CDISC SDTM IG for Associated Persons 1.0 Overview March 2016 – DC CDISC user group - PowerPoint Presentation

CDISC SDTM IG for Associated Persons 1.0 Overview March 2016 – DC CDISC user group Michael DiGiantomasso -

1 Foreword Data about people …. not in the study …. that could affect study subjects … or who have been accidentally exposed There are hardly any new concepts. The key is separation to avoid ambiguity and confusion

Conformance Rules CDISC SDTM IG 3.2 2 Contents Overview Document Keys History Conclusion Goals Questions Examples Specification table Single-Domain Special Purpose Interventions Events Findings Multi-Domain Purpose Prerequisites Definition Intention Data not applicable Use Cases Core Concepts Data Separation Domains Variables Identifiers Relationships

3 Document Keys Key Value Version 1.0 Publication Date 2013-12-12 Public Review Ending n/a Draft Status FINAL Doc Organization CDISC SDTM Principal Contributors Fred Wood Shannon Labout Jan Hess Gary Walker Carey Smoak Joyce Hernandez Diane Wold CDISC SDS Team Other AP members & Contributors Sharmila Chaudhuri Jennifer Roessler Donna Sattler Darcy Wold FDA Contributor(s) Lynn Henley, CDRH

4 History Publish / PR Date SDTM Model Ver. 2004-06-25 1.0 2005-04-28 1.1 2008-11-12 1.2 2012-07-16 1.3 2013-11-26 1.4 2015-08-24 1.5 Ver. Standard Publish / PR Date Corresponds to 3.1 SDTM IG 2004-07-14 3.1.1 2005-08-26 3.1.2 2008-11-12 3.1.3 2012-07-16 3.2 2013-11-26 1.0 SDTM IG MD 2012-12-04 1.0 SDTM IG AP 2013-12-12 SDTM 1.4 1.0 SDTM IG PGx 2015-06-01 3.0 SEND IG 2011-05-19 3.12015-09-101.0SEND DART2015-09-10 AP is already 2+ years old

5 Prerequisites CDISC states the order of reading and understanding as follows… SDTM – for building block concepts SDTM IGs SDTM IG 3.x (Demographics data, etc..) SDTM IG MD (medical devices, when applicable) SDTM IG PGx (pharmacogenetics /genomics data, when applicable)

6 Definition An AP classifies the non-subject participants in a clinical study, and the data of interest about them. non-subject participants may be associated with the study itself (when no clear relationship exists) a study devicea study subject family member caregiver organ donor

7 Data Not Applicable Domains which describe a subject’s progress through a study ( SE, SV, DS) are not allowed for associated persons since they…. are not in the study (screened nor randomized) do not have their own visits do not participate in epochs are not dispositioned

8 Basic Intention Guide the organization & format of data about non-study subjects. Describe an implementation method based on SDTM Introduce SDTM variables for use in AP data. Provide use cases for implementers. Prepare for regulatory submission

9 Technical Intention data collected about APs may be important to understand & analyze study data follow an AP for potential adverse reactions AP-IG is based on the assumptions that it may be necessary … to distinguish Associated Persons (AP) data from subject data to physically keep AP and study subject data separate in a data submission data warehouse It is unknown how a user might query for data in a data warehouse… the mechanisms for keeping AP data from being confused with subject data will be described in “Core Concepts” slides

10 Use Cases CDISC Disclaimer : “The use cases listed below have come to the attention of the SDS Team. It is probably not an exhaustive list.” Family history: data about a subject’s family members Donor info : original owners of donated organs, blood, tissues, etc. Caretakers : a questionnaire administered to the caretaker of a study subject about their experience as a caretakerSexual partners: demographics, sexual and/or pregnancy history of the study subject’s sexual partnerEnvironmental exposure: the smoking habits of a person who lives in the same household as a study subjectDiagnostic samples: when study topic is an investigational device,…the tissue or blood sample dasta used in a diagnostic test can be an APAccidental study treatment exposure: someone (not enrolled) is exposed to a study treatment (e.g., estrogen cream, radioactive isotope),…their exposure (APEX) and adverse event data (APAE) is neededThere may not be a study subject to whom the associated person is related. Operator accidents: adverse event data (APAE) about research staff who are injured while using an investigational device, especially when the device is the topic of the studyThere may not be a study subject to whom the associated person is related.Source / Contact Case Investigation (highlighted in TAUG TB 2.0) – not mentioned in AP IG

