the strengths, weaknesses, opportunities, and threats
TRANSCRIPT
DOI: 10.4018/IJBDAH.2019010101
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
Copyright©2019,IGIGlobal.CopyingordistributinginprintorelectronicformswithoutwrittenpermissionofIGIGlobalisprohibited.
1
The Strengths, Weaknesses, Opportunities, and Threats Analysis of Big Data Analytics in HealthcareChaojie Wang, The MITRE Corporation, McLean, USA
https://orcid.org/0000-0001-8521-9420
ABSTRACT
ImprovingtheperformanceandreducingthecostofhealthcarehavebeenagreatconcernandahugechallengeforhealthcareorganizationsandgovernmentsateverylevelintheUS.Measurestakenhaveincludedlaws,regulations,policies,andinitiativesthataimtoimprovequalityofcare,reducecostsofcare,andincreaseaccesstocare.CentraltothesemeasuresisthemeaningfulandeffectiveuseofBigDataanalytics.Toreap thebenefitsofbigdataanalyticsandalignexpectationswithresults,researchers,practitioners,andpolicymakersmusthaveaclearunderstandingoftheuniquecircumstancesofhealthcareincludingthestrengths,weaknesses,opportunities,andthreats(SWOT)associatedwiththeuseofthisemergingtechnology.ThroughdescriptiveSWOTanalysis,thisarticlehelpshealthcarestakeholdersgainawarenessofbothsuccessfactorsandissues,pitfalls,andbarriersintheadoptionofbigdataanalyticsinhealthcare.
KeyWORDSBig Data, Data Analytics, Data Breaches, Data Ethics, Electronic Health Records, Health Information Exchange, Health IT, Healthcare, Machine Learning, SWOT Analysis
1. INTRODUCTION
TheUShealthcaresystemhasbothstrengthsandweaknesses.Itenjoysalarge-scale,well-trained,andhigh-qualityworkforceofclinicians,nurses,andspecialists,robustmedicalresearchprograms,andtheworld’sbestclinicaloutcomesinselectmedicalservices.Yet,itsuffersfromhighexpenditure,lowperformance,anddisparityinhealthstatus,accesstocare,andoutcomesofcare(Barnes,Unruh,Rosenau,&Rice,2018).
1.1. High Cost of the US Healthcare SystemAccording to a recent report published by The Organization for Economic Co-operation andDevelopment(2018),in2017theUSspendingonhealthcarewasthelargest,measuredbyboththespendingpercapitaandthepercentageofthegrossdomesticproduct(GDP)amongits37membernations.Figure1showsthattheUSspentover$10,000percapitaonhealthcarethatyear,orabout17%ofGDP.
EvenmorealarmingistherapidgrowthinUShealthcarespending.AccordingtotheCentersforMedicareandMedicaidServices(CMS),healthcarespendingisprojectedtogrowatanaveragerateof5.8percentfrom2012-2022,1.0percentagepointfasterthantheexpectedaverageannualgrowthintheGDP.By2022,UShealthcarespendingisprojectedtobenearly20%ofGDP(CentersforMedicareandMedicaidServices,2012).
Thisarticle,originallypublishedunderIGIGlobal’scopyrightonMay24,2019willproceedwithpublicationasanOpenAccessarticlestartingonJanuary20,2021inthegoldOpenAccessjournal,InternationalJournalofBigDataandAnalyticsinHealthcare(convertedtogoldOpenAccessJanuary1,2021),andwillbedistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://cre-
ativecommons.org/licenses/by/4.0/)whichpermitsunrestricteduse,distribution,andproductioninanymedium,providedtheauthoroftheoriginalworkandoriginalpublicationsourceareproperlycredited.
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
2
1.2. Low Performance of the US Healthcare SystemThisextremelyhighspendingissharplycontrastedwiththelowperformanceintheUShealthcaresystem.In2000,theWorldHealthOrganization(WHO)publishedareportthatmeasuredandrankedthehealthsystemperformanceof191countries.Accordingtothisreport,theUShealthcaresystemwasunimpressivelyrankedat37,belowmost industrializedcountries includingFrance, theUK,andCanada,andevenbelowsomelessdevelopedcountriessuchasColombiaandChile(Tandon,Murray,Lauer,&Evans,2000).Almosttwodecadeslater,therehasnotbeenmuchimprovementintheperformanceoftheUShealthcaresystem.Accordingtoa2017reportfromtheCommonwealthFund,theUSisrankedlastoutof11high-incomeindustrializedcountriesbasedonmeasuresincludingcareprocess,accesstocare,administrativeefficiency,equity,andhealthoutcomes(Schneider,Sarnak,Squires,Shah,&Doty,2017).Figure2showsthattheUShealthcaresystemperformsatthebottomonfourofthefivemeasures.
1.3. efforts to Improve the US Healthcare SystemIn2007,theInstituteforHealthcareImprovement(IHI)launchedtheTripleAiminitiativetoimprovethepatientexperienceofcare(includingqualityandsatisfaction),improvethehealthofpopulations,andreducethepercapitacostofhealthcare.TheTripleAiminitiativedirectlytargetsthecriticalmeasuresofhealthcareperformanceincludingbothqualityofcareandefficiencyofcare(InstituteforHealthcareImprovement,2007).
In2008,CongressenactedtheMedicareImprovementsforPatientsandProvidersAct(MIPPA).AspartoftheimplementationofMIPPA,CMSintroducedthevalue-basedpurchasing(VBP)planlinkingpaymentsdirectlytothequalityoutcomesofthecareprovided.Thispay-for-qualityorpay-for-performanceplanaimstomoveawayfromthetraditionalfee-for-serviceplanwhichprovidesnoincentivesfortheproviderstoimprovecarequalityandcontributestothehighcostofhealthcare(Terhaar,2018).
In2010,toaddressthedisparityandinequityinhealthcareandtoincreaseaccesstocarefortensofmillionsofuninsuredandunderinsuredAmericans,CongressenactedthePatientProtectionandAffordableCareActof2010,alsoknownas“Obamacare”(“PatientProtectionandAffordableCareActof2010,”2010).
