Eapm report big data workshop european health forum[1]

16th European Health Forum Gastein

W o r k s h o pB I G D A T A

REPORT - OCTOBER 2013

COLOPHON

Production VitalTransformationBVBA www.vitaltransformation.com

Words Nuala Moran

Design Peter Koekoek

Photos European Health Forum Gastein BigStockPhoto

The Big Data Workshop was organised by the European Alliance for Personalised Medicine and supported by the Lithuanian Health Forum, EFPIA, IBM and Pfizer in Gastein, Austria from 2 – 4 October 2013.

ORGANISED BY

SUPPORTED BY

B I G D A T A W O R K S H O P 3

Harnessing big data to improve R&D and the translation of new therapies

With health care under pressure from many angles, the key issue is to ensure sustainability of health systems in the face of these challenges. Rational application of big data has an important role to play.

Itcanbedifficulttotalkaboutthepromiseofinvestinginbigdataattimesoffallingresourcesforhealthcare,saidLeonas Kaletinas, Member of the Board of Directors

of the Lithuanian Health Forum.However,“Thereclearlyisopportunityforcountrieslikeours,andforpatients,professionalsandpayerstousethebasicsofbigdatatobecomecost-effective,toavoidwaste,toincreasevalueandincreaseefficiency,”KaletinassaidopeningthedebateintheBig Data Workshop,heldaspartofthe16th European Health Forum in Gastein,Austriafrom2–4October.

“Foroursystembigdatacouldbearevolution,becausewecouldseehugechangesinhealthcaremanagement,”Kaletinassaid,inthefirstsession,‘Big Data and Best Practice for Public Health’.

AnimportantaspectforLithuaniainitscurrentEUPresidencyroleistostressthepotentialbenefitsofbigdatainhelpingtostrengthenhealthsystems.‘Sustainable Health Systems for Inclusive Growth in Europe’isthetopicfortheLithuanianPresidencyConferenceinVilniuson19–20November.KaletinassaidshehopestheconferenceinVilniuswillprogressthingssothatnextyear,“bestpracticeandnotbarriers”arethefocusofattempttomovethedeploymentofbigdataforward.

TwoparticularareaswheretheEuropeanCommissioncouldhelpopendoorstobigdata,areinnewdataprotectionlegislationandthroughfundingprogrammestopromoteadoption,Kaletinasbelieves.ForcountrieswithlimitedfundingandashortageoftheappropriateskillsbeingpartofaEuropeannetworkwouldbeveryhelpful.

LeonasKaletinas,MemberoftheBoardofDirectorsoftheLithuanianHealthForum

41 6 T H E U R O P E A N H E A L T H F O R U M G A S T E I N

Itisnecessarytotakeameasuredapproachtobigdata–bearinginmindthatqualitycountsmorethan

quantity,Terje Peetso, Unit H1 – Health and Well Being, DG Connect at the European Commission suggested.Theremaynowbedifferencesinscale,butmedicalsciencehasalwaysbeenaboutcollectingdatasaidPeetso.“It’snotabouttakingbloodit’saboutanalysingit,”shesaid.

Similarly,bigdatashouldnotbeseenassomesortofendinitself,butshouldbeappliedinwaysthatcontributetopublichealthandpatientempowerment.Agoodplacetostartwouldbeexistingdatarepositories.“It’slikerefiningcrudeoiltoaddvalue;weshouldlookintodatawealreadyhaveandgetvaluefromthat,”Peetsosaid.

Thiswouldprovidethefoundationforlookinginto“new”datathatis

accumulating.Inhealth,itisimportantnottofocussolelyondataperse,butalsoonquestionsaboutownership,trustandaccess,andoninteroperabilitybetweendifferentcomputersystemstoensureallrelevantdatacanbeassembledandanalysed.

Indeed,interoperabilityisseenasaparticularissueinapplyingbigdatatopublichealth,withmanydisparatesystems,bothwithindifferenttiersofnationalhealthservicesandbetweenmemberstates.

