The Basics of VHA Data - Stanford Medicine · 3/12/18 6 •Additional documents that may be...
Transcript of The Basics of VHA Data - Stanford Medicine · 3/12/18 6 •Additional documents that may be...
-
3/12/18
1
TheBasicsofVHADataAccessing,Requesting,and
AnalyzingLauraA.Graham,PhD,MPH
HealthServicesResearchFellow,Ci2i,VAPaloAltoPostdoctoralFellow,SPIRE,StanfordUniversity
PriorSteps
• Beforerequestingdataaccess…
• Writeaclearprotocol• Knowyourstudyteam• GetVAAccess(Employed,WOC’ed)• GetIRBApproval!
ResearchIdea
Protocol
Funding?
IRBApplication
DUAs/DTAs
BeginResearch
-
3/12/18
2
SeekingInformationonVAData?
• VHADataPortal• http://vaww.vhadataportal.med.va.gov/• GreatresourceforVHAdata
• VHADataArchitectureRepository• https://vaausdarmul81.aac.dva.va.gov/• DetaileddatadictionaryforCDWtables• SearchstringcanbeusefulforfindingdataburiedinCDWtables
SeekingInformationonVAData?
• VAInformaticsandComputingInfrastructure(VINCI)• https://vaww.vinci.med.va.gov/vincicentral/
• VAInformationResourceCenter(VIReC)• http://vaww.virec.research.va.gov/
• VAHealthEconomicsResearchCenter(HERC)• http://vaww.herc.research.va.gov/
• OfficeofPerformanceMeasurement(OPM)• https://vaww.car.rtp.med.va.gov/
• PlanningSystemSupportGroup(PSSG)• http://vaww.pssg.med.va.gov/
• Data.gov/CorporateDatabasesMonograph• www.data.gov
-
3/12/18
3
ReadilyAvailableDataSources
• VHASupportServiceCenter• https://vssc.med.va.gov/• Facility-levelinformation
• BusinessOperations• ClinicalCare• Quality&Performance
• NationalCenterforVeteransAnalysisandStatistics• https://www.va.gov/vetdata/
DART
DataAccessRequestTracker
-
3/12/18
4
RequestingVHADataAccess
• MostofyourdataaccessrequestswillbecoveredbytheDataAccessRequestTracker(DART)• https://dart.vha.med.va.gov/
• OtherdatasourcesexistandtheirrequestprocessescanbefoundintheVHADataPortal• http://vaww.vhadataportal.med.va.gov/
DataAccessRequestTracker(DART)
• Onlineworkflowapplication• DatarequestsandPreptoResearchRequests
-
3/12/18
5
DataAccessRequestTracker(DART)
• DARTisyourstartingplaceforVINCIdataaccessamongothersources• AcquiresNationalDataService(NDS)approval• OnceapprovedanassignedVHAdatamanagerwillprovideaVINCIWorkspaceandaccesstodata• WillhelpyounavigateVINCI
VINCIhelpswithorangeboxes
• https://dart.vha.med.va.gov/
• CompleteanduploadtherequireddocumentsaslistedinDARTforthedatasource(s)ordatatool(s)selected.• Someofthefollowingdocumentswillberequired:
• ResearchProjectDocumentsandApprovals• Researchstudyprotocol• ResearchandDevelopmentCommitteeApprovalLetter(s)• InstitutionalReviewBoard(IRB)ApprovalLetter(s)• IRBApprovedHIPAAAuthorizationorWaiverofHIPAAAuthorization• IRBApprovedSampleInformedConsentorWaiverofInformedConsent
DataAccessRequestTracker(DART)
-
3/12/18
6
• Additionaldocumentsthatmayberequired:• DataRequestForms
• ResearchRequestMemo – Replacesthe9957forALL researchrequestssubmittedthroughDART afterMarch25,2015
• RealSSNAccessRequest – RequiredforaccesstorealSSNs• CDWDomainCheckList – CDWdataonly• VitalStatusRulesofBehavior – VitalStatusFileonly• SpecialUserAccessRequestFormforResearchers – CAPRIandVistAWeb only• SurgeryDataResearchProposalTemplate – SurgicalQualityDataUseGroup(SQDUG)dataonly
• NationalSurgeryOffice(NSO)DataUseAgreementandSecurityInformation – SQDUGdataonly
DataAccessRequestTracker(DART)
DataAccessRequestTracker(DART)
-
3/12/18
7
• Onceallrequireddocumentsareuploaded andtherequesthasbeensubmitted,itwillundergoaprivacyreview.
• Additionalreviewsarerequiredforthefollowing:• RequestsfordatawithrealSSNswillundergoreviewbytheOfficeofResearch&Development.
• RequestswillundergoasecurityreviewusingtheinformationprovidedintheResearchRequestMemo.
DataAccessRequestTracker(DART)
DataAccessRequestTracker(DART)
• WhendoIneedtorequestRealSSNs?
• You’replanningtousechartabstraction• You’relinkingtoanotherexistingdatasourcebySSN• You’replanningtousetheTIU(TextIntegrationUtilities)
• EverythingelseislinkedbyeitherSta3n/PatientSID orSCRSSN andyouwillreceiveacrosswalkfileonceyourrequestisapproved
-
3/12/18
8
DataAccessRequestTracker(DART)
• AfterdataaccessapprovalaVINCIDataManagercanhelpwiththefollowing:• Cohortselection• DataExtraction• Formatting,Indexing,andSecurity
OtherDataAccessExamples
• OfficeofReportingAnalyticsPerformanceImprovementandDeployment(RAPID)• DUAForms
• ProjectInformationSheet• PersonalAgreementStatement• DataAccessList
• StudyProtocol• IRBApproval• DataRequestSpecifications
-
3/12/18
9
VINCITheVAInformaticsandComputingInfrastructure
VINCIIntroduction
• Web-basedplatformforaccessingandanalyzingdatawithavarietyoftools• ProvidesconsultationservicesforIRBResearchfortheentirelife-cycleofaresearchproject• https://vaww.vinci.med.va.gov/vincicentral/
-
3/12/18
10
VINCIIntroduction
• https://vaww.vinci.med.va.gov/vincicentral/
• 105high-performanceserversand1.5petabytesofdatastorage• 450+newprojects– 17%increaseYOY• Provideddatato840+projects• VINCIDataAnalyststouchabout50projects/week• Structure
• Windows2012R2Workspace• SQLServers(Currently3)
• RB01:vhacdwdbs01.vha.med.va.gov• RB02:vhacdwdbs02.vha.med.va.gov• RB03:vhacdwdbs03.vha.med.va.gov
VINCIUse
-
3/12/18
11
VINCIUserGuides
-
3/12/18
12
VINCIStandardWorkspaceApplications
-
3/12/18
13
VINCIStandardWorkspaceApplications
• IconsontheStandardWorkspacelinktoremotedesktopservers• Yourlog-inandpasswordarethesameasforyourVAcomputer(i.e.vhapal …)• Theoveralldesktopapplicationhaslimitedsoftwareavailability.• Logintoeachapplicationseparately• Eachwillbeaseparatedesktop/server
• TheSAS9.xiconisagoodplacetostart.
VINCIFolderSetUp
• Sameasonadesktop• HardDiskDrives(K,L,M,N,T,S,U,V,X)• NetworkDrives
• P:/drive(Projects)• O:/drive(Projects2)• H:/drive
• AllprojectfoldersaretypicallyfoundintheP:/drive• H:/isyourpersonalfolder• Allprojectfoldershaveasub-foldernamedUploads whereuploadedinformationisincluded
-
3/12/18
14
VINCIFolderSetUp
VINCI– MoreInformation
• MoreinformationonaccessingandusingtheVINCIWorkspacearefoundintheVINCIWorkspaceUserGuide
-
3/12/18
15
SQLServerManagementStudio
ViewYourData
SQLServerManagementStudio
• 2017version• MethodforaccessingandviewingCDWdomaintableswithSQLcode• Canalsobeusedtopulldataina.txtformatifneeded• RequiressomeinitialsetuptolinktoserversandsomeSQLknowledgebutalsohaspoint-and-clickoptions thatmakeiteasytouse
• Thebestplacetostartlookingatdatayou’vebeengivenaccesstoforyourstudy
-
3/12/18
16
SQLServerManagementStudioSetUp• VINCIDatabaseUserGuide• Section4.2.1(pages3-4)providesdetailedinformationonSQLserversetup
DefaultSettingsToUse
ServerType: DatabaseEngine
ServerName: .vha.med.va.gov
Authentication: WindowsAuthentication
SQLServerManagementStudioDataAccess• VINCIDatabaseUserGuide• Section4.2.2(pages4-7)providesdetailedinformationonaccessingdatabases
• Overview• Databases>CDWWork containsyourdimensiontables• Databases>ORD_XXXX>Views containsyourSQLdatatables• Right-clickand“SelectTop1000Rows”foreasyviewingofadatasource
-
3/12/18
17
SQLServerManagementStudioDataPulls• NotasefficientasaSASpullbutcanbedone
• First,setupyourquery.YourinitialrunswillbeoutputtoGrid.
• Onceyourqueryissetup• Use“Query>ResultsTo”tochangeto“ResultstoFile”• Then“Query>QueryOptions…”tosetupyourfilebutnavigatingthrough“Results>Text”
ApplicationsforAnalysis
AnalyzeYourData
-
3/12/18
18
VINCIStandardWorkspaceApplications
VINCISAS
• LotsofUserGuidesforreference• Thiscourseisonlyanintrobutthesewillansweryourmoreadvancedquestions.
-
3/12/18
19
VINCISAS
• SeveralversionsofSASareavailableforuse.• SAS9.2• SAS9.4• SASEnterpriseGuide7.1
• Allofferdifferentcapabilitiessocheckwithallofthem
VINCISAS
• “Grid”referstotheLinuxoperatingsystemwhereSASEnterpriseGuide7.1 runs• Youwillneedaslightlydifferentlibname reference:“/data/dart/2014/ORD_Smith_201401001D/“
• YouwillalsoneedtouseWINSCPfordatatransferandaPIV-exemptiontoaccesstheGrid
• Checkout“Grid– WheretoBegin?”formoreinformation• SAS9.2and9.4 arenotonthegridandsometimeseasiertoworkfrom.• Plus,SAS9.2istheonlyprogramwithPROCTRAJ!
-
3/12/18
20
VINCISASDataPulls
• ItendtouseSAS9.2orSAS9.4totunnelintotheCDWandpullmydatadirectlyintoSAS.
• Thisrequiresthefollowingcode:
• TheCatalogischangedtoreferenceyourprojectfolder.• “Datasource”mayalsoneedtobechangedbasedonyourprojectserver.• Schemawillchangebasedonwhatviewyouarelookingat(SRC,Dim,…)
VINCISASDataPulls
• PROCSQListypicallyhowIpullandmergethedatainSAS9.4• IinitiallysetupthetestdatapullsinSQLManagementStudiousingapullofthetop100 totestthequeries• OncemySQLcodelooksIgoodthenImoveittoSAS,removethebrackets,updatethelibnames andrun.
-
3/12/18
21
VINCISASDataPulls
• ThiscangetcomplicatedwithmultiplejoinsandWherestatements.
VINCISASDataPulls
• SQLpass-throughlanguageisalsoaquickwayoftunnelingintotheSQLtablesandpullingoutdata.
-
3/12/18
22
VINCIEnterpriseGuide7.1
• VINCISAS9.4and9.2arelimitedinthetypesofanalysesthatcanbedone
• VINCIEnterpriseGuideofferseverythingandisbyfarthemostcapableoftheSASproductsbutyouhavetosetuplinkstotheGridandmoveyourfilestotheLinuxsystem
• Itendtodomostofmydatamanagementandcleaningin9.4andthenswitchtheanalyticdatasettoLinuxformyanalyses
VINCIEnterpriseGuide7.1
• ThefirststeptousingEnterpriseGuide7.1istocheckyourLinuxsystemregistrationbysettingupWinSCP• OncesetupWinSCP isyourmethodformovingfilesfromPCtoLinux• ThiscanbedonebyfollowingtheinstructionsinUsingEnterpriseGuidewiththeGrid
-
3/12/18
23
WinSCP
VINCIEnterpriseGuide7.1
• TouseSASEnterpriseGuide7.1youwillneedtosetupyourconnection
• ThiscanbedonebyfollowingtheinstructionsinUsingEnterpriseGuidewiththeGrid
-
3/12/18
24
Policies
ProtectYourData
PoliciesforElectronicData
• VAHandbook6500• DefinitionsandguidancefortheVHA
• KeyPoints• USBdrivesareforbidden• DonotstoresensitiveinformationonaHardDrive(C:/Drive).Trynottostoreanythingonaharddrive.
• Sensitiveinformationcanneverbetransported,accessed,processedorotherwiseusedoffsite.DON’TEMAILDATA!
-
3/12/18
25
TypesofData
1. IndividuallyIdentifiableHealthInformation2. ProtectedHealthInformation(PHI)3. PersonallyIdentifiableInformation(PII)4. VASensitiveInformation/Data
• VHAHandbook6500 providesguidanceondefinitionsandmanagement
DisposalofData
• GovernedbytheVHARecordControlSchedule (RCS10-1)
• Currentguidance(Item8300.6,PageIII-8-6)• Temporary;cutoffattheendofthefiscalyearaftercompletionoftheresearchproject.• Destroy6yearsaftercutoff,mayretainlongerifrequiredbyotherFederalregulations.(DAA-0015-2015-0004,item0032)
• IftheinvestigatorleavesVA,allresearchrecordsareretainedbytheVAfacilitywheretheresearchwasconducted.
• IfthegrantisongoingandtheinvestigatorleavesoneVAfacilitytogotoanotherVAfacility,theinvestigatormustobtainapprovalforacopyofrelevantmaterialstobeprovidedtothenewVAfacility'sresearchoffice.
• Theinvestigatorisnotthegrantee,nordoestheinvestigatorownthedata.
-
3/12/18
26
FunThingstoDo
FullOrganizationalView
• >9MillionEnrolledPatients
• 170VAMedicalCenters(141withSurgicalCapacity)
• 1,063OutpatientCenters
-
3/12/18
27
LongitudinalData
• AllvitalsandlaboratoryvaluescollectedintheVAareeasilyaccessible
• Pharmacyprescriptions–inpatient,outpatient,andevenevidenceofnon-VAprescriptionsareavailable
JAMA Surg. 2014;149(11):1113-1120. doi:10.1001/jamasurg.2014.2044
NaturalLanguageProcessingCapabilities
• Linguistic-computerscienceresearch• Abilityofacomputerprogramtounderstandhumanlanguageasitiswritten infree-textofclinicalnotes• SeveralVAtoolstoworkwith
-
3/12/18
28
Thankyou!