Whilehealthcareinitiatives,regulations,andpoliciesmayhelpdrivethequalityandperformanceimprovementatthemacrolevel,effectiveimplementationsrequireconcertedeffortsatthemicrolevelbyallstakeholdersincludingpolicymakers,providers,payer,patients,andthepublic.Inaddition,thesediversestakeholdersmustbeempoweredandenabledbyinnovativesolutionsandtechnologies
Figure 1. 2017 health spending per capita as share of GDP (Organization for Economic Co-operation and Development, 2018, p. 2)
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
3
todealwiththecomplexproblemsinhealthcare.Bigdataanalyticsasanemergingtechnologyisoneofmanyindispensabletoolsinthetoolboxandcanplayanimportantroleinimprovinghealthcare.However,aswithanynewtechnology,benefitsoftencomewithlimitations,opportunitieswithrisks,andhopeswithhypes.Itisimportanttounderstandthefullspectrumofthisnewtechnologysothatitcanbeefficientlyandeffectivelyadoptedandapplied.
MotivatedbythetremendouschallengesfacingtheUShealthcaresystemandthegreatpotentialbigdataanalyticshasinimprovingitsperformance,thispaperpresentsahigh-levelanalysisofthestrengths,weaknesses,opportunities,andthreats(SWOT)associatedwiththeuseofbigdataanalyticsinhealthcare.Thegoalofthispaperistohelppolicymakers,careproviders,researchers,andthepublicgainbroaderanddeeperunderstandingofthemanydimensionsoftheuseofthisemergingtechnologyinhealthcaresothattheycanmakeinformedandsensibledecisionsintheireffortstoimprovehealthcarequalityandperformance.
2. BIG DATA ANALyTICS AS eNABLeR TO IMPROVe HeALTHCARe
2.1. Big Data AnalyticsDuringthepastdecade,bigdataemergedasanewtechnologytrendthankstotherapidinnovationandadvancementininformationandcommunicationtechnology(ICT).BigdatawasinitiallydefinedwiththreeessentialcharacteristicsknownastheThreeV’s–Volume,Velocity,andVariety.Sincethen,moreV’shavebeenadded.ASixV’smodelwidelyusedinhealthcareaddsthreeadditionalV’s–Veracity,Variability, and Value (Senthilkumar, Rai, Meshram, Gunasekaran, & Chandrakumarmangalam,2018).Dataanalyticsisanumbrellatermcommonlyusedtoreferencebusinessintelligence,businessanalytics,datamining,knowledgediscoveryindatabases,anddatascience.Aslargevolumesofdatainwidevarietiesarecollectedandmadeavailable,thedemandincreasesforextractingvaluesfrombigdatatoimprovebusinessperformanceanddriveorganizationalchanges.Dataanalyticsanswersthechallengebyintegratingcomputerscience,informationtechnology,statistics,domainknowledge,andhumancollaborationinastreamlinedprocessofknowledgediscovery,creation,andapplication(Wang,2018).
2.2. Applications of Big Data Analytics in HealthcareHealthcaredata analytics is the applicationofbigdata anddata analytics tohealthcare.Similartermswithminordifferencesalsoappearbothinacademiaandindustry.Amongthemarehealthcareanalytics, health analytics, healthcarebigdata analytics, bigdata analytics in health, andhealth
Figure 2. 2017 healthcare systems performance ranking (Schneider et al., 2017)
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
4
informatics.Knowledgecanbediscovered,andinsightscanbegainedfrombigdataanalyticstoimprovehealthcarequalityandperformance.Basedonthecontentanalysisof26casestudies,Wang,Kung,andByrd(2016)identifiedfivepotentialbenefitsofhealthcaredataanalytics:ITinfrastructure,operational, organizational, managerial, and strategic benefits. Lebied (2018) provided detailedaccountsoftwelveapplicationsofbigdataanalyticsinhealthcaresuchaspreventionofopioidabuse,telemedicine,preventionofERvisits,andfrauddetection.Theseserveasexemplarsofthemanyapplicationscurrentlyrecognizedandmanymoreinnovativeapplicationscontinuingtoemerge.
TheUniversityofPennsylvaniaHealthSystem(UPHS)’suseofmachinelearningandElectronicHealthRecords(EHR)datatopredicttheonsetofseveresepsisisanexampleofusingbigdataanalyticsinhealthcare.Sepsisisalife-threateningillnessduetobloodinfection.AccordingtotheCenterforDiseaseControl&Prevention(CDC),eachyear,1.7millionAmericanadultsdevelopsepsis,270,000Americansdieofsepsis,andoneinthreepatientswhodieinhospitalshassepsis(Dantes&Epstein,2018).SepsiscostsAmericanhospitals$27billionannually,makingitthenumberonecostdriverforhospitals(Reinhart,2018).UPHSdevelopedapredictivemodelbytrainingamachinelearningalgorithmusinghistoricpatientdatastoredin itsEHRincludinglabs,clinical,anddemographicdata.Thetrainedmachinelearningmodelwasthenusedtopredictseveresepsis12hourspriortotheclinicalonsetbycontinuouslypullingreal-timepatientdatafromtheEHR.Alertsweresenttocliniciansandnursestoenableadditionalmonitoringandearlyintervention(Gianninietal.,2017).
2.3. The Adoption Model for Analytics Maturity (AMAM)Whenitcomestotheadoptionandutilizationofbigdataanalytics,eachorganizationisconstrainedbyitsuniqueorganizationalcharacteristicsincludingsize,financialresources,technicalcapability,andleadershipandenvironmentalcharacteristicssuchasgeographicallocation,community,andpatientpopulations.Tohelpassessthematurityofhealthcareorganizationsinadoptingbigdataanalytics,HIMSSAnalytics,awhollyownedsubsidiaryoftheHealthInformationandManagementSystemSociety(HIMSS),developedtheAdoptionModelforAnalyticsMaturity(AMAM)(HIMSSAnalytics,n.d.). AMAM describes eight stages representing increasing levels of maturity. An organizationtypicallystartsoutatstagezeroinwhichitonlyusesanalyticsinsparseandsiloedpointsolutionsandgraduallyclimbsuptostageseven,whereanalyticsisusedforprescriptiveandpersonalizedcareforindividualpatients.