Thequestionofinteroperabilityofmedicaldatasystemskeepscroppingup,andtherehavebeenmanyattemptstodealwiththis.Whileitis“amonumentaltask”therearesomecompellingexamplesofthevalueofpersisting,saidJohn Crawford, Healthcare Industry Leader Europe, IBM. The European Union’sepSOS(EuropeanPatientsSmart

OpenServices)system,whichnotonlytransmitsmedicaldatafromonecountrytocountry,butalsotranslatesfromonelanguagetoanother,“isagreatexample,”Crawfordsaid.

AnotherisScotland’sEmergencyCareSummarydatabasecontainingasummaryofdemographic,allergyandmedicationinformationfor5.5millionpeople.Itenableshealthcareprofessionalstoaccesstoimportantpatientinformationinemergencyandunscheduledcaresituations.

Quality not quantity

B I G D A T A W O R K S H O P 5

TerjePeetso,UnitH1–HealthandWellBeing,DGConnectattheEuropeanCommission

Amelia Andersdotter, MEP (Greens/EFA, Sweden)hasadeep

personalunderstandingoftheissuesofownership,trustandsecurityintheageofbigdatabecausesomeofherownchildhoodmedicalinformationhasbeenpublishedinSwedenwithoutherconsent.TheinformationisavailableviatheInternet,whichAndersdotterpointedout,“isnotaplacethatissecure.”

ForAndersdotter,therearetwoparticularissueshere:thatinformationtechnologiesarebeyondthecompetencesofmostmedicalpractitionersandthattheInternetisoutsidethelaw.AsAndersdotterpointedout,“The[US]NationalSecurityAgencycan’tinvademyhousebuttheycaninvademyInternet.”

HavingnocontroloverthepublishedmedicalinformationismakingAndersdotter“uncomfortable”andleadinghertoavoidcontactwithhealthinstitutions,asituationwhichisexactlytheoppositeofhowanyonewouldliketoseetheapplicationofbigdataplayingout.

AsAndersdotternoted,inallothercircumstancesthereishighconfidencethatmedicalrecordsaresecure,withnoaccessforthecourts,thepoliceorthepress.Evenifanindividualgivesinformedconsentfortheirdatatobeused,theproblemofsecurityofdataremains.“TheInternetisfordistributinginformation,soifthere’sinformationyoudon’twantdistributing,it’snottherighttool-likeahammer,theInternetsolvessomeproblemsbutnotothers,”Andersdottersaid.

Afurtherquestionrevolvesaroundthenatureofconsent.Whenanindividualagreestheirhealthinformationcanbemadeavailableinsomeway,theydonotknowwhattheyareconsentingto.“Thedoctorinputsthedata,sothepatientdoesn’tknowifthere’sarisk,”saidAndersdotter,adding,“therearesomeseriouschallengesforpoliticiansandindustrytopreservetheconfidenceofcitizens.”

TheEuropeanCommissionneedstoaddresspointsoffailureinthesystembeforebigdatamovesahead,Andersdottersaid,suggestingamoregranularapproachtouseofdatamaybeappropriateandthattheimmediatefocusshouldbeonopeningupdatathatisnotpersonalbutisneverthelessrelevanttopublichealth,suchasclinicaltrialsdataandinformationinsubscriptionmedicaljournals.

JohnCrawfordagreedthedesignprinciplesoftheInternet,whicharebasedonsharinginformation,makeitinherentlyunsuitableforsensitiveprivatedata.Buthenoted,privatedatastoresintheCloud,withappropriateencryptionandothersecuritymeasures,couldprovideasolution.

Theappropriatesecuritymeasuresdependonhowtheinformationisused,Andersdotterbelieves.Pseudonymisationmayworkwithmachinereadingofdata,butwouldnotnecessarilypreserveconfidentialityiftheinformationisbeingreadbyahuman.“Youneedtothinkaboutissueslikethisinimplementinganytechnicalsystems,”Andersdottersaid.