11 TA Use Cases Risk and disease transmission for Infectious Disease (S/C Case) TAUG TB ( Tuberculosis) 2.0 TAUG EVD (Ebola Virus Disease) 1.0 the non-study subject from whom the study subject likely contracted the disease. the non-study subject who may have contracted the disease from a study subject . Family HistoryTAUG Dyslipidemia 1.0Cancer StudiesPolycystic Kidney Disease (PKD)Immunosuppressive Therapy / Solid Organ TransplantsAny trials collecting data about organ donors Any study that solicit pregnant women, since by default their offspring would be an associated person

12 Core Concepts Data Separation Domains Variables Identifiers Relationships

13 Data Separation datasets are given a prefix of AP-- to distinguish them from study subject data. AP domain codes belong in the DOMAIN columnprefixed with AP 4 characters long . Domain label names begin with “Associated Persons” records require the population of the APID variable . That’s it…

14 Domains AP domains parallel … general observation class Demographics (DM) Comments (CO) potentially AP domain structure is same as those for study subject domains Exception: variables that only apply to study subjects are prohibited (USUBJID, SPDEVID) all other general assumptions about SDTM and SDTMIG variables and domains will apply to AP data Domain naming conventions by example AP DM - demographic data collected about an associated person APLB - lab data about an associated person APMH - medical history about an associated person APFASU - findings about the substance use of an associated personSQ AP DM – supplemental qualifiers for associated person demographicsSQAPFAMH – supplemental qualifiers for findings about medical history of an associated person AP Variable prefixes will NOT include AP There is one special purpose domain APRELSUB which will be discussed in the “Relationships” section

15 Identifier and Qualifier Variables VARIABLE Core Description STUDYID Required Study Id DOMAIN Required Domain Code APID Required Associated Person ID --SEQ Required Sequence Number RSUBJID Expected Related Subject   RDEVID Permissible Related Device SREL Required Relationship If RSUBJID is populated, the relationship of the AP to Subject If RDEVID is populated, the relationship of the AP to Device IF RSUBJID & RDEVID are null, the AP relationship is to the StudySDTM IG 3.2 Conformance Rules include these

16 Timing Variables All general date variables are allowed. Study-based timing variables can be used to place an AP record in the study context for the related study-subject not appropriate if the AP is associated only with the study Non study-subject device If an AP has relationships with multiple study subjects, study-based timing variables may be ambiguous: use with caution

17 Variables not to be used RF STDTC – Subject Reference Start date RF ENDTC – Subject Reference End date RFX STDTC – First study treatment date RFX ENDTC – Last study treatment date RFI CDTC - Informed Consent date * RFP ENDTC – End of Participation date ARM CD - Planned Arm Code ARM - Planned Arm ACT ARMCD – Actual Arm code ACTARM - Actual Arm 4 sets of Demographics variables that only apply to study subjects * Informed Consent only applies to being in the study. An associated person is not in the study, so consent in that context does not apply. Naturally, if a person is identified as an AP after they have been accidentally exposed, or for which the protocol wants to collect information for other reasons after the study started, then the concept of “consent” probably occurs in that context. Post study informed consent would be in SQAPDM ( suppqual for associated person demographics). With regard to the information for a donor, they most likely already gave consent or license to use their organ(s), and in doing so provided any information needed that would be part of an AP domain for a donor.