3. THe SWOT FRAMeWORK AND SUMMARy OF ANALySIS
3.1. The SWOT FrameworkStrengths-Weaknesses-Opportunities-Threats(SWOT)isamanagementscienceframeworkoriginallyusedforcorporatestrategicplanningdatingbacktothe1960s.Itisananalysistooltoassessabusiness’sfitnessasmeasuredbyhowwellitsinternalqualities(thestrengthsandweaknesses)matchupwithexternalfactors(theopportunitiesandthreats)(Hill&Westbrook,1997).
Strengthsandopportunitiesareconsideredpositive,favorable,andsynonymoustosuccessfactorswhile weaknesses and threats are considered negative, unfavorable, and synonymous to pitfalls,challengesorbarriers.Theobjectiveofstrategicplanningistomitigateorminimizetheunfavorablefactorsandutilizeormaximizethefavorablefactors.TheSWOTframeworkhasbeenappliedtotheuseofbigdataanalyticsingeneral(Wang,Wang,&Alexander,2015)andinaspecificindustryordomain(Collins,2016).ThispaperappliesSWOTanalysistotheuseofbigdataanalyticsinhealthcare.Figure3showstheframeworkinfourquadrants,knownasaSWOTmatrix.
3.2. Strengths and Weaknesses of the SWOT FrameworkSWOTisasimpleyetpowerfulmodelforanalyzingacomplexsituation.Itreflectsthecomplexityoftherealityandpresentsadialecticallybalancedperspectiveforanalyzinganddealingwithcomplex
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
5
problems.However,SWOTanalysisisnotconsideredarigorousempiricalresearchmethod.Itisdifferentfromthequalitativemethodaimingatthegenerationofnewtheoriesorthequantitativemethodaimingattheconfirmationorfalsificationofexistingorproposedtheories.SWOTprovidesawayof thinkingakin tosystems thinkingorcritical thinking.Thedefinitionandclassificationofwhatconstitutesstrengths,weaknesses,opportunities,andthreatsaresubjectiveandreflecttheresearchersandpractitioners’personalexperienceandunderstandingofthecomplexsocioeconomic,cultural,andmanagerialproblems.SWOT’spowerliesinitssimplicityindealingwithcomplexity.Itprovidesasimpleframeofreferenceforanalyzinganddealingwithcomplexsituationssothatpotential opportunities can be explored, potential risks can be assessed, and potential outcomes(intendedorunintended,desirableorundesirable)canbeilluminatedbeforeinterventionscanbedevelopedandactionscanbetaken.
3.3. SWOT Analysis of Big Data Analytics in HealthcareThis SWOT analysis was performed based on review of literature, industry reports, and expertopinions.Theauthoralsodrawsuponhisprofessionalexperiencesinsystemsengineering,healthIT,dataanalytics,andhealthcarequalitymanagement.AligningwiththefourquadrantsoftheSWOTmatrix,thispapersetsouttoanswerthefollowingfourquestions:
1. Whatarethestrengthsofhealthcarethatmakeitfavorablefortheadoptionofbigdataanalytics?
Figure 3. The SWOT matrix
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
6
2. Whataretheweaknessesofhealthcarethatmayhindertheadoptionofbigdataanalytics?3. Whataretheopportunitiesthatfavortheadoptionofbigdataanalyticsinhealthcare?4. Whatarethethreatsassociatedwiththeadoptionofbigdataanalyticsinhealthcare?
Theextantliteratureonsimilarsubjectstendstobeatalowerlevelorwithtechnicalimplementationdetails.Collins(2016)performedaSWOTanalysisofbigdataandhealtheconomicsintheUKwithafocusoncostsandincludedcoverageofdrugs,biomonitoring,anddatarepositories.Yangetal.(2016)performedaSWOTanalysisofwearabledevicesinhealthcare.WhiletheflexibilityoftheSWOTframeworkallowsfordifferentlevelsofanalysis,thispaperfocusesonthebigpicture(theforest)insteadoflow-leveldetails(thetrees).TheresultsoftheSWOTanalysisaresummarizedinFigure4.Detaileddescriptionsareprovidedinthefollowingfoursections.
4. STReNGTHS
4.1. Tradition of evidence-Based MedicineWesternmedicinehasthetraditionofapplyingscienceandtechnologyintheprevention,control,diagnosisandtreatmentofdiseases.FromthedevelopmentofmedicaldevicessuchasMagneticResonanceImaging(MRI)andtheAutomatedExternalDefibrillator(AED)totheuseofexperimentaldesignandstatisticalinferenceinclinicaltrials,scienceandtechnologyremainattheheartofmodernmedicine.
A spirited movement called Evidence-Based Medicine (EBM) began within the healthcarecommunityintheearly1990s.EBMisdefinedas“theconscientious,explicit,andjudicioususeofcurrentbestevidenceinmakingdecisionsaboutthecareofindividualpatients.Thepracticeofevidence-basedmedicinemeansintegratingindividualclinicalexpertisewiththebestavailableexternalclinicalevidencefromsystematicresearch”(Sackett,Rosenberg,Gray,Haynes,&Richardson,1996).EBMhasbecomeanessentialcodeintheDNAofhealthcareeversince.Itsprinciples,processes,andpracticesarewellalignedwiththoseofbigdataandanalytics.Theincreasingvolume,variety,andavailabilityofdataalongwiththeadvancementofdatastorage,computingpower,andanalyticstoolsandtechniqueswillmakeEBMmorepracticalandpowerful.