The Internet is a hammer

AmeliaAndersdotter,MEP(Greens/EFA,Sweden)

Thechallengeofbigdataandhealthinformaticsliesnot

onlyincapturingandstoringinformationsecurely,butalsoindevisingthetoolsandanalyseandmanageit,saidBarbara Kerstiens, Head of Sector, Public Health, DG Research at the European Commission,openingthesecondsessionoftheworkshop on ‘Data Sharing for Improved Research and Translation’.

TheCommissionisfundingresearchonboththeseaspects.Anumberofissuesneedtobeaddressedinharnessingthepowerofbigdatatoimproveresearchandspeeduptranslationofresearchoutputstoimprovehealthbothatapublichealthlevelandinthedevelopmentofpersonalisedmedicines.

Theseinclude:standardisation,integration–especiallytoachieveeconomiesofscale,forexampleinresearchintorarediseases;thechallengeofopenaccess,andnotjustmakingdataavailablebutensuringitisreadableanduseable;theneedfornewstatisticalmethodsandtools;andprovidingthemeanstotrackclinicaloutcomes.

Dealingwiththis‘todo’listcallsforinternationalcollaboration.“No

individualcountrycandealwiththesechallengesorgetthebenefits[alone],Kerstienssaid.TheEuropeanCommissioniswellplacedtosupportthenecessaryresearch,buildingonpreviousinvestmentssuchastheEuropeanBioinformaticsInstituteandasignificantnumberofinternationalcollaborationsithasfundedinthisfield.

ThereisalsoaneedforanEU-levelpublicprivatepartnershipinvolvingallstakeholderstoconsiderallaspectsofdatasharingandaccess,toensurethereisaparticipant-centredapproach.

Fromaresearchperspective,itiscriticaltoavoidperpetuatingdatasilosthataredisconnectedfromoneanother,sincethiswilllimitthepotentialforbigdataanalyses.“It’saworkinprogress

andcontinuingtalksareneeded,”Kerstienssaid.“Providersofdataneedtounderstandthechallenges.”

TheInnovativeMedicinesInitiativeisanexampleofanEU-fundedprogrammethataimstoimprovedrugdevelopmentandregulationthroughtheuseofpooleddata.Meanwhile,newEUClinicalTrialsandDataProtectionrulesthatarebeingformulatedcurrentlymeantheissuesrelatingtobigdata,healthresearchandprivacy,areonthetable.“Itisaconversationthathasstartedandistobecontinued,”saidKerstiens.

The issues are on the table

BonnieWolff-Boenisch,HeadofResearchAffairsatScienceEuropeandJohnCrawford,HealthcareIndustryLeaderEurope,IBM

81 6 T H E U R O P E A N H E A L T H F O R U M G A S T E I N

B I G D A T A W O R K S H O P 9

Bonnie Wolff-Boenisch, Head of Research Affairs at Science Europe,thebodyestablishedin2012torepresenttheviewsoftheleadingresearchfunding

organisationsaroundEurope,toldtheworkshopthatallresearchuniversitiesaretryingtostrikeabalancethatensuresthepotentialofbigdataisrealised,butthatprivacyisnotcompromised.

Openingupdataisimportant:“Ifyouhavebrightmindsaccessingit,youdon’tknowwhatwillcomeoutofit,”Wolff-Boenischsaid.ScienceEuropeconsiderssomerulesaretoostrictandcouldhampercertaintypesofresearch.Forexample,arequirementtogetinformedconsentforeachindividualpieceofresearchwouldmakebiobanksinefficient.

Similarly,itisimportantnottobetooprescriptive,sincethekeytoextractingvaluefromdataistobeabletoapplynewtools,to“playaround”andcomeupwithnewmethodsandapproachesforconvertingdatatousefulinformation.“Weneedtobeabletoreadthebookofbigdata,”saidWolff-Boenisch.

Thepowerinherentinbigdataisthatitcanprovide“individualisedevidence”leadingtothedevelopmentoftruelypersonalisedmedicine,saidAngela Brand, Professor of Health, Medicines and Life Sciences at Maastricht University and Co-chair of the workshop on behalf of the European Alliance for Personalised Medicine.

Bigdatawillprovidethemeansfordecisionsupportacrossallaspectsofhealthcare–rangingfromassessingsafety

andefficacyofdrugs,tocarryingouthealthtechnologyassessments,andprevention,diagnosisandtreatment-toberefocusedfromthepopulationlevel,onesizefitsallparadigm,totheindividual.“Weneedtogetindividualevidence,”Brandsaid,“andthisinformationmustbeavailableonajust-in-basis.”

Achievingthisambitionraisespresentschallengesaroundthegovernanceandthequalityofimplementationofbigdatainhealth,andcallsforstandardsforconsolidating,characterising,validatingandprocessingdata.However,Brandsaid,itsinherentdiversityandcomplexitymeanshealthinformation“willalwaysbemessy”,raisingthequestionofhowtosetthebarinassessingqualityofimplementation.

Whiledatausersshouldbeaccountableforthecustodianshipofpersonalmedicalinformation,itisimpossibletoguaranteecompletedatasecurity,anditwouldbedishonesttodoso.Giventhis,Brandsuggestedamoreappropriateapproach–toreplacetherequirementforindividualinformedconsenteverytimesomeone’sdataisused–wouldbethatindividualdatasetsareaggregatedintobigdataalgorithms.

Inthediscussion,delegatesraisedanumberofotherissues,includinghowtoguaranteethequalityofdatagoingintoshareddatabases,agreeingtechnicalapproachestowhichallstakeholderscansignup,anddevelopingsustainablebusinessmodelsforthedeploymentofbigdatainhealth.

Reading the book of big data

1 01 6 T H E U R O P E A N H E A L T H F O R U M G A S T E I N

It’sclearthatestablishingtrustisessentialifthetechniquesandtoolsofbigdataaretobesuccessfully

appliedtohealth.Thekeytothisistransparencysothat,“peoplewhogivedataknowwhathappenstoitandfollowit,”saidErnst Hafen of the Institute of Molecular Systems Biology, ETH Zurich,openingthethirdsessionon‘Big Data and improved evaluation models for efficacy and efficiency’.

Hafensuggested‘ThePeople’sHealthDatabank’asamodelforengenderingtrustandtransparency.Thiswouldbeasafeandsecureplacetostoredata,whichpeopletrustinthesamewayastheytrustabanktostoretheirmoney,andtotransmitittothirdpartiesontheinstructionsoftheaccountholder.

Suchhealthandgenomicdatabankscouldberunascooperatives,withrequestsfordataaccessforresearchpurposeshandledcentrally,andindividualshavingtherighttowithholddatafromparticularpiecesofresearch.Forcompaniesrequestinginformationforadrugdevelopmentproject,therewouldbeacharge,withthemoneyinvestedbackintotherunningofthedatabank.

ThehugepotentialthatbigdataholdstodrivethedevelopmentofpersonalisedmedicinemakesitappearcountertotheprincipleofsolidaritythatunderpinsEurope’shealthcaresystems.OrganisingthePeople’sHealthDatabankasacooperativewouldenshrinesolidarityinthenewageofpersonalisedmedicine.Individualswouldsharetheirdatatogetcuresforthemselvesandforeveryoneelse.

Hafenproposedtherewouldbeacooperativehealthdatabankineverycountry,eachusingthesamedatastandardssoinformationcouldbesharedbetweenthem.

TheideaofcollectivelycreatingconsentinthePeople’sHealthDatabankisverycompelling,believesAdam Heathfield, Director of Science Policy Europe at Pfizer.Thecooperativemodelisagoodoneifpeoplebuyin.

Asacompany,Pfizerismakingaconcertedefforttomakebetteruseofreallifedataandhasputanewteaminplacetolookatthis.InR&Dthefirststepwillbetobuildonepidemiologicalinformationtobecomesmarterintargetselection,andthenoverlaygenomicstolinkgenotypesandphenotypes.Forexistingproducts,bigdatawillbeusedtoanswerquestionsabouthowwellmedicinesactuallyperforminthemarketandtoprovideinputsforhealthtechnologyassessments.

Post-marketingstudiesandhealthtechnologyassessmentsarebecomingamuchbiggerburden,requiringinformationthatcannotbegeneratedinclinicaldevelopment,andbigdatapromisestoprovidesomerelief.Issuesremainingtoberesolvedincludeguaranteeingdataqualityanddevelopingrobustmethodsforframingandansweringquestions.“Wearealongwayfromhavingthedataanalysistoolsweneed,”Heathfieldsaid.

While,assuggestedinHafen’sPeople’sHealthDatabankmodel,Pfizerispreparedtopayforaccesstoanonymiseddatasets,Heathfieldnoted

thatpharmaceuticalcompaniescannotpaypeopletotakepartinclinicaltrials(thoughtheycanpayexpenses).“Therewouldbeaproblemofacooperativegenuinelygettingconsentandbeingpaidfordata,withoutskewingthatissue,”saidHeathfield.

Hafensuggestedthiscouldbefinessedbyagatekeepingfunction.Forexample,inadatabasewith10millionrecords,theremightbe30,000womenwithaBRCAgenemutationwhohaveagreedtoshareinformationontheirstatus.IfPfizerpaidforaccess,thecooperativewouldfilterthedatabaseandthenapproachthewomenandaskiftheywantedtoparticipateinaclinicaltrial.

Thetechnologyisathandtoapplybigdatatohealth,buttheremustbeapublicdebateabouttherisksassociatedwithdatasharing,saidJohnCrawford.Furthermore,thereisnopointinaccumulatingdataunlessitisthenanalysedandtheresultstranslatedintoaction.

Healthcertainlyfitsthebigdataparadigmintermsofthevolumeofdataitgenerates.However,itremainsthe

We are all health billionaires

AngelaBrand,ProfessorofHealth,MedicinesandLifeSciencesatMaastrichtUniversity

B I G D A T A W O R K S H O P 1 1

casethatthemajorityofthisisheldintextfiles.“Youhavetogetitintoastatewhereyoucandosomethingusefulwithit,”Crawfordsaid.

Informationheldinhealthrecordsmaynothavetheotheressentialbigdatapropertyofvelocity,butotherformsofhealthdatado.Inonefamousexample,GoogleclaimedtohavetrackedtheoutbreakofseasonalflubeforetheUSCentersforDiseaseControlandPreventionbecausepeoplestartedusingthecompany’ssearchenginetolookupsymptoms.

Similarly,analysisofdatageneratedbymonitoringdevicesinintensivecareunitscanpickupsignsofnosocomial(hospital-acquired)infectionsbefore

thereareobservablesymptoms.Healthdataalsofitsthebigdatamouldintermsofvariability,withinformationoftenbeinginconsistent,incompleteandcontradictory,Crawfordnoted.IBM’sWatsoncomputer,withitsabilitytoreadandunderstandnaturallanguageandweighevidence,ismovingdecisionsupporttoanewlevel,allowingdoctorstoaccessandinterpretallthelatestevidenceandmakebetterdecisionsasaresult.Thisalsohighlightsthewayinwhichbigdatacanshiftanalyticsfromretrospectivetorealtime.

Summingup,Ralf Sudbrak, Scientific Coordinator, Max Planck Institute for Molecular Genetics notedthatconcernsaboutdataprotectionvaryfromtoomuchtotoolittle,dependingonthe

individual’sperspective.“Weneedtomakesurewearenottooprotective;itisnecessarytofindthebalancebetweentheneedfordataprotectionandtheuseofbigdata.Thereisahugeopportunityforbenefitstopatientsandsociety.”

Apre-requisitetorealisingthesebenefitsistocreatetherightframeworkfordatasharinganddataaccesstoenableresearch.Datacollectionanddataaccesscanbeforanumberofdifferentpurposes.Giventhis,thereisaneedforharmonisationofdataandharmonisationofpatientrecords.“Dataneedstobeintherightformat,”Ralfconcluded.

ErnstHafenoftheInstituteofMolecularSystemsBiology,ETHZurich

1 21 6 T H E U R O P E A N H E A L T H F O R U M G A S T E I N

EHFG.ORG