18 AP Identifiers not always collected and often must be generated by sponsors. ~ artificial process to allow APs to relate to study subjects outside the CRFsunique items (subject IDs, domain codes, visit #) can assist in the creation CRFs often collect data across domains. Sponsors should be careful to preserve known data relationships avoid an appearance of a relationship where none exists or is unknown. sponsors who already have a method of identifying APs…may need to assign new APIDs for submission (APIDs are not required to be unique outside of a study)If uniqueness is intendted, that should be indicated in the SDRG3 Points to consider: same APID should always identify the same associated person. allows sponsors to maintain relationships between data split into different domainsallows reviewers to see that the relationship existsdistinct APIDs should be used when there is no known relationship AP Identifies a single unit: (a single person OR a group/pool of people)Differentiation between single person vs. group should be apparent. APID naming conventions will aid reviewers when data contains a mix of single and group APs

19 Relationships It’s all about

20 CDISC Codelist for Relationship CDISC/NCI provides an Extensible vocabularyC100130 – Relationship to Subject ( RELSUB) 54 terms 4 categories FamilialSpouse/Sexual Partner/Friend Health Care Providers Donors http://evs.nci.nih.gov/ftp1/CDISC/SDTM/ AP data may be collected because they bear some relation to an investigational device (i.e. SAMPLE DONOR) If AP is related to a subject, the RELSUB codelist should be consulted aboveIf AP is related to an investigational device, other terminology may be needed. CDISC does not provide

21 CDISC RELSUB Codelist Items C41256 FAMILY MEMBER C21481 RELATIVE , FIRST DEGREE C19811 RELATIVE , SECOND DEGREE C100808 RELATIVE , THIRD DEGREE C96580 MOTHER , BIOLOGICAL C96572 FATHER , BIOLOGICALC100807CHILD, BIOLOGICALC100805 GRANDCHILD, BIOLOGICAL C73427 TWIN C73428 TWIN , DIZYGOTIC C73429TWIN, MONOZYGOTICC96571COUSIN, BIOLOGICAL C96576 COUSIN, BIOLOGICAL MATERNAL C96582 COUSIN , BIOLOGICAL PATERNAL C100806GRANDPARENT, BIOLOGICALC111248 GRANDPARENT, BIOLOGICAL MATERNAL C111286 GRANDPARENT , BIOLOGICAL PATERNAL C96573 GRANDFATHER , BIOLOGICALC96577GRANDFATHER, BIOLOGICAL MATERNAL C96583 GRANDFATHER , BIOLOGICAL PATERNAL C96574 GRANDMOTHER , BIOLOGICAL C96578 GRANDMOTHER , BIOLOGICAL MATERNAL C96584GRANDMOTHER, BIOLOGICAL PATERNALC96569AUNT, BIOLOGICALC96575AUNT, BIOLOGICAL MATERNALC96581AUNT, BIOLOGICAL PATERNAL C96587 UNCLE , BIOLOGICAL C96579 UNCLE, BIOLOGICAL MATERNALC96585UNCLE, BIOLOGICAL PATERNALC100809SIBLING, BIOLOGICAL    C71390SIBLING, FULLC71391SIBLING, HALFC96570BROTHER, BIOLOGICALC96656BROTHER, BIOLOGICAL MATERNAL HALFC96655BROTHER, BIOLOGICAL PATERNAL HALFC111201BROTHER, FULLC71402BROTHER, HALFC96586SISTER, BIOLOGICALC96658SISTER, BIOLOGICAL MATERNAL HALFC96657SISTER, BIOLOGICAL PATERNAL HALFC111202SISTER, FULLC71403SISTER, HALFC62649SPOUSEC53262DOMESTIC PARTNERC100812SEXUAL PARTNERC100811PREGNANCY PARTNERC72884FRIENDC48284DONORC100810DONOR, ORGANC25168DONOR, TISSUEC63861TECHNICIANC85499CLINICIANC17445CAREGIVERC20821NURSEMany terms for targeted specificity

22 Relationship Rationale The relationship to a study subject is often the reason for collecting data about that AP and is reflrected in SREL variable. Multiple relationships may be collected in additional values of SREL, The most relevant relationship should be first.. Example: a child with a debilitating disease is a subject and the associate person is the Aunt and Caregiver Study 1 is concerned with paternal relatives’ family history only AP is an AUNT , BIOLOGICAL PATERNAL who is also a CAREGIVER Study 2 is concerned with Affect of disease on subjects' caregivers' QoLAP is a CAREGIVER who is also a RELATIVE.APEX and APAE collection reason is implicitAPEX - study treatment exposure, most likely by accident. APAE - adverse events, most likely from exposure to study treatment or an accident involving an investigational devicerelationship should explain how the accidental exposure occurred. If no relationship exists, default is “ACCIDENTAL ASSOCIATION” with the study, (i.e. if study treatment was delivered to the wrong hospital patient through a clerical error)Relationship vocabulary can overlap with evaluators. same values can appear in the variables SREL and EVAL. unless data are collected about the evaluator, that person is not an AP. (don’t get overzealous about using AP)Ex: a caregiver provides evaluations of a subject <> not an AP data about the caregiver was collected, such as a questionnaire assessing the caregiver’s quality of life.