Figure 4. Summary of the SWOT analysis on big data analytics in healthcare
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
7
4.2. Prevalence of electronic Health RecordsElectronic Health Records (EHR) evolved from traditional Electronic Medical Records (EMR).WhileEMRwaslimitedtomedicalrecords,EHRexpandsthescopetoallrecordsgermanetohealthincludingmedical, administrative, claim, and socioeconomic records. In2009, theUSCongressenactedtheHealthInformationTechnologyforEconomicandClinicalHealth(HITECH)ActaspartoftheAmericanRecoveryandReinvestmentAct(ARRA)afterthe2008globalfinancialcrisis.HITECHwascreatedtopromotetheadoptionofhealthITandimplementationofEHRtomodernizetechnologyinfrastructureandcapabilitiesofhealthcare.FederalinvestmentwasprovidedtohealthcareproviderstoincentivizehealthITadoptionandafederalcertificationprocesswasputinplacetoensureITsystemsaredevelopedaccordingtotechnological,functional,andsecurityrequirements.
SincetheenactmentoftheHITECHAct,morehealthcareorganizationshaveadoptedhealthITandimplementedEHRs.AccordingtotheOfficeoftheNationalCoordinatorforHealthInformationTechnology(2018),“In2017,96percentofallnon-federalacutecarehospitalspossessedcertifiedhealthIT.”ThepenetrationofhealthITandEHRsprovidesthetechnicalfoundationanddatasourcesfortheadoptionofbigdataanalyticstoimprovehealthcare.Atthesametime,bigdataanalyticsaimstodelivertherightamountofinformationtocareprovidersattherighttimeintherightplaceandhelprealizetheintendedbenefitsofEHRsandalleviatetheunintendedburdenassociatedwithEHRssuchasadministrativeoverheadandinformationoverload.
4.3. Widespread Use of Mobile TechnologyTheubiquitousInternetcoupledwithGlobalPositioningSystems(GPS),mobiletelecommunicationnetworks,andcloudcomputingleadtotheproliferationofmobiledevices(includingwearablefitnesstrackers)andmobileappsthatarebecominganintegralpartofhealthcareandourdailylife.AccordingtoStatista(2018),asofJanuary2018therewere3.7billionmobileusersworldwide.Whiletheglobalmobilebroadbandsubscriptionpenetrationrateisaround50%,theAmericasandEuropehavethehighestrates,around78.2percentand76.6percent,respectively.
Amongmillionsofmobileapps,healthandwellnessappsarebecomingincreasinglypopular.Therearecloseto48,000mobilehealthandwellnessappsavailablefromtheAppleAppStorealone(Statista,2018).Mobilehealthandwellnessappsnotonlychangethewayhealthcareisdeliveredandhowhealthismonitoredandmanaged,butalsoprovideadditionalbiometricsandlifestyledatatotheclinicians,researchers,andpolicymakersintheireffortstoimprovecareforindividualpatientsandpopulationsatlarge.
5. WeAKNeSSeS
5.1. Complex System of HealthcareTheUShealthcaredeliveryandpaymentsystemiscomplexwithmanyinterdependent,interactingstakeholders.Asacomplexadaptivesystem, itexhibitsnon-linear,dynamic,and indeterministicbehaviorsthatareunpredictableanddifficulttomanageandcontrol(Rouse,2008).Figure5showsthemanystakeholdersandhoweachof themplaysadifferent role inacomplex relationship todeliverhealthcare.
Improvinghealthcarerequiresthecollaborationandconcertedeffortsofallstakeholdersthroughconsensus building. Government policymaking should be conducted by taking inputs from allstakeholdersandanypotentialunintendedadverseconsequencesshouldbeevaluatedandremedied.Forexample,CMS’sbundledpaymentinitiativeintendedtoreducecostofcaremayleadtocareproviders’discriminationinpatientselectioncommonlyknownas“lemondropping”and“cherrypicking”,wherehighriskpatientsareturnedawayinfavoroflowriskpatients.Bigdataanalyticscanprovideevidenceandinsightstohelppolicymakersassesstheintendedbenefitsandunintendedharmofhealthpolicies.
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
8
5.2. Complex Determinants of HealthWhilemodernmedicinehasgreatsuccessattreatingsymptomsandmanagingbothacuteandchronicconditions,itisnotassuccessfulinpreventingandcuringdiseases.Thisisduetothecomplexnatureofdiseasesandthemultiplecontributingfactorsincludingclinical,behavioral,andsocioeconomicfactors.Thereisagrowingbodyofresearchlinkingsocioeconomicfactors,suchasanindividual’slifeconditionandsocial standing, tohisorherhealth statusunder thegeneral schemeof socialdeterminantsofhealth(Barr,2014;Marmot,2005;McGovern,Miller,&Hughes-Cromwick,2014).Thereisamyriadoffactorsthataffecthealthandinfluencehealthcare.Somearenaturalwhileothersarecultural;somearemeasurablewhileothersarenot.Machinelearningandartificialintelligencemaybeeffectiveinanalyzingnaturalormeasurablefactorsbutmustrelyonhumanintelligenceandhumanjudgementtodealwiththeculturalorunmeasurablefactors.
5.3. Complex Process of CareHealthcareisateamsportwheremultipleprofessionalsmustworktogethertodeliverqualitycaretopatients(Nancarrowetal.,2013).Forexample,thedialysiscareforpatientssufferingfromEnd-StageRenalDisease(ESRD)requirescoordinatedeffortsbyadministrative,medical,andsocialprofessionalsincluding facility administrators, medical directors, nephrologists, nurses, dialysis technicians,dietitians,andsocialworkers(MarylandDepartmentofHealth,n.d.).Inaddition,ESRDpatientstendtohavecomorbiditiessuchasdiabetesandhypertensionandarevulnerabletohospitalizationsandERvisits.Theirqualityoflifedependsonthecoordinatedcarebydialysisfacilities,hospitals,communitiesandtheirfamilies.Thisprocessofcareiscomplexandmakesthequalityofcaremuchhardertoquantitativelymeasureandanalyze.
Inaddition,patientengagementplaysacriticalroleinachievingoptimalhealthoutcomesinthepatient-centeredcareprocess.Patientsmustbeinformedofhowtheirtreatmentisbeinginfluencedbydata,evidence,andanalyticsandbeengagedinthehealthcaredecisionmakingprocess.Thisrequiresongoingpatienteducationtoimprovebothhealthliteracyanddataliteracy.
Thecomplexityarticulatedintheabovethreeinterrelatedareasposesgreatchallengestotheeffectivenessofbigdataanalytics.Asmuchaswewouldliketousetheinsightsgainedfrombigdataanalyticstoimprovetheperformanceofhealthcare,actionscannotbetakensolelybasedontheoutputsofnondeterministiccomputeralgorithms.Forexample,IBMWatsongeneratedconsiderablebuzzforitspurportedutilityinhelpingdoctorsmorerapidlyandaccuratelydiagnoseillness,butarecentarticlerevealedthatWatsonfailedtoliveuptothoseexpectations(Hernandez&Greenwald,2018).
Figure 5. Stakeholders and interests in health care (Rouse, 2008, p. 19)
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
9
6. OPPORTUNITIeS
6.1. Promise of Health Information exchangesWhileEHRsenablehealthcareorganizationstocentralizethemanagementofpatienthealthdata,theHealthInformationExchanges(HIEs)goonestepfurtherbyprovidingthetechnicalinfrastructuretoconnectthesedisparateanddiverseEHRssothatpatienthealthdatacanbeexchanged,aggregated,andsharedacrossthewholehealthcaredeliverysystem.“Electronicexchangeofclinicalinformationallowsdoctors,nurses,pharmacists,otherhealthcareproviders,andpatientstoaccessandsecurelyshare a patient’s vital medical information electronically - improving the speed, quality, safety,coordination,andcostofpatientcare”(TheOfficeoftheNationalCoordinatorforHealthIT,n.d.).HIEsmakeitpossibletoestablishcomprehensive,inclusive,andlongitudinalpatientandpopulationhealthdatafortheeffectiveapplicationofbigdataanalytics.
Since2009,theSocialSecurityAdministration(SSA)hasbeenusingHIEstoautomaticallyandtimelyrequestandobtainmedicalevidencerecords(MERs)fromhealthcareprovidersnationwidetosupportdisabilityclaimsadjudication(Feldman&Horan,2011).AsofDecember2018,over18,000healthcareproviders representedby150healthcareorganizationsparticipated in theexchangeofMERswithSSA.Comparedtoapaperprocess,theuseofHIEsgreatlyspeedsupmedicalevidenceacquisition to support disability determination. Faster availability of benefits resulting from theexpediteddeterminationhelpsdisability claimants pay formuchneededhealthcare services andsupportshealthcareprovidersinthedeliveryofhealthcare(SocialSecurityAdministration,n.d.a).Inaddition,SSAusesthevastamountofelectronichealthdataalongwithmachinelearningtoprovidedata-drivendecisionsupporttoitsdisabilityadjudicators.Thisuseofbigdataanalyticsfurtherspeedsupthedisabilityadjudicationprocessandhelpsincreasetheaccuracyandconsistencyofdisabilitydeterminationdecisions.SSAestablishedinteroperabilityguidelinesandacertificationprocesstoensureparticipatingorganizationscomplywithindustryinteroperabilitystandards(SocialSecurityAdministration,n.d.b).
6.2. Abundance of Data from All SourcesOneofthemanyV’sofbigdataisvariety.Forbigdataanalyticstobeeffective,manykindsofdatathatarerelatedandrelevanttohealthandhealthcareshouldbeincluded,especiallythesocioeconomicdatathatmeasuretheimportantsocialdeterminantsofhealth.EHRsmaycontainindividualpatient’sdemographicdatabutlacksthemacrosocioeconomicdatathatcanbeobtainedfromgovernmentcensusandsurveydata.
Forexample,theAmericanCommunitySurveydatafromtheUSCensusBureaucontainrichsetsofsocioeconomicdataaboutcommunitiesintheUSandcouldbeusedinconjunctionwithclinical,administrative,andclaimsdatatogainabetterunderstandingofthestateofhealthandhealthcare.Inaddition,datafromlawenforcement,non-governmentalorganizations,andsocialmedianetworkscanalsobeleveragedforbigdataanalyticsinhealthcare.
In January 14, 2019, President Trump signed into law the Foundations for Evidence-BasedPolicymaking(FEBP)Act.AspartoftheFEBPAct,theOpen,Public,ElectronicandNecessary(OPEN)GovernmentDataActrequiresallnon-sensitivegovernmentdatatobemadeavailableinopenandmachine-readableformats.Thesuccessful implementationof this federal lawwillhelppropeltheadoptionofbigdataanalytics.
6.3. Availability of Technology InnovationsThepastdecadehasseena rapidadvancement incomputerscienceand information technology.Theconfluenceofcloudcomputing,mobilecomputing,machine learning,artificial intelligence,andInternetofThingshasgivenrisetoaplethoraofnascenttools,techniques,andplatformsforperformingproductiveandeffectivebigdataanalytics.Therearemanyreadilyavailablechoicesof
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
10
bothcommercial-off-the-shelf(COTS)productsandfreeopensourcesoftware(FOSS)products.Therearemanyvirtualcommunities,blogs,forums,andtutorialsavailableontheweb.Datasciencehasemergedasaburgeoningvibrantprofession,andhundredsofdatascienceprogramshavesproutedupincollegesaroundtheworld.
7. THReATS
7.1. Cyberattacks and Data BreachesWhilebigdataanalyticshasthepotentialtohelpimprovethequalityandperformanceofhealthcare,therisksofcyberattacksanddatabreachesarehighandshouldnottobeoverlooked.Accordingtoarecentstudyoverafive-yearperiodfrom2012to2017,therewere1,512reporteddatabreachesofprotectedhealthinformationaffectingatotalofmorethan154millionpatientrecords(Ronquillo,ErikWinterholler,Cwikla,Szymanski,&Levy,2018).Thesestolendatacanbeusedfor“identitytheft,criminalimpersonation,taxfraud,healthinsurancescams,andahostofothercriminaloffenses”(Goodman,2016,p.109).
ThecyberattacksonhealthcareITsystemsnotonlyputpatients’privacyandsecurityatrisk,theyalsoposeeconomicharmstohealthcareproviders,insurers,andtaxpayers.A2016studybythePonemonInsfitute(2016)estimatedthatabout90%ofthehealthcareorganizationsrepresentedinthestudysuffereddatabreachesinthepasttwoyearsandtheaveragecostofdatabreachesforcoveredentitiessurveyedwasmorethan$2.2million,whiletheaveragecosttobusinessassociatesinthestudywasmorethan$1million.
7.2. Unethical Use of Health DataExternalmaliciouscyberattacksarecertainlygraveconcerns,howeverinsiderthreatsandunethicaluseofhealthdatatodiscriminatehealthinsurancecoverageandtoboostcorporaterevenuesandprofitsshouldnotbeoverlooked.Bigdataanalyticsisadouble-edgedsword.Whenappliedproperlyandethically,ithasthepotentialtodogood;otherwise,itmaycauseharmtopatients,providers,andtaxpayers.Oneemergingapplicationistheuseofpredictiveanalyticstodrawinsightsforprofitswithoutregardtoprivacylawsandprotectionofpatients’personalandhealthinformation.
A2017reportbyTheCenturyFoundation(Tanner,2017)paintedagloomypictureofbigdatainhealthcare.Thereexistsamulti-billion-dollarindustrythatcollects,mines,buys,andsellsanonymizedpatienthealthdata.Thepatienthealthdataaretradedroutinelyforprofit.Whilethesharinganduseofde-identifiedpatienthealthdataforsecondaryuseinmedicalandpolicyresearchareallowedbytheHealthInsurancePortabilityandAccountabilityAct(HIPAA)(Cohen&Mello,2018),thereisarealdangerforthepatienthealthdatatobere-identifiedthroughtheprocessofdatalinkingwithadditionaldatasourcesfromsocialmedianetworksandmobilehealthandwellnessappsandtheuseofmachinelearningalgorithms.Thisunethicalfor-profitdataminingalongwiththeupsurgeinmalicioushackinganddatabreachescanresultindevastatingimpacts.
7.3. Biases in Data and AlgorithmsRealityiscomplexandunknowable.AsstatedbyLaoziinthe2500-yearoldTaoisttextTao Te Ching,“ThetaothatcanbetoldisnottheeternalTao.ThenamethatcanbenamedisnottheeternalName.Theunnamableistheeternallyreal”(Laozi,Mitchell,Roig,&Little,1989).EchoingLaozi,statisticianGeorgeBox(1976)wasfamouslyquotedforthemaximthat“allmodelsarewrong”becausetheyareonlyapproximationsoftruereality.Bigdataanalyticsisinherentlybiasedsinceitreliesondataasinput,andalgorithmsastheenablers.Limitationsandbiasesexistinbothdataandalgorithmswhichcanbetracedbacktothecognitivelimitationsandbiasesinhumanminds(Gianfrancesco,Tamang,Yazdany,&Schmajuk,2018).
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
11
Whethercollectedthroughhumanobservationsorsensorydevices,dataareapproximatemeasuresoftheactualpropertiesoftheobservedentities.Evencorrectlymeasuredandcurateddatacannotescapefromthebiasesofthosewhodefinethemeasurementandthosewhodesignthedatacollectioninstruments. Inaddition, incomplete, inaccurate,andmissingdataare typicalandcandistort theoutcomeoftheanalytics(affectionatelyknownas“garbagein,garbageout”).
8. CONCLUSION
Thispaperpresentedahigh-levelanalysisofthestrengths,weaknesses,opportunities,andthreats(SWOT)intheapplicationofbigdataanalysisinhealthcare.Whilethestrengthsandopportunitiesarepositive,desirable,andeasiertocomprehend,theweaknessesandthreatsarenegative,undesirable,anddemanddiligenceandvigilance.Althoughthispaperprovidesevencoverageofallfourfactors,itshouldbestressedthatmoreattentionmustbepaidtotheweaknessesandthreats.Thisisespeciallytrueintheeraofrapidinnovationsininformationtechnologywhichoftengiverisetoanunrealisticexpectationofapanaceawithquick,autonomous,andmagicalcuresforacomplexsocialproblem.
Technologysuchasmachinelearningandartificialintelligencehavebeeneffectiveindealingwith“hard”problemssuchaswinning“Go”games,recognizinghumanvoicesandfaces,andeventranslatinglanguages.Theseproblemshavewell-definedboundaries,well-understoodrulesofgameorcausalmechanisms,andcanbesimulatedorprogrammedusingsoftwarewithhighdegreesofaccuracyandcertainty.However,socialproblemsaremuchmorecomplexwithill-definedboundariesandpoorly-understoodcausalmechanisms(Kirk,1995).Healthcareisaperfectexampleofacomplexsystemwithnon-linear,dynamic,andindeterministicbehaviorsinvolvingmultiplestakeholderswithconflictsofinterestsandundertheinfluenceofmultitudesofintertwinednaturalandsocialforces.TheseweaknessesareinherentinhealthcareandshouldbekeptinmindwhenapplyingtechnologiessuchashealthITandbigdataanalytics.Thereisnosimpleandstraight-forwardsolutiontothehighcostandlow-qualitychallengesfacingtheUShealthcaresystem.Bigdataanalyticsispromisingandcanhelpuncoverpartialandlimitedknowledgefrombigdatatoinformclinicalandpolicydecisionmaking,butitdoesnotprovidethewholetruthabouthealthcareandisnotasilverbullet;overcomingthe weaknesses requires the conscious and judicial use of political skills, business acumen, andcollaborativespiritinadditiontotechnicalandanalyticalcompetency.
Whilethispaperprovidesonlycursoryexposuretothesubjectofbigdataanalyticsinhealthcare,oneofitsobjectivesistodrawattentiontosomeofthemanyinterrelatedfactorsthatmustbeconsideredwhenseekingtoleveragebigdataanalyticstoimprovehealthcarequalityandperformance.Amoredetailedanddeeperanalysisonanyofthefactorsmentionedintheabovehigh-levelanalysiswouldrepresentaworthwhileresearchsubject.Forexample,healthcareisahighlypersonalizedprocessinvolvingthecoordinationofmultipleproviderssuchasacutecarehospitals,primarycarephysicians,andnursepractitioners.Carecoordinationisaknownfactorinfluencingsuchhealthcareoutcomesas unplanned hospital readmission (Fluitman et al., 2016). Quantitatively measuring health carecoordinationisachallengingtask.Comparedtoreadily-quantifiableclinicalmeasuressuchasbloodpressureorbodymass index, individualhumanbehaviorsaremuchharder toquantify,andcarecoordinationinvolvesmultipleparties.Moreover,fairandjustattributionofqualitymeasurescorestomultipleprovidersisalsoadifficulttask.Thisisfurthercomplicatedbypatientbehaviorssincepatientsarealsopartofthecoordinationprocess.
DISCLAIMeR
Theauthor’saffiliationwithTheMITRECorporationisprovidedforidentificationpurposesonlyandisnotintendedtoconveyorimplyMITRE’sconcurrencewith,orsupportfor,thepositions,opinionsorviewpointsexpressedbytheauthor.ApprovedforPublicRelease;DistributionUnlimited.CaseNumber19-0249.
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
12
ReFeReNCeS
Analytics, H. I. M. S. S. (n.d.). Adoption Model for Analytics Maturity. Retrieved from https://www.himssanalytics.org/amam
Barnes,A.J.,Unruh,L.Y.,Rosenau,P.,&Rice,T.(2018).HealthSystemintheUS.InE.vanGinneken&R.Busse(Eds.),Health Care Systems and Policies.NewYork,NY:Springer.doi:10.1007/978-1-4614-6419-8_18-2
Barr,D.A.(2014).Health disparities in the United States: Social class, race, ethnicity, and health.JHUPress.
Box,G.E.(1976).Scienceandstatistics.Journal of the American Statistical Association,71(356),791–799.doi:10.1080/01621459.1976.10480949
Centers for Medicare and Medicaid Services. (2012). National health expenditure projections 2012-2020.Retrieved from https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/Downloads/Proj2012.pdf
Cohen,I.G.,&Mello,M.M.(2018).HIPAAandProtectingHealthInformationinthe21stCentury.Journal of the American Medical Association,320(3),231.doi:10.1001/jama.2018.5630PMID:29800120
Collins, B. (2016). Big data and health economics: Strengths, weaknesses, opportunities and threats.PharmacoEconomics,34(2),101–106.doi:10.1007/s40273-015-0306-7PMID:26093888
Dantes,R.B.,&Epstein,L.(2018).CombattingSepsis:APublicHealthPerspective.Clinical Infectious Diseases,67(8),1300–1302.doi:10.1093/cid/ciy342PMID:29846544
Feldman,S.S.,&Horan,T.A.(2011).Collaborationinelectronicmedicalevidencedevelopment:AcasestudyoftheSocialSecurityAdministration’sMEGAHITSystem.International Journal of Medical Informatics,80(8),e127–e140.doi:10.1016/j.ijmedinf.2011.01.012PMID:21333588
Fluitman,K.S.,vanGalen,L.S.,Merten,H.,Rombach,S.M.,Brabrand,M.,Cooksley,T.,&Nanayakkara,P.W.B.et al.(2016).Exploringthepreventablecausesofunplannedreadmissionsusingrootcauseanalysis:Coordinationofcareistheweakestlink.European Journal of Internal Medicine,30,18–24.doi:10.1016/j.ejim.2015.12.021PMID:26775179
Gianfrancesco,M.A.,Tamang,S.,Yazdany,J.,&Schmajuk,G.(2018).Potentialbiasesinmachinelearningalgorithmsusingelectronichealthrecorddata.JAMA Internal Medicine,178(11),1544–1547.doi:10.1001/jamainternmed.2018.3763PMID:30128552
Giannini,H.,Chivers,C.,Draugelis,M.,Hanish,A.,Fuchs,B.,Donnelly,P.,&Schweickert,W.et al.(2017).D15CriticalCare:DoWeHaveACrystalBall?PredictingClinicalDeteriorationandOutcomeinCriticallyIllPatients:DevelopmentAndImplementationOfAMachine-LearningAlgorithmForEarlyIdentificationOfSepsisInAMulti-HospitalAcademicHealthcareSystem.American Journal of Respiratory and Critical Care Medicine,195.
Goodman,M.(2016).Future crimes: Inside the digital underground and the battle for our connected world.RandomHouse.
Hernandez,D.,&Greenwald,T. (2018). IBMHasaWatsonDilemma.The Wall Street Journal.Retrievedfrom https://www.wsj.com/articles/ibm-bet-billions-that-watson-could-improve-cancer-treatment-it-hasnt-worked-1533961147
Hill,T.,&Westbrook,R.(1997).SWOTanalysis:It’stimeforaproductrecall.Long Range Planning,30(1),46–52.doi:10.1016/S0024-6301(96)00095-7
InstituteforHealthcareImprovement.(2007).TripleAim-TheBestCarefortheWholePopulationattheLowestCost.Retrievedfromhttp://www.ihi.org/Engage/Initiatives/TripleAim/Pages/default.aspx
Kirk,D.(1995).Hardandsoftsystems:Acommonparadigmforoperationsmanagement.International Journal of Contemporary Hospitality Management,7(5),13–16.doi:10.1108/09596119510090708
Laozi,M.S.,Roig,J.V.,&Little,S.(1989).Tao teaching.KyleCathie.
Lebied,M.(2018).12examplesofbigdataanalyticsinhealthcarethatcansavepeople.Datapine.Retrievedfromhttps://www.datapine.com/blog/big-data-examples-in-healthcare/
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
13
Marmot,M.(2005).Socialdeterminantsofhealthinequalities.Lancet,365(9464),1099–1104.doi:10.1016/S0140-6736(05)74234-3PMID:15781105
MarylandDepartmentofHealth.(n.d.).CodeofMarylandregulations:Sec.10.30.02.04.freestandingdialysisfacilities–staffing.Retrievedfromhttp://mdrules.elaws.us/comar/10.30.02.04
McGovern,L.,Miller,G.,&Hughes-Cromwick,P.(2014).The relative contribution of multiple determinants to health outcomes.ProjectHOPE.
Nancarrow,S.A.,Booth,A.,Ariss,S.,Smith,T.,Enderby,P.,&Roots,A. (2013).Tenprinciplesofgoodinterdisciplinary team work. Human Resources for Health, 11(1), 19–19. doi:10.1186/1478-4491-11-19PMID:23663329
OfficeoftheNationalCoordinatorforHealthInformationTechnology.(2018).Percentofhospitals,byType,thatpossesscertifiedhealthIT.Retrievedfromhttps://dashboard.healthit.gov/quickstats/pages/certified-electronic-health-record-technology-in-hospitals.php
PatientProtectionandAffordableCareActof2010,42U.S.C§18031.(2010).USCongress.
PonemonInstitute.(2016).Six annual benchmark study on privacy & security of healthcare data.Retrievedfromhttps://www.ponemon.org/local/upload/file/Sixth%20Annual%20Patient%20Privacy%20%26%20Data%20Security%20Report%20FINAL%206.pdf
Reinhart,K.(2018).Sepsis:Aglobalpublichealthburdenforhospitals.Becker’s Hospital Review.Retrievedfromhttps://www.beckershospitalreview.com/quality/sepsis-a-global-public-health-burden-for-hospitals.html
Ronquillo,J.G.,ErikWinterholler,J.,Cwikla,K.,Szymanski,R.,&Levy,C.(2018).Health IT, hacking, and cybersecurity: national trends in data breaches of protected health information.JAMIAOpen.
Rouse,W.B. (2008).Healthcareasacomplexadaptivesystem: Implications fordesignandmanagement.Bridge-Washington-National Academy of Engineering,38(1),17.
Sackett,D.L.,Rosenberg,W.M.,Gray,J.M.,Haynes,R.B.,&Richardson,W.S.(1996).Evidencebasedmedicine: What it is and what it isn’t. BMJ: British Medical Journal, 312(7023), 71–72. doi:10.1136/bmj.312.7023.71PMID:8555924
Schneider,E.C.,Sarnak,D.O.,Squires,D.,Shah,A.,&Doty,M.M.(2017).MirrorMirror2017:InternationalComparisonReflectsFlawsandOpportunitiesforBetterUSHealthCare.Retrievedfromhttps://interactives.commonwealthfund.org/2017/july/mirror-mirror/
Senthilkumar,S.,Rai,B.K.,Meshram,A.A.,Gunasekaran,A.,&Chandrakumarmangalam,S.(2018).BigDatainHealthcareManagement:AReviewofLiterature.American Journal of Theoretical and Applied Business,4(2),57–69.doi:10.11648/j.ajtab.20180402.14
SocialSecurityAdministration.(n.d.a).HealthITPartners.Retrievedfromhttps://www.ssa.gov/hit/partners.html
SocialSecurityAdministration.(n.d.b).SSAgov/HealthIT.Retrievedfromhttps://github.com/ssagov/healthit
Statista. (2018). Mobile Internet - Statistics & Facts. Retrieved from https://www.statista.com/topics/779/mobile-internet/
Tandon,A.,Murray,C.J.,Lauer,J.A.,&Evans,D.B.(2000).Measuring overall health system performance for 191 countries.Geneva:WorldHealthOrganization.
Tanner,A.(2017).Our bodies, our data: How companies make billions selling our medical records.BeaconPress.
Terhaar,M.F.(2018).Value-BasedPurchasing.Clinical Analytics and Data Management for the DNP,1,27.
TheOfficeoftheNationalCoordinatorforHealthIT.(n.d.).Whatishealthinformationexchange?Retrievedfromhttps://www.healthit.gov/faq/what-health-information-exchange
TheOrganizationforEconomicCo-operationandDevelopment.(2018).Spending on Health: Latest Trends.Retrievedfromhttp://www.oecd.org/health/health-systems/Health-Spending-Latest-Trends-Brief.pdf
Wang,C.(2018).Integratingdataanalytics&knowledgemanagement:Aconceptualmodel.Issues in Information Systems,19(2),208–216.
International Journal of Big Data and Analytics in HealthcareVolume 4 • Issue 1 • January-June 2019
14
Chaojie Wang is a seasoned software engineer, systems architect, data scientist, researcher, and project manager with over thirty-year experience in industry, government, and academia. Dr. Wang currently works for The MITRE Corporation, an international thinktank and operator of several Federally Funded Research and Development Centers (FFRDC). In his capacity as a principal systems engineer, Dr. Wang advises the federal government on IT Acquisition & Modernization, Data Analytics & Knowledge Management, and Emerging Technology Evaluation & Adoption. Previously, Dr. Wang worked for Lockheed Martin Corporation and participated in multiple large-scale IT projects for the federal government. Dr. Wang is a certified Project Management Professional (PMP), a certified SAFe Agilist (SA), and currently serves on the Editorial Review Board for the International Journal of Patient-Centered Healthcare (IJPCH) by IGI Global and the Issues in Information System (IIS) by the International Association for Computer Information Systems (IACIS). Dr. Wang holds a Bachelor of Engineering in MIS from Tsinghua University, a Master of Art in Economics and a Master of Science in Statistics both from the University of Toledo, an MBA in Finance from Loyola University Maryland, and a Doctor of Science in Information Systems and Communications from Robert Morris University.
Wang,L.,Wang,G.,&Alexander,C.A.(2015).Bigdataandvisualization:Methods,challengesandtechnologyprogress.Digital Technologies,1(1),33–38.
Wang,Y.,Kung,L.,&Byrd,T.A.(2016).Bigdataanalytics:Understandingitscapabilitiesandpotentialbenefitsforhealthcareorganizations.Technological Forecasting and Social Change.doi:10.1016/j.techfore.2015.12.019
Yang,Y.,Yu,S.,Hao,Y.,Xu,X.,&Liu,H.(2016).TheSWOTAnalysisoftheWearableDevicesinSmartHealthApplications.Paper presented at theInternational Conference on Smart Health.