with the health 111

care professionals in the RCT [11]. The information provided covered demographics (sex and age) and the 112

following patient-reported measures: pain intensity (11-point numeric rating scale) [15], whiplash trauma (yes/no) 113

[16], education (≥ 3 years, yes/no), worker's compensation claim (yes/no) [17], and Örebro Musculoskeletal Pain 114

Screening Questionnaire (ÖMPSQ) [18]. ÖMPSQ was originally developed as a screening tool to identify 115

psychosocial risk factors for prolonged disability and sick leave; the higher the score (maximum = 210), the higher 116

the risk of prolonged disability. 117

Prognostic variables: Clinical domain 118

Data on radiculopathy and number of tender points [19, 20] stemmed from the clinical examination performed by 119

the rheumatologist prior to enrolment in the RCT [11]. Radiculopathy describes a clinical assessment based on 1) 120

pain radiating from the neck to the upper extremity and 2) one or more positive neurological signs: weakened deep 121

tendon reflexes, decreased muscle strength, dermatomal sensory deficits or a positive foraminal compression test. 122

The reliability of the rheumatologist's tender points count has previously been reported as excellent [21] and a 123

threshold of ≥ 11 tender points was chosen, since this is still commonly used among clinicians to assess 124

fibromyalgia [22]. In the clinical examination, the tender points count was used as a measure of sensitization of the 125

nociceptive system [23]. Since data were not sufficient to support the diagnosis of fibromyalgia, this term was not 126

used and a dichotomized threshold of the tender points count was used instead. The interventions in the original 127

RCT [11] (multidisciplinary and brief intervention) were also included. 128

6

Prognostic variables: MRI domain 129

MRI of the cervical spine was ordered for all participants unless clinical examination revealed unambiguous signs of 130

shoulder pathology (n = 7). For different reasons, MRI of the cervical spine was only available for 97 of the 131

participants (no access to MRI: n = 32 and unsatisfactory signal-to-noise ratio assessed by a senior radiologist: n = 132

32 (unpublished). All MRIs were de-identified and assessed by the first author (physician with 9 years of post-133

graduate clinical experience) who was blinded to the study outcomes. The inter-rater reliability of the MRI 134

assessments has previously been reported with un-weighted kappa values ranging from 0.56 to 0.79 (unpublished) 135

while intra-rater reliability yielded kappa values ranging from 0.67 to 0.90 (unpublished). The MRI variables listed 136

in Table 1 were evaluated at disc levels C2/C3 to C7/T1. For each participant, an MRI finding was defined as 137

positive if the finding was present at ≥ 1 disc level, i.e. two levels with disc height reduction were counted only 138

once. 139

Statistical Analyses 140

Crude logistic regression analyses were performed for all baseline variables hypothesized to affect WPS, and 141

variables yielding p-values < 0.2 were carried forward as has been done in other studies [24, 25]. In the 142

multivariable analyses, the prognostic variables were included using a domain-wise approach resembling the typical 143

order of information obtained in a clinical setting. First, demographic variables were included, second patient-144

reported, third clinical, and finally MRI variables – these are referred to as Models 1, 2, 3 and 4. For each model, the 145

probability of s-WPS was calculated and dichotomized, i.e. the participants were classified as having ≥ 50% chance 146

of s-WPS or not. By comparing the estimated chance of s-WPS with whether s-WPS actually happened or not, 147

sensitivity, specificity, positive and negative predictive values (PPV and NPV), and area under the curve (AUC) 148

were calculated. In this study, sensitivity denotes the proportion of u-WPS correctly classified as '<50% chance of s-149

WPS', and specificity, the proportion of s-WPS that is correctly classified as '≥50% chance of s-WPS'. In accordance 150

with this, NPV denotes the proportion of sickness absentees classified as ' ≥50% s-WPS chance' who actually 151

achieved s-WPS. PPV denotes the proportion of sickness absentees classified as '<50% s-WPS chance'' who did not 152

achieve s-WPS. AUC is a measure of discriminatory ability which describes the probability that a worker achieving 153

s-WPS will be classified as having a better chance than a worker not achieving s-WPS. To ensure comparability 154

between the models, estimates were presented both for the number of individuals with complete data and for the 155

number of individuals with complete data on the adjacent model (e.g. in Model 3: estimates were presented for the 156

7

161 individuals with complete data on demographic, patient-reported and clinical domains and for the 93 individuals 157

with complete data on the MRI domain (Model 4)). For each added domain of information, sensitivity, specificity, 158

PPV and NPV were compared between the models. To further appraise the properties of the models, the chance of s-159

WPS was divided into four categories (< 30%, ≥ 30% and < 50%, ≥ 50% and < 70%, ≥ 70%) and tabulations were 160

made to assess whether adding domains of information improved prediction for those with the lowest (< 30%) and 161

highest (≥ 70%) chance of s-WPS. Estimates were reported using 95% confidence intervals (CIs). 162

The statistical analyses were performed by the first author who was not involved in the clinical assessment or 163

interventions given in the original RCT [11]. STATA15 statistical software package was used for all statistical 164

analyses (STATA Corp., College Station, Texas, USA). 165

Nomenclature 166

This study explored the associations between prognostic variables and s-WPS. The term "prognostic variable" 167

denotes a baseline variable which is associated with the outcome s-WPS but without making inferences about 168

causality [26]. The terms 'predict'/'prediction' relate to the ability of the models to forecast the prognosis regarding 169

work participation. 170

Ethics 171

All participants provided written informed consent. The study was approved by The Danish Data Protection Agency 172

(J. no. 1-16-02-86-16) and by The Central Denmark Region Committees on 173

Health Research Ethics (M-20090027). 174

Results 175

One hundred-and-sixty-eight sickness absentees were included. S-WPS was achieved by 80 (47.6%) of the sickness 176

absentees, i.e. 47.6% of the sickness absentees had a work participation score ≥ 75% from Week 30 to 104 after 177

enrolment. The proportion achieving s-WPS was similar (47.4%) in the MRI sample. Baseline characteristics are 178

presented in Table 1, and were similar in the MRI sample compared with the entire cohort (data not shown). 179

The crude associations between potential prognostic variables and s-WPS are presented in Table 2. For those 180

variables yielding p < 0.2, the following associations were seen: Increased odds of s-WPS were found for male sex, 181

radiculopathy and kyphosis. Decreased odds of s-WPS were found for increased sick leave duration (months), 182

8

increased ÖMPSQ score, ≥ 11 tender points, and spinal canal stenosis. The remaining variables from Table 2 did not 183

fulfill the criterion of p < 0.2 for inclusion in the multivariable analyses. 184

Table 3 presents the predictive values from Models 1- 4. As shown, knowledge about only the sex offered an NPV 185

of 0.60 (0.46; 0.74) and a PPV of 0.58 (0.49; 0.67). By adding patient-reported variables (duration of sick leave and 186

ÖMPSQ score), the NPV increased to 0.70 (0.59; 0.80) and the PPV to 0.72 (0.62; 0.81) along with increased 187

specificity. Adding clinical variables elicited only minor changes in NPV and sensitivity while the AUC changed in 188

the second decimal. Further addition of MRI variables did not improve any of the values. 189

Additional detail regarding Models 1- 4 is offered by Table 4 where the chance of s-WPS in four categories is 190

compared with actual WPS outcomes. Model 1 did not identify anybody with < 30% or ≥ 70% of s-WPS, whereas 191

Model 2 performed better than Model 1, discriminating among those whose chance of s-WPS was very low or very 192

high. The discriminating properties of the models were similar for Models 2, 3 and 4 (assessed by the distribution of 193

chance). 194

Discussion 195

To our knowledge, this is the first study exploring the contribution of demographics, patient-reported, clinical and 196

MRI information to the prediction of work participation in sickness absentees with neck or shoulder pain. The 197

results suggest that demographics and patient-reported measures include important prognostic information (sex, 198

duration of sick leave prior to enrollment and ÖMPSQ score) and that neither clinical nor MRI information provide 199

substantial additional information to the prediction of work participation within 2 years. That is, Model 2 performed 200

similarly to Models 3 and 4 regarding sensitivity, specificity, NPV, PPV and AUC. Model 2 also performed 201

similarly to Models 3 and 4 at discriminating between those having the lowest and highest chance of s-WPS (Table 202

4). 203

Comparison with other studies: 204

Only 47.6% of the sickness absentees achieved s-WPS within two-year follow up, which is indicative of a 205

population with a high level of work disability. In a large study of neck and back pain [27], only 8% reported sick 206

leave the previous year when asked at 3-year follow up. The study [27] was conducted in a workplace setting, hence, 207

study participants were presumably less disabled than those referred to secondary care settings as in the current 208

study. From secondary care settings however, previous studies [6, 28] have also found what seemed to be better 209

9

work outcomes than in the current study. In a study of low back pain, 72% achieved RTW within 1 year, however 210

RTW was defined as 4 consecutive weeks of work [6]. Another study of primarily musculoskeletal pain used a more 211

sustainable RTW outcome, namely 3 consecutive months with increased working time compared to baseline [28]. 212

Sixty percent had achieved this outcome at 1 year follow up and percent-wise more comparable with our study 213

findings. Differences in population characteristics, setting or the outcome measures chosen may explain the wide 214

range of work participation following musculoskeletal pain. Biering et al. suggested that for prognostic purposes the 215

WPS was superior to time-to-event-based measures and measures at fixed time-points [14]. As we used the WPS in 216

the current study and included the study population from a secondary care setting, we believe that the fact that less 217

than half of the study participants achieved s-WPS is indicative of work disability that hampers work participation. 218

A comprehensive literature search strategy adapted from a 2017 review [2] was performed using Medline and yet, 219

no studies were identified which categorized prognostic variables and used an analytical approach resembling the 220

order of information obtained in a clinical setting. For this reason, direct comparison with previous studies was not 221

possible, but some studies reported measures that can be compared with the current study [29, 30]. In a Norwegian 222

study [29], the sickness absentees' own prediction of sick leave duration ≥ 26 weeks yielded a slightly better PPV 223

(0.78) than any of the models in the current study. However, sensitivity based on self-prediction was 0.28 and 224

sensitivity based on dedicated medical consultants was 0.07. Both measures are notably lower than the sensitivity 225

estimates of the current study suggesting that 1) perhaps the sickness absentees in Fleten et al.'s study were over-226

optimistic, leading to low sensitivity of self-prediction, and 2) the information that was available to the medical 227

consultants [29] was inadequate to result in a sensitivity as high as in the current study. 228

Rehabilitation professionals in another study [30] predicted the chance of RTW based on sick leave duration, reason 229

for sick leave, unemployment, age > 45 years, female sex and ‘gut feeling’. Their prediction was concordant with 230

actual RTW in 73% and thus similar to the specificity estimates of the current study. Of interest is that the 231

rehabilitation professionals in that study [30] had access to some similar information as was available in the current 232

study, namely sex and sick leave duration. Their access to further information about reason for sick leave, 233

unemployment, age > 45 and "gut feeling" did not yield a higher specificity than in the current study. 234

Regarding the association between sex and work outcomes, results in previous studies have been conflicting. In a 235

population-based study [31] and in studies from secondary care on spinal pain [32] and shoulder pain [33], sex was 236

not associated with work outcomes, while a primary care study [24] did find predictive value of sex. In summary, 237

10

reviews have found conflicting evidence on the impact of sex on work outcomes [1, 3] and with this in mind, the 238

current study's estimates of PPV and NPV when based on sex alone are not surprising. 239

The impact of sick leave duration corroborates the findings in numerous reviews [1, 3, 7, 34, 35] and cohort studies 240

of shoulder pain, spinal pain and whiplash trauma [24, 36, 37]. Due to the non-modifiable nature of sick leave 241

duration, our findings highlight the need for action that prevents and/or addresses sick leave early enough to 242

minimize the duration. 243

The negative association between the ÖMPSQ score and s-WPS is in line with previous findings [18, 38]. While 244

studies from primary care settings and workplaces [39, 40] have suggested ÖMPSQ cut-off values of 90 and 105, 245

the current study suggests that in a secondary care setting, ÖMPSQ has important properties regarding prognosis of 246

work participation also as a numerical variable. 247

Interestingly, a study of primarily chronic spinal pain suggested cut-offs of 90 and 105 [40] but did not yield 248

sensitivity and specificity that were concurrently as high as any of the models including ÖMPSQ in the current 249

study. In contrast, a primary care study of patients with acute or sub-acute spinal pain [39], a cut-off of 90 yielded 250

better sensitivity and slightly lower specificity [39], thereby demonstrating the properties of ÖMPSQ in early 251

detection of poor prognosis. 252

In the context of clinical variables, a cohort study on low back pain [41] found a crude association between tender 253

points count (numeric variable) and unsuccessful RTW. However, this was not maintained in the multivariable 254

model. Moreover, no association was found between radiculopathy and work outcomes [41], which is supported by 255

a review presenting moderate evidence for no association between radiating pain and RTW [2]. Therefore, our 256

finding of the limited improvement of predictive values in Model 3 is in line with these previous findings [2, 41]. 257

Finally, regarding MRI, only one study was identified which explored the association between MRI findings and 258

work outcomes [6]. In this study of sickness absentees with low back pain, the presence of Modic type 1 changes 259

was associated with unsuccessful RTW. Unfortunately, we were unable to assess the impact of Modic changes in the 260

current study since the inter- and intra-rater reliability for this pathology was not established owing to too low 261

prevalences (unpublished). 262

Methodological considerations 263

WPS was chosen as the outcome measure in an attempt to capture both work reintegration and work maintenance 264

although we are aware that other RTW outcomes reflect other aspects regarding the process and context of RTW 265

11

[42]. The chosen cut-off value (s-WPS ≥ 75% or not) showed moderate to substantial agreement with other RTW 266

measures (RTW yes/no at a given time point and time to 4 weeks of self-support without relapses, respectively) in a 267

previous study [14]. As previously argued, the same study [14]found WPS suitable for prognostic purposes. 268

When classifying the chance of s-WPS, a 50% threshold was chosen for two reasons: 1) A reasonable balance 269

between sensitivity and specificity measures was desired and 2) Our clinical experience suggests that sickness 270

absentees with neck or shoulder pain who are seen in secondary care are troubled with many biopsychosocial issues. 271

Hence, choosing higher cut-offs for the chance of s-WPS might be too optimistic. A different cut-off would 272

probably be relevant if the study population had comprised patients from primary care on short-term sick leave. 273

Strengths 274

First, the use of registry data for the outcome assessment ensured 100% follow up thereby reducing the risk of 275

attrition bias. Second, the use of registry data reduced the risk of measurement bias since the outcome assessment 276

was unaffected by knowledge of the prognostic variables. Likewise, the assessment of prognostic factors was 277

unaffected by knowledge of the outcome. Third, the risk of attrition bias was minimized by the low number of 278

missing values for the majority of variables (demographic, patient-reported and clinical variables). Only for MRI 279

was the number of missing values substantial. However, the distribution between exposure and outcome was 280

assessed for all the variables in Table 1, which revealed that attrition was not skewed for any of the variables, hence 281

attrition bias was not suspected (data not shown). Fourth, the duration of follow-up reaching 2 years constitutes a 282

realistic long-term outcome. A fifth strength is the a priori decision to take an analytical approach resembling the 283

working conditions of the clinicians who are responsible for appraisal of work prognosis. This reduced the risk of 284

purely data-driven results. Finally, when bearing in mind the well-established impact of external societal factors on 285

the process of sick leave [43, 44], it is a further strength of the current study that the Danish legislation on sickness 286

benefits did not undergo major changes from 2009-2014. A change in legislation in 2014 reduced the right for 287

sickness benefits from 52 to 22 weeks. It affected only six sickness absentees (equally distributed between s-WPS 288

and u-WPS) and is thus not suspected to bias the estimates. 289

Limitations 290

First, since the study population was originally included in an RCT, the results may not be generalizable to all 291

sickness absentees with neck or shoulder pain. The possibility of sampling bias due to referral patterns of the general 292

12

practitioners cannot be refuted nor elucidated. Hence, the results are expected to be representative of sickness 293

absentees with neck or shoulder pain who are seen in a secondary care setting. Generalizability to primary care 294

settings should be made with caution and confirmatory studies including all sickness absentees with neck or 295

shoulder pain would be needed to improve generalizability. 296

Second, the representativeness of the results should be considered in the light of the above-mentioned change in 297

legislation, implying that the right to sickness benefits is now limited to 22 weeks compared with 52 weeks during 298

most of the study. Confirmatory studies would be desirable to assess the impact of this change in legislation. But 299

given that the literature over the past decades is corroborated (i.e. major impact of sick leave duration and 300

psychosocial factors [45]), we expect that similar results would be found. 301

Third, the limited MRI sample affects precision of the estimates; an issue that was further attenuated for certain MRI 302

variables (disc bulge/protrusion/extrusion, neural foraminal stenosis, zygapophyseal osteoarthritis and uncovertebral 303

osteoarthritis). For these variables, the available sequences did not allow for evaluation on all 97 MRIs (Table 1 and 304

(unpublished). Had MRI been available for the entire cohort, confidence intervals for MRI variables would have 305

been narrowed down. 306

Fourth, the use of registry data is usually an advantage, but the data source also warrants consideration since 307

registration procedures imply a risk of unequal registration of short-term sick leave. Sick leave registration in 308

DREAM begins at the end of the employer paid period and backward adjustments are made, so that the number of 309

sickness benefit weeks in DREAM equal the total number of sick leave weeks. Since registration is initiated at the 310

end of the employer paid period, multiple absences due to sick leave lasting only days or a few weeks are usually 311

not registered (because they are within the employer-paid period). In the current study, some participants may have 312

been misclassified as having achieved s-WPS although they had multiple short-term absences due to sick leave. A 313

related problem applies to employees with a §56-agreement, i.e. their employers are entitled to reimbursement of 314

sickness benefits from day one; such employees may be misclassified as u-WPS in the current study due to sick 315

leave registrations exceeding actual duration. These possible misclassifications are considered non-differentiated 316

since they are not suspected to be associated with the exposure (i.e. the prognostic variables). Unfortunately, the 317

data do not offer any insight as to the possible distribution of §56-agreements in the current study. However, 318

explorative post-hoc analyses were performed in which single weeks of sickness benefit reimbursement were 319

considered as §56-agreements, i.e. regarded as working weeks. This resulted in the distribution of s-WPS/u-WPS 320

changing from 80/88 to 83/85 (data not shown). It did not change the results of the study. 321

13

Sixth, the limited sample size implies a risk of type 2 errors. That is, important prognostic factors may not be 322

discovered as statistically significant owing to the limited sample size. Other factors that could affect the prognosis 323

for work outcomes include fear avoidance beliefs [2, 3, 45] and physical workload [1, 2, 35]. These factors were not 324

isolated for analysis in the current study since they were covered by the ÖMPSQ score and we wished for all 325

variables to be as mutually exclusive as possible. Furthermore, the aim was not to investigate a complete list of all 326

possible factors affecting WPS prognosis but rather to explore the contributions of demographic, patient-reported, 327

clinical and MRI variables for which purpose we believe the current sample size was adequate. 328

Conclusion 329

In sickness absentees with neck or shoulder pain, clinical and MRI information provide no additional information 330

for the prediction of work participation compared with only demographic and patient-reported information. 331

Though Model 2 performs similarly to Models 3 and 4, there are limitations with respect to discriminating between 332

those with the lowest and highest chance of s-WPS (Table 4). This means that clinical applicability requires caution 333

if the aim is identification of those who have the lowest/highest chances of s-WPS. The results do, however, provide 334

valuable knowledge to clinicians both in the assessment of work prognosis and in dialogue with patients and other 335

stakeholders: prediction should primarily be based on demographic and patient-reported information, not on clinical 336

and MRI findings. 337

Due to the exploratory nature of the current study, the generalizability of findings needs to be assessed in 338

independent studies, preferably in both primary and secondary care settings. Finally, from a public health 339

perspective, there is a need for management of neck and shoulder pain and identification of those at risk of poor 340

work prognosis early enough to avoid harmful long-term absences due to sick leave. 341

342

All procedures followed were in accordance with the ethical standards of The Central Denmark Region Committees 343

on 344

Health Research Ethics. Informed consent was obtained from all participants included in the study. 345

Abbreviations 346

AUC: Area Under the Curve, MRI: Magnetic Resonance Imaging, NPV: Negative Predictive Value, ÖMPSQ: 347

Örebro Musculoskeletal Pain Screening Questionnaire, PPV: Positive Predictive Value, RCT: Randomized 348

14

Controlled Trial, RTW: Return To Work, s-WPS: successful Work Participation Score, u-WPS: unsuccessful Work 349

Participation Score, WPS: Work Participation Score. 350

Tables 351

Table 1 Baseline characteristics in sickness absentees with neck or shoulder pain. 352

Table 2 Odds ratios for work participation score ≥ 75% (s-WPS) in sickness absentees with neck or shoulder pain. 353

Table 3 Predictive values, sensitivity, specificity and AUC for the prediction of work participation score < 75% (u-354

WPS) or ≥ 75% (s-WPS) in 4 models reflecting type of prognostic information. 355

Table 4 Cross-tabulations of the distribution of chance vs. achieved work participation score ≥ 75% (s-WPS) or < 356

75% (u-WPS). 357

Authors' contributions 358

Line Thorndal Moll: Design of the study, analysis and interpretation of data, primary responsible for the manuscript. 359

Anne Mette Schmidt: Interpretation of data, critical appraisal of the manuscript. 360

Christina Malmose Stapelfeldt: Design of the study, interpretation of data, critical appraisal of the manuscript. 361

Merete Labriola: Design of the study, interpretation of data, critical appraisal of the manuscript. 362

Ole Kudsk Jensen: Design of the study, interpretation of data, critical appraisal of the manuscript. 363

Morten Wasmod Kindt: Interpretation of data, critical appraisal of the manuscript. 364

Tue Secher Jensen: Design of the study, interpretation of data, critical appraisal of the manuscript. 365

Berit Schiöttz-Christensen: Design of the study, interpretation of data, critical appraisal of the manuscript. 366

Conflicts of interest 367

Author Line Thorndal Moll declares that she has no conflict of interest. 368

Author Anne Mette Schmidt declares that she has no conflict of interest. 369

Author Christina Malmose Stapelfeldt declares that she has no conflict of interest. 370

Author Merete Labriola declares that she has no conflict of interest. 371

Author Ole Kudsk Jensen declares that he has no conflict of interest. 372

Author Morten Wasmod Kindt declares that he has no conflict of interest. 373

Author Tue Secher Jensen declares that he has no conflict of interest. 374

Author Berit Schiöttz-Christensen declares that she has no conflict of interest. 375

15

Table 1 Baseline characteristics in sickness absentees with neck or shoulder pain. 376

377

378

379

380

Demographics N

Male sex, n (%) 168 53 (31.5)

Age, median (IQR) 168 40.5 (34.5 ; 48.9)

Patient-reported information

Pain intensity (0-10) last week, median (IQR) 158 7 (5 ; 8)

Whiplash trauma, n (%) 168 24 (14.3)

Sick leave duration (months), median (IQR) 168 2.3 (1.2 ; 3.0)

ÖMPSQ score, mean (SD) 161 122 (23.7)

Education ≥ 3 years, n (%) 155 29 (18.7)

Ongoing worker's compensation claim, n (%) 139 36 (25.9)

Clinical information

Radiculopathy, n (%) 168 41 (24.4)

≥ 11 tender points, n (%) 168 57 (33.9)

Intervention group

Brief intervention, n (%)

Multidisciplinary intervention, n (%) 168

83 (49.4)

85 (50.6)

MRI information

Kyphosis, n (%) 97 23 (23.7)

Disc height reduction, n (%) 97 71 (73.2)

Bulge, protrusion or extrusion, n (%) 83 69 (83.1)

Spinal canal stenosis, n (%) 97 14 (14.4)

Neural foraminal stenosis, n (%) 84 46 (54.8)

Zygapophyseal osteoarthritis, n (%) 83 39 (47.0)

Uncovertebral osteoarthritis, n (%) 78 37 (47.4)

16

381 382

17

383 384

18

Table 4. Cross-tabulations of the distribution of chance vs. achieved work participation score ≥ 75% (s-WPS) or <

75% (u-WPS).

Domains included N* Outcome

Demographics* 168

Chance of s-WPS s-WPS u-WPS

< 30% 0 0

30% to 49% 48 67

50% to 70% 32 21

≥ 70% 0 0

Demographic + patient-

reported** 161

Outcome


< 30% 11 33

30% to 49% 13 30

50% to 70% 28 19

≥ 70% 24 3


reported + clinical*** 161

Outcome


< 30% 11 33

30% to 49% 14 34

50% to 70% 27 14

≥ 70% 24 4


reported + clinical +

MRI****

93

Outcome


< 30% 5 23

30% to 49% 8 12

50% to 70% 17 12

≥ 70% 13 3

*: The reduction of N in all models is explained by the number of individuals with available ÔMPSQ score for the

given models.

385

386

19

References 387

1. Cancelliere C, Donovan J, Stochkendahl MJ, Biscardi M, Ammendolia C, Myburgh C, et al. Factors affecting return 388

to work after injury or illness: Best evidence synthesis of systematic reviews. Chiropr Man Therap. 2016 Sep 389

8;24(1):32,016-0113-z. eCollection 2016. 390

2. Steenstra IA, Munhall C, Irvin E, Oranye N, Passmore S, Van Eerd D, et al. Systematic review of prognostic factors 391

for return to work in workers with sub acute and chronic low back pain. J Occup Rehabil. 2017 Sep;27(3):369-81. 392

3. Rinaldo U, Selander J. Return to work after vocational rehabilitation for sick-listed workers with long-term back, 393

neck and shoulder problems: A follow-up study of factors involved. Work. 2016 Sep 27;55(1):115-31. 394

4. Hulshof C, Pransky G. Handbook of work disability, chapter 13. In: Loisel P, Anema JR, editors. Handbook of 395

Work Disability. 1st ed. New York: Springer Science+Business Media; 2013. p. 203. 396

5. Palstam A, Mannerkorpi K. Work ability in fibromyalgia: An update in the 21st century. Curr Rheumatol Rev. 397

2017;13(3):180-7. 398

6. Jensen OK, Nielsen CV, Sorensen JS, Stengaard-Pedersen K. Type 1 modic changes was a significant risk factor for 399

1-year outcome in sick-listed low back pain patients: A nested cohort study using magnetic resonance imaging of 400

the lumbar spine. Spine J. 2014 Nov 1;14(11):2568-81. 401

7. Carroll LJ, Hogg-Johnson S, Cote P, van der Velde G, Holm LW, Carragee EJ, et al. Course and prognostic factors 402

for neck pain in workers: Results of the bone and joint decade 2000-2010 task force on neck pain and its 403

associated disorders. Spine (Phila Pa 1976). 2008 Feb 15;33(4 Suppl):S93-100. 404

8. Christensen AI, Davidsen M, Juel K. The national health profile 2013. 2014 March 5th 2014. 405

9. Jensen HAR, Davidsen M, Christensen AI. The national health profile 2017. 2018 March 6th 2018. 406

10. National danish patient registry [homepage on the Internet]. . Available from: 407

http://www.esundhed.dk/sundhedsregistre/LPR/Sider/LPR04_Tabel.aspx. 408

11. Moll LT, Jensen OK, Schiottz-Christensen B, Stapelfeldt CM, Christiansen DH, Nielsen CV, et al. Return to work 409

in employees on sick leave due to neck or shoulder pain: A randomized clinical trial comparing multidisciplinary 410

and brief intervention with one-year register-based follow-up. J Occup Rehabil. 2017 Aug 23. 411

12. Danish sickness benefit act [homepage on the Internet]. . 2014. Available from: 412

https://www.retsinformation.dk/Forms/R0710.aspx?id=161736#Kap8. 413



20

13. Stapelfeldt CM, Jensen C, Andersen NT, Fleten N, Nielsen CV. Validation of sick leave measures: Self-reported 414

sick leave and sickness benefit data from a danish national register compared to multiple workplace-registered sick 415

leave spells in a danish municipality. BMC Public Health. 2012 Aug 15;12:661,2458-12-661. 416

14. Biering K, Hjollund NH, Lund T. Methods in measuring return to work: A comparison of measures of return to 417

work following treatment of coronary heart disease. J Occup Rehabil. 2013 Sep;23(3):400-5. 418

15. Bolton JE, Humphreys BK, van Hedel HJ. Validity of weekly recall ratings of average pain intensity in neck pain 419

patients. J Manipulative Physiol Ther. 2010 Oct;33(8):612-7. 420

16. Biering-Sorensen S, Moller A, Stoltenberg CD, Holm JW, Skov PG. The return-to-work process of individuals 421

sick-listed because of whiplash-associated disorder: A three-year follow-up study in a danish cohort of long-term 422

sickness absentees. BMC Public Health. 2014 Feb 4;14:113,2458-14-113. 423

17. Loisel P, Cote P. Handbook of work disability, chapter 5. In: Loisel P, Anema JR, editors. 1st ed. New York: 424

Springer Science+Business Media; 2013. p. 59. 425

18. Karran EL, McAuley JH, Traeger AC, Hillier SL, Grabherr L, Russek LN, et al. Can screening instruments 426

accurately determine poor outcome risk in adults with recent onset low back pain? A systematic review and meta-427

analysis. BMC Med. 2017 Jan 19;15(1):13,016-0774-4. 428

19. Kasch H, Qerama E, Kongsted A, Bach FW, Bendix T, Jensen TS. Deep muscle pain, tender points and recovery in 429

acute whiplash patients: A 1-year follow-up study. Pain. 2008 Nov 15;140(1):65-73. 430

20. Jensen OK, Nielsen CV, Stengaard-Pedersen K. One-year prognosis in sick-listed low back pain patients with and 431

without radiculopathy. prognostic factors influencing pain and disability. Spine J. 2010 Aug;10(8):659-75. 432

21. Jensen OK, Callesen J, Nielsen MG, Ellingsen T. Reproducibility of tender point examination in chronic low back 433

pain patients as measured by intrarater and inter-rater reliability and agreement: A validation study. BMJ Open. 434

2013 Feb 26;3(2):10.1136/bmjopen,2012-002532. Print 2013. 435

22. Wolfe F, Smythe HA, Yunus MB, Bennett RM, Bombardier C, Goldenberg DL, et al. The american college of 436

rheumatology 1990 criteria for the classification of fibromyalgia. report of the multicenter criteria committee. 437

Arthritis Rheum. 1990 Feb;33(2):160-72. 438

23. Clauw DJ. Chapter 77. fibromyalgia and related syndromes. In: Hochberg MC, Silman AJ, Smolen JS, Weinblatt 439

ME, Weisman MH, editors. Rheumatology. 5th ed. Philadelphia: Elsevier Ltd; 2011. p. 769. 440

21

24. Kuijpers T, van der Windt DA, van der Heijden GJ, Twisk JW, Vergouwe Y, Bouter LM. A prediction rule for 441

shoulder pain related sick leave: A prospective cohort study. BMC Musculoskelet Disord. 2006 Dec 6;7:97,2474-7-442

97. 443

25. Rudbeck M, Jensen SL, Fonager K. Arthroscopic subacromial decompression and predictors of long-term sick 444

leave benefit and permanent benefits. J Shoulder Elbow Surg. 2013 Sep;22(9):1167-72. 445

26. Herbert RD. Cohort studies of aetiology and prognosis: They're different. J Physiother. 2014 Dec;60(4):241-4. 446

27. Bergstrom G, Bodin L, Bertilsson H, Jensen IB. Risk factors for new episodes of sick leave due to neck or back 447

pain in a working population. A prospective study with an 18-month and a three-year follow-up. Occup Environ 448

Med. 2007 Apr;64(4):279-87. 449

28. Brendbekken R, Vaktskjold A, Harris A, Tangen T. Predictors of return-to-work in patients with chronic 450

musculoskeletal pain: A randomized clinical trial. J Rehabil Med. 2018 Feb 13;50(2):193-9. 451

29. Fleten N, Johnsen R, Forde OH. Length of sick leave - why not ask the sick-listed? sick-listed individuals predict 452

their length of sick leave more accurately than professionals. BMC Public Health. 2004 Oct 12;4:46,2458-4-46. 453

30. von Celsing AS, Svardsudd K, Wallman T. Predicting return to work among sickness-certified patients in general 454

practice: Properties of two assessment tools. Ups J Med Sci. 2014 Aug;119(3):268-77. 455

31. Haukka E, Ojajarvi A, Kaila-Kangas L, Leino-Arjas P. Protective determinants of sickness absence among 456

employees with multisite pain-a 7-year follow-up. Pain. 2017 Feb;158(2):220-9. 457

32. Elfving B, Asell M, Ropponen A, Alexanderson K. What factors predict full or partial return to work among 458

sickness absentees with spinal pain participating in rehabilitation? Disabil Rehabil. 2009;31(16):1318-27. 459

33. Engebretsen K, Grotle M, Bautz-Holter E, Ekeberg OM, Brox JI. Predictors of shoulder pain and disability index 460

(SPADI) and work status after 1 year in patients with subacromial shoulder pain. BMC Musculoskelet Disord. 2010 461

Sep 23;11:218,2474-11-218. 462

34. Carroll LJ, Hogg-Johnson S, van der Velde G, Haldeman S, Holm LW, Carragee EJ, et al. Course and prognostic 463

factors for neck pain in the general population: Results of the bone and joint decade 2000-2010 task force on neck 464

pain and its associated disorders. Spine (Phila Pa 1976). 2008 Feb 15;33(4 Suppl):S75-82. 465

35. Desmeules F, Braen C, Lamontagne M, Dionne CE, Roy JS. Determinants and predictors of absenteeism and 466

return-to-work in workers with shoulder disorders. Work. 2016 Sep 27;55(1):101-13. 467

22

36. Carstensen TB, Fink P, Oernboel E, Kasch H, Jensen TS, Frostholm L. Sick leave within 5 years of whiplash 468

trauma predicts recovery: A prospective cohort and register-based study. PLoS One. 2015 Jun 22;10(6):e0130298. 469

37. Lindell O, Johansson SE, Strender LE. Predictors of stable return-to-work in non-acute, non-specific spinal pain: 470

Low total prior sick-listing, high self prediction and young age. A two-year prospective cohort study. BMC Fam 471

Pract. 2010 Jul 20;11:53,2296-11-53. 472

38. Hockings RL, McAuley JH, Maher CG. A systematic review of the predictive ability of the orebro musculoskeletal 473

pain questionnaire. Spine (Phila Pa 1976). 2008 Jul 1;33(15):E494-500. 474

39. Linton SJ, Boersma K. Early identification of patients at risk of developing a persistent back problem: The 475

predictive validity of the orebro musculoskeletal pain questionnaire. Clin J Pain. 2003 Mar-Apr;19(2):80-6. 476

40. Bergstrom G, Hagberg J, Busch H, Jensen I, Bjorklund C. Prediction of sickness absenteeism, disability pension 477

and sickness presenteeism among employees with back pain. J Occup Rehabil. 2014 Jun;24(2):278-86. 478

41. Jensen OK, Stengaard-Pedersen K, Jensen C, Nielsen CV. Prediction model for unsuccessful return to work after 479

hospital-based intervention in low back pain patients. BMC Musculoskelet Disord. 2013 Apr 19;14:140,2474-14-480

140. 481

42. Wasiak R, Young AE, Roessler RT, McPherson KM, van Poppel MN, Anema JR. Measuring return to work. J 482

Occup Rehabil. 2007 Dec;17(4):766-81. 483

43. Loisel P, Durand M-, Berthelette D, Vézina N, Baril R, Gagnon D, et al. Disability prevention. new paradigm for 484

the management of occupational back pain. Dis Manage Health Outcomes. 2001;9(7):351-360. 485

44. Werner EL, Cote P. Low back pain and determinants of sickness absence. Eur J Gen Pract. 2009;15(2):74-9. 486

45. Bültmann U, Brouwer S. Handbook of work disability, chapter 10. In: Loisel P, Anema J, editors. 1st ed. New 487

York: Springer Science+Business Media; 2013. p. 149. 488

489

Evaluation manual for the inter- and intra-rater reliability study on MRI of the cervical

spine

This evaluation manual covers all the definitions and classifications that are used for the above mentioned reliability study. In this evaluation manual, the names of the variables are so-called 'generic' names, whereas variables in Epidata have appropriate prefixes to determine the cervical level in question. Likewise, for the neural foramina, the uncovertebral and zygapophyseal joints, suffixes in Epidata are used to denote whether it is on the left or the right hand side. Six disc levels are examined from C2/C3 to C7/T1. ____________________________________________________________________ Kyphosis 0 : normal or reduced lordosis 1 : kyphosis 9 : missing Cervical alignment is assessed on mid-sagittal T2-weighted images. A line from the posteroinferior aspect of vertebra C2 to the posteroinferior aspect of vertebra C7 is drawn. If any part of C3-C6 lie behind/posterior to this line, the alignment is classified as kyphosis (1). ____________________________________________________________________ Disc height 0 : normal disc height 1 : reduced disc height 9 : missing Disc height is assessed on T2-weighted sagittal images.

inghor

Typewritten Text

inghor

Typewritten Text

inghor

Typewritten Text

Appendix 4

inghor

Typewritten Text

inghor

Typewritten Text

inghor

Typewritten Text

inghor

Typewritten Text

inghor

Typewritten Text

This classification is inspired by Jacobs et al. (2) but modified since the relatively small anatomical structures in the cervical spine complicate assessments of nucleus signal. Reduced height is assessed by comparison with neighbouring discs that look morphologically normal (3). In case of extensive disc degeneration thus complicating between-disc comparison, the assessment is based on the experience of the reader.

Both of these have normal disc height = grade 0. From Jacobs et al. (2).

__________________________________________________________

Reduced height = grade 1

Reduced height = grade 1

Disc contour 0 : normal 1 : bulge or protrusion 2 : extrusion 9 : missing Disc contour is assessed by use of both sagittal and axial images. We use a modified version of the classification by Fardon et al. 2014 (4).

A normal disc is defined as morphologically normal. I.e., there is no sign of illness, trauma or ageing. The disc is located within the "disc space", the limits of which are determined by the vertebral endplates craniocaudally. The peripheral limits of "disc space" are constituted by the apophyses of the vertebrae (excluding possible osteophytes) (4).

Disc bulge describes a general (in some cases asymmetric) bulging of disc material. The material has moved beyond "disc space" and comprises > 25% of the disc circumference. Usually, a bulge does not exceed 3 mm beyond "disc space" (4).

Protrusion: A focal displacement of disc material of which the base involves < 25% of the circumference. The maximum measure of the displaced disc material is smaller than the measure of the base of the displaced material at the disc space of origin, measured in the same plane (4).

Extrusion: A focal displacement of disc material where the base involves < 25% of the disc circumference. The maximal measure of the displaced disc material is greater than the measure of the base of the displaced material at the disc space of origin, measured in the same plane. Extrusion also comprises what was formerly denoted as sequestration where the displaced material has lost continuity with the disc space of origin (4).

Normal discs on axial and sagittal images

Normal discs (4)

______________________________

Bulge

Bulge (4)

___________________________________

_______________________________________________

Protrusion, axial

Protrusion (4)

Protrusion, sagittal

Extrusion, sagittal

Extrusion (4)

Extrusion, axial

_______________________________________________________________ Spinal canal stenosis 0 : no stenosis 1 : >50% obliteration of CSF, no cord deformity 2 : >50% obliteration of CSF with cord deformity but no signal change 3 : >50% obliteration of CSF with cord deformity and signal change 9 : missing The spinal canal is assessed on three T2-weighted images; the mid-sagittal and the two neighbouring slices (one on each side of the mid-sagittal image). The classification by Kang et al. is used (5).

________________________________________________________________

Grade 0-3 (5):

Vertebral endplate signal change (VESC) 0 : normal

1 : VESC type 1

2 : VESC type 2

3 : VESC type 3

4 : mixed VESC type 1 and 2

9 : missing Primarily assessed on sagittal images. VESC constitute changes in the endplates of two neighbouring vertebral endplates surrounding a degenerated disc (4). If only visible on one slice, the finding is excluded (6). Likewise, changes only related to osteophytes or Schmorl nodes will be excluded. Type 1 is hyperintense on T2-weighted and hypo-/isointense on T1-weighted images (7). Type 2 is hyperintense on T1-weighted images and hyper-/isointense on T2-weighted images (7). Type 3 er hypointense on T1 og T2 (bone sclerosis where hydrogen protons are so tightly bound and thus difficult to bring in motion no MR signal on T1 and T2) (4).

VESC type 1. T1-weighted




Drawing of type 1, 2 og 3 (4)

____________________________________________________________________ Neural foraminal stenosis 0 : normal or < 50% obliteration of perineural fat 1 : >= 50% obliteration of perineural fat with/without nerve root compromise. 9 : missing The extent of perineural fat obliteration is determined on T2 weighted oblique images. The assessment is made on the slice with most severe stenosis. Meanwhile, on the axial images, the reader ensures that he/she is actually looking at the neural foramen. The boundaries of the neural foramen are: to the anterior the disc and posterior aspect of the vertebral bodies; to the posterior the zygapophyseal joint; and cranially and caudally the pediculi (8). On axial images, the medial and lateral boundaries are the medial and lateral aspect of the pediculi, respectively (9). If no oblique images are available, the assessment is based on the axial images where stenosis is most severe, i.e. where the distance between the processus articularis superior (posterolaterally) and corpus vertebra/discus (anteromedially) is identified (10).

Grade 1

In the following, foraminal stenosis on axial images is depicted. While assessing the axial images, the sagittal images are used to demonstrate the location in relation to the processes articulares. The severity of stenosis is determined by comparing the relevant number of subsequent axial slices in the neural foramen in question.

Grade 0 (15):

0 = figure a +b

Grade 1(15):

1 = figure c+d

____________________________________________________________________ Uncovertebral osteoarthritis 0 : normal 1 : definite osteoarthritis (primarily assessed by the presence of osteophytes which take up place in the neural foramen). 9 : missing Only if >= 50% foraminal stenosis is settled on, the reader can consider the classification definite osteoarthritis. (i.e. if foraminal stenosis is <50%, any possible irregularities of the uncinate process are not articulate enough to be classified as osteoarthritis). The classification was chosen because degenerative changes of the uncovertebral joint primarily comprise the growth of osteophytes (11). Assessment is only done on oblique images which allow for the best assessments (11,12). If these are not available, the value '9' = missing is used due to the risk of partial volume effect on axial images.

Grade 0

____________________________________________________________________ Zygapophyseal osteoarthritis 0 : normal (no definite joint space narrowing, osteophytes or hypertrophy of the processus articularis) 1 : definite osteoarthritis (definite joint space narrowing, osteophytes or hypertrophy of the processus articularis) (13,14) 9 : missing The assessment is done on oblique images (12). If these are not availabe, axial and sagittal images are used. Only if >= 50% foraminal stenosis is settled on, the reader can consider the classification definite osteoarthritis. (i.e. if foraminal stenosis is <50%, any possible irregularities of the zygapophyseal joint are not articulate enough to be classified as osteoarthritis.

Grade 0, normal uncovertebral joint

Grade 1, definite uncovertebral osteoarthritis

____________________________________________________________________ Comments If any other relevant pathological findings are identified. This may be spondylolistesis, fracture, cord pathology etc. ____________________________________________________________________

Grade 0, normal zygapophyseal joint

Grade 1, definite zygapopyseal osteoarthritis

(1) Nouri A, Martin AR, Mikulis D, Fehlings MG. Magnetic resonance imaging assessment of degenerative cervical myelopathy: a review of structural changes and measurement techniques. Neurosurg Focus 2016 Jun;40(6):E5.

(2) Jacobs LJ, Chen AF, Kang JD, Lee JY. Reliable Magnetic Resonance Imaging Based Grading System for Cervical Intervertebral Disc Degeneration. Asian Spine J 2016 Feb;10(1):70-74.

(3) Fu MC, Webb ML, Buerba RA, Neway WE, Brown JE, Trivedi M, et al. Comparison of agreement of cervical spine degenerative pathology findings in magnetic resonance imaging studies. Spine J 2016 Jan 1;16(1):42-48.

(4) Fardon DF, Williams AL, Dohring EJ, Murtagh FR, Gabriel Rothman SL, Sze GK. Lumbar disc nomenclature: version 2.0: Recommendations of the combined task forces of the North American Spine Society, the American Society of Spine Radiology and the American Society of Neuroradiology. Spine J 2014 Nov 1;14(11):2525-2545.

(5) Kang Y, Lee JW, Koh YH, Hur S, Kim SJ, Chai JW, et al. New MRI grading system for the cervical canal stenosis. AJR Am J Roentgenol 2011 Jul;197(1):W134-40.

(6) Maatta JH, Karppinen J, Paananen M, Bow C, Luk KD, Cheung KM, et al. Refined Phenotyping of Modic Changes: Imaging Biomarkers of Prolonged Severe Low Back Pain and Disability. Medicine (Baltimore) 2016 May;95(22):e3495.

(7) Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology 1988 Jan;166(1 Pt 1):193-199.

(8) Bojsen-Moeller F. Chapter 8: Hvirvelsoejlen (The Spine). Bevaegeapparatets Anatomi. 12th ed. Copenhagen: Munksgaard Danmark; 2001. p. 89.

(9) Wiltse LL, Berger PE, McCulloch JA. A system for reporting the size and location of lesions in the spine. Spine (Phila Pa 1976) 1997 Jul 1;22(13):1534-1537.

(10) Kim S, Lee JW, Chai JW, Yoo HJ, Kang Y, Seo J, et al. A New MRI Grading System for Cervical Foraminal Stenosis Based on Axial T2-Weighted Images. Korean J Radiol 2015 Nov-Dec;16(6):1294-1302.

(11) Yochum TR, Rowe LJ. Chapter 10: Arthritic Disorders. Essentials of Skeletal Radiology; 1996. p. 795-807.

(12) Shim JH, Park CK, Lee JH, Choi JW, Lee DC, Kim DH, et al. A comparison of angled sagittal MRI and conventional MRI in the diagnosis of herniated disc and stenosis in the cervical foramen. Eur Spine J 2009 Aug;18(8):1109-1116.

(13) Kalichman L, Suri P, Guermazi A, Li L, Hunter DJ. Facet orientation and tropism: associations with facet joint osteoarthritis and degeneratives. Spine (Phila Pa 1976) 2009 Jul 15;34(16):E579-85.

(14) Xu C, Ding ZH, Xu YK. Comparison of computed tomography and magnetic resonance imaging in the evaluation of facet tropism and facet arthrosis in degenerative cervical spondylolisthesis. Genet Mol Res 2014 May 30;13(2):4102-4109.

(15) Park HJ, Kim SS, Lee SY, Park NH, Chung EC, Rho MH, et al. A practical MRI grading system for cervical foraminal stenosis based on oblique sagittal images. Br J Radiol 2013 May;86(1025):20120515.

inghor

Typewritten Text

inghor

Typewritten Text

Table5. PrevalenceofpositiveMRI findings ReaderA,1st

assessment ReaderB ReaderCReaderA,2ndassessment

MRIfindingNumberofMRIsassessed

Positivefindingsn(%)

NumberofMRIsassessed






Kyphosis 50 15(30.0) 50 17(34.0) 50 20(40.0) 50 13(26.0)

DischeightC2/C3 50 3(6.0) 50 0(0.0) 50 1(2.0) 50 3(6.0)

DischeightC3/C4 50 8(16.0) 50 4(8.0) 50 8(16.0) 50 10(20.0)

DischeightC4/C5 50 12(24.0) 50 15(30.0) 50 21(42.0) 50 17(34.0)

DischeightC5/C6 50 26(52.0) 50 24(48.0) 50 32(64.0) 50 36(72.0)

DischeightC6/C7 50 21(42.0) 50 16(32.0) 50 22(44.0) 50 28(56.0)

DischeightC7/T1 50 1(2.0) 50 1(2.0) 50 2(4.0) 50 3(6.0)

DisccontourC2/C3 30 1(3.3) 50 12.00 50 1(2.0) 31 1(3.2)

DisccontourC3/C4 38 11(28.9) 50 9(18.0) 50 8(16.0) 35 9(25.7)

DisccontourC4/C5 45 15(33.3) 50 13(26.0) 50 15(30.0) 45 16(35.6)

DisccontourC5/C6 48 31(64.6) 50 26(52.0) 50 32(64.0) 48 35(72.9)

DisccontourC6/C7 46 26(56.5) 50 19(38.0) 50 24(48.0) 46 26(56.5)

DisccontourC7/T1 38 2(5.2) 50 3(6.0) 50 1(2.0) 34 2(5.9)

inghor

Typewritten Text

inghor

Typewritten Text

inghor

Typewritten Text

Appendix 5

inghor

Typewritten Text

SpinalcanalstenosisC2/C3 50 0(0.0) 50 0(0.0) 49 1(2.0) 50 0(0.0)





SpinalcanalstenosisC7/T1 50 0(0.0) 50 0(0.0) 50 0(0.0) 50 0(0.0)

VESCC2/C3 50 0(0.0) 50 1(2.0) 50 0(0.0) 50 0(0.0)

VESCC3/C4 50 1(2.0) 50 1(2.0) 50 1(2.0) 50 1(2.0)

VESCC4/C5 50 0(0.0) 50 1(2.0) 50 1(2.0) 50 0(0.0)

VESCC5/C6 50 2(4.0) 50 2(4.0) 50 3(6.0) 50 3(6.0)

VESCC6/C7 50 3(6.0) 50 4(8.0) 50 5(10.0) 50 5(10.0)

VESCC7/T1 50 0(0.0) 50 0(0.0) 50 0(0.0) 50 0(0.0)

VESC:vertebralendplatesignalchange(Modicchange)

Table5. PrevalenceofpositiveMRIfindings continued ReaderA,1st

assessment ReaderB ReaderCReaderA,2ndassessment

MRIfindingNumberofMRIsassessed








RightuncovertebralosteoarthritisC2/C3 44 1(2.3) 43 0(0.0) 48 0(0.0) 43 0(0.0)RightuncovertebralosteoarthritisC3/C4 47 1(2.1) 44 0(0.0) 48 3(6.3) 46 2(4.3)RightuncovertebralosteoarthritisC4/C5 47 9(19.1) 46 5(10.9) 49 6(12.2) 46 5(10.9)RightuncovertebralosteoarthritisC5/C6 47 12(25.5) 46 8(17.4) 49 12(24.5) 44 10(22.7)RightuncovertebralosteoarthritisC6/C7 48 11(22.9) 46 10(21.7) 49 7(14.3) 46 14(30.4)RightuncovertebralosteoarthritisC7/T1 45 0(0.0) 46 0(0.0) 49 0(0.0) 45 0(0.0)LeftuncovertebralosteoarthritisC2/C3 44 0(0.0) 44 0(0.0) 48 0(0.0) 43 0(0.0)LeftuncovertebralosteoarthritisC3/C4 47 3 (6.4) 44 0(0.0) 48 2(4.2) 46 4(8.7)LeftuncovertebralosteoarthritisC4/C5 48 9(18.8) 46 4(8.7) 49 3(6.1) 46 7(15.2)LeftuncovertebralosteoarthritisC5/C6 47 11(23.4) 46 9(19.6) 49 9(18.4) 45 12(26.7)LeftuncovertebralosteoarthritisC6/C7 48 8(16.7) 46 5(10.9) 49 10(20.4) 46 10(21.7)LeftuncovertebralosteoarthritisC7/T1 45 0(0.0) 46 1(2.2) 49 0(0.0) 43 0(0.0)Rightzygapophysealosteoarthritis C2/C3 46 1(2.2) 43 1(2.3) 49 3(6.1) 44 1(2.2)Rightzygapophysealosteoarthritis C3/C4 48 0(0.0) 44 3(6.8) 49 1(2.0) 46 2(4.3)RightzygapophysealosteoarthritisC4/C5 48 6(12.5) 46 6(13.0) 49 2(4.1) 47 3(6.4)Rightzygapophysealosteoarthritis C5/C6 48 7(14.6) 46 6(13.0) 49 5(10.2) 47 11(23.4)Rightzygapophysealosteoarthritis C6/C7 48 10(20.8) 46 7(15.2) 49 5(10.2) 47 12(25.5)Rightzygapophysealosteoarthritis C7/T1 47 0(0.0) 46 0(0.0) 49 0(0.0) 45 0(0.0)LeftzygapophysealosteoarthritisC2/C3 46 0(0.0) 44 0(0.0) 49 0(0.0) 43 0(0.0)LeftzygapophysealosteoarthritisC3/C4 48 0(0.0) 44 0(0.0) 49 1(2.0) 46 3(6.5)LeftzygapophysealosteoarthritisC4/C5 48 5(10.4) 46 6(13.0) 49 4(8.2) 48 8(16.7)Leftzygapophyseal osteoarthritisC5/C6 48 9(18.8) 46 7(15.2) 49 6(12.2) 47 12(25.5)LeftzygapophysealosteoarthritisC6/C7 48 8(16.7) 46 5(10.9) 49 3(6.1) 47 10(21.3)LeftzygapophysealosteoarthritisC7/T1 47 0(0.0) 46 0(0.0) 49 0(0.0) 45 0(0.0)RightneuralforaminalstenosisC2/C3 42 1(2.4) 42 0(0.0) 46 3(6.5) 43 1(2.3)RightneuralforaminalstenosisC3/C4 46 1(2.2) 44 1(2.3) 48 3(6.3) 46 2(4.3)

RightneuralforaminalstenosisC4/C5 48 10(20.8) 45 8(17.8) 49 8(16.3) 48 6(12.5)RightneuralforaminalstenosisC5/C6 48 13(27.1) 46 10(21.7) 49 14(28.6) 48 15(31.3)RightneuralforaminalstenosisC6/C7 48 15(31.3) 46 11(23.9) 49 11(22.4) 48 15(31.3)RightneuralforaminalstenosisC7/T1 45 0(0.0) 46 0(0.0) 48 2(4.2) 45 0(0.0)LeftneuralforaminalstenosisC2/C3 42 1(2.4) 44 0(0.0) 47 0(0.0) 43 0(0.0)LeftneuralforaminalstenosisC3/C4 46 3(6.5) 44 0(0.0) 49 4(8.2) 46 4(8.7)LeftneuralforaminalstenosisC4/C5 48 9(18.8) 46 6(13.0) 49 6(12.2) 48 10(20.8)LeftneuralforaminalstenosisC5/C6 48 13(27.1) 46 13(28.3) 49 14(28.6) 48 16(33.3)LeftneuralforaminalstenosisC6/C7 47 9(19.1) 46 8(17.4) 49 12(24.5) 48 12(25.0)LeftneuralforaminalstenosisC7/T1 45 0(0.0) 46 1(2.2) 48 0(0.0) 45 0(0.0)

inghor

Typewritten Text

inghor

Typewritten Text

Answer to request 86 / 2017

Dear Line

You have asked the Regional Ethical Committee of Central Denmark Region whether the described project needs committee approval.After specific assessment of your study, it is not considered a health research project but rather a methodological study (following the definitions in the law of the committee §2, no. 1). Thus, the study does not require approval from the committee. Please be aware that you might need approval from the Data Protection Agency.

The law in mention is law no. 593 from June 14th 2011 about scientific ethical assessment of health research.

Kind regards,

Helle NikkelSecretaryScientific Ethical Committees, Central Denmark Region

inghor

Typewritten Text

Appendix 6

inghor

Typewritten Text

TableA.OverviewofRCTsaimedatreturntowork.WorkplacesinvolvedSickleaveduration(months)

0 0.5 1 1.5 2 3 4 5 6 7 8 9 10 11 12>12

Loisel1997,thoracicpainorLBP(125)

Anema2007,LBP(115)

Bültmann2009,LBP,neckpainandpaininlowerlimb(126)

Lambeek2010,LBP (116)

Vermeulen2011*,

MSK(142)

Myhre2014,neckandbackpain (128)Linton2016,LBP(117)

*Workoutcomespoorer<90daysbutbetter>90 daysGreen:WorkoutcomesstatisticallysignificantlyimprovedRed:NodifferenceinworkoutcomesbetweennewinterventionandreferenceinterventionLBP:LowBackPain;MSK:Musculoskeletalpain;MHD:MentalHealthDisorders

inghor

Typewritten Text

Appendix 7

inghor

Typewritten Text

TableB.OverviewofRCTsaimedatreturntowork.Workplacesnot/scarcelyinvolved

Sickleaveduration(months)

0 0.5 1 1.5 2 3 4 5 6 7 8 9 10 11 12 >12

Indahl1995,

LBP**(38)

Indahl1998,LBP***(39)

Haldorsen1998,MSK(118)

Hagen2000,LBP(40)

Marhold2001,MSK(121) Marhold2001,MSK(121)

Hagen

2010,LBP(143)

Jensen2011,LBP(114) Jensen2012,LBP(130)

Andersen2015,paininbackorupperbody****

(123)

Andersen2015,paininbackorupperbody*****

(124)

Reme2016,LBP (122)

Moll2017,neckorshoulderpain(97)

Lytsy2017,

MSK,MHD(120)

Aasdahl2017,LBP,MHD,other(119) Brendbekken2017,MSK(127) **Followup1year***Followup5years****Followup3months*****Followup11monthsGreen:WorkoutcomesstatisticallysignificantlyimprovedRed:NodifferenceinworkoutcomesbetweennewinterventionandreferenceinterventionLBP:LowBackPain;MSK:Musculoskeletalpain;MHD:MentalHealthDisorders

inghor

Typewritten Text

Sick assessment and prognosis - MarselisborgCentret

Documents