23 Multiple Relationships and APRELSUB A relationship must exist for a non-subject to be an associated person. Single Relationships fairly common and are easy to comprehend does not require APRELSUB domain relationship is specified explicitly in the AP domain using SREL Multiple Relationships 1 Associated person ... having 1 to many relations to 1 to many Subjects/Devices 1 AP to many relations: SREL = MULTIPLE 1 AP to many subjects: RSUBJID = MULTIPLEAPRELSUB must be used to record 1 record per relationship and has 4 variables of interestAPID – Associated PersonSREL - RelationshipRSUBJID – Related SubjectRDEVID – Related DeviceAPRELSUB is similar to the RDF model that uses Subject-Predicate-Object triples:SUBJECT PREDICATE OBJECT APID SREL RSUBJID AP-007 DONOR, ORGAN AB-01 AP-007 DONOR, ORGAN XY-02 James Is organ donor to Mike James Is organ donor to Paul

24 APRELSUB assumptions Allowed relationships associated person  study subject associated person  study device Prohibited relationships s tudy subject  associated person s tudy device  associated personassociated person  associated person

25 APRELSUB Examples Ex 1 : two study subjects received donated organs from same AP, who was biologically related to one of them. critical for immunosuppressive therapies for solid organ transplants STUDYID APID SREL RSUBJID NVM_0896 NS51 DONOR, ORGAN NVM10051 NVM_0896 NS51 DONOR, ORGAN NVM10082 NVM_0896 NS51 RELATIVE, BIOLOGICAL NVM10082 Ex 2 : include all APs even though only one has multiple relationships to the subject STUDYID APID SREL RSUBJID ZY_098 A0101 HOUSEHOLD MEMBER ZY_098_01 ZY_098 A0101 SIBLING, FULL ZY_098_01 ZY_098 A0102 HOUSEHOLD MEMBER ZY_098_01 ZY_098 A0103 HOUSEHOLD MEMBER ZY_098_01 ZY_098 A0104 HOUSEHOLD MEMBER ZY_098_01 Ex 3: caregivers and family members. The sponsor chose to include only those APs with multiple relationships STUDYIDAPIDSRELRSUBJIDVIIAP005CAREGIVERVII_02VIIAP005AUNT, BIOLOGICAL PATERNALVII_02VIIAP006MOTHERVII_02VIIAP006COUSIN, BIOLOGICAL PATERNALVII_03VII AP027 CAREGIVER VII_07 VII AP027 AUNT, BIOLOGICAL MATERNAL VII_07 VII AP027 MOTHER, STEP VII_07 VII AP030 TWIN, DIZYGOTIC VII_07 VII AP030 BROTHER, FULL VII_07

26 AP Domain Examples Specification table Single-Domain Special Purpose (DM) Interventions (EX,SU) Events (AE, MH x 2, ) Findings (8: [LB,QS,RP,SC] x 2 Multi-Domain (2) AP domains are based on domains from other IG therefore documentation is generally limited to examples . The domains in this IG do not include all possible AP domains, those involved in the use cases which came to the attention of the SDS Team and led to the development of this IG. Additional domains may be included in future versions

27 Example: AP Specification Table

28 Example: Single Domain – Special Purpose APDMAssociated Persons DemographicsNot required Only submit if data is collected about APs If POOLDEF records exist and demographics about the pool is collected, otherwise submit for each individual in pool Example: CRF collects organ donor demographics D456 = donor TRS_0520DS_056 = Recipient study subject

29 Example: Single Domain – Interventions APEXAssociated Persons Exposure Assumptions accidental exposures to study treatment If an AP is related to a study subject, that relationship should be captured If no relation between an AP and any study subject, SREL = ACCIDENTAL ASSOCIATION Example : study treatment was inadvertently dispensed to someone other than the intended study subject

30 Example: Single Domain – Interventions APSUAssociated Persons Substance Use family history: CRF asks about the subject’s mother’s drinking and smoking habits document environment exposure: second hand smoke CRF asks about smoking habits of up to five household members. Subject lives with only 3 people x x

31 Example: Single Domain – Events: APAE Associated Persons Adverse Events - CRF collects data about a Device Operator (DEV_2011_OP04) with no related study subject X X 05 04 2008 x Operator exposed to radioactive contrast agent X X CT Scan

32 Example: Single Domain – Events: APMH Associated Persons Medical History study subject 2011-02-02-031 has family members who have been diagnosed with Pompe Disease The CRF collects data about both single-person APIDs ( mother, father) – naming conventions with “N” group APIDs (siblings, cousins ) – naming conventions with “NS” 2

33 Example: Single Domain – Findings: APLB Associated Persons Laboratory Test Results CRF is collecting lab information about an organ donor D456 = donor TRS_0520DS_056 = Study Subject Example: FDA submission of Viral Serologies and other related data for donors in a Tacrolomius study – extremely complicated and hard to understand without AP

34 Other Examples for Single Findings Domain Associated Persons Questionnaires (APQS)Caregiver Quality of Life (CGQOL) questionnaire that the caregiver of a study subject completes Associated Persons Reproductive Systems (APRP)find out if the subject’s female partner became pregnant during the studyAssociated Persons Subject Characteristics (APSC)CRF is collecting blood-group information about a donor

35 Multi-Domain Example 1 Accidental exposure of a study treatment to non-study subject Exposure triggers and adverse event

36 Multi-Domain Example 2CRF collects data about a subject’s family history relating to Polycystic Kidney Disease (PKD). data collected about them falls into 6 domains AP MH - Medical History AP DM – Demographics AP SC – Characteristics AP PR - ProcedureAPSS - Survival StatusAPDD - Death DetailsAPRELSUB - AP to Subject Relation

Source Case   Infected with TB   Contact Case TB infected Associated Person   Study Subject   Healthy Associated Person 37 TB Exposure and Risk Factors Two Types of investigation: Source case : identify from whom the subject contracted TB Contact case: identify those who came in contact with a subject diagnosed with TB Passive – AP self refers / presents them self Active – Investigator looks for contacts Evaluate risk of exposure Evaluate TB strain characteristics Associated Persons Domains are essential for Infectious Disease studies. The examples in the TAUG could be applied to any study requiring source and/or contact cases

38 Source Case Investigation DOMAIN USUBJID ERLNKID ERTERM ERCAT ERSTDTC ERENDTC   SETTING ER Mary 1 Exposure to TB SOURCE CASE INVESTIGATION 2012-08-06 2012-08-10   Child Care Center STUDYID APID SREL RSUBJID TB-001 SC_Teacher EDUCATOR Mary A pediatric study subject, Mary, has been enrolled and exposed to TB. Mary, a child, has difficulty producing sputum… makes bacteriologic confirmation of infection difficult. it is necessary to identify the source case (adult) capable of producing the sputum sample. After identification, investigators can collect and test a sputum sample to learn about the TB strain suspected APID SREL RSUBJID MBCAT MBTSTDTL MBTEST MBLOC MBSPEC MBMETHOD MBORRES APMB SC_TeacherEDUCATORMarySOURCE CASE INVESTIGATIONIdentificationMycobacterium TuberculosisLUNGSPUTUMMICROBIAL CULTURE, SOLIDPRESENTMicrobiology susceptibility tests (MS) can be run on the sample to determine drug resistance, if any

39 Contact Case Investigation Investigate an associated person (not in the study) who may have come into contact with the infected study subject. E.g. a mother with TB is enrolled, her child would be an associate person for whom a contact case is initiated DOMAIN APID SREL RSUBJID ER TERM ER CAT ER STDTC ER ENRTPT ER ENTPT AP ER Sue CHILD, BIOLOGICAL Kate Exposure to TB CONTACT CASE INVESTIGATION 2012-08-06 ONGOING 2012-11-05 CASEFIND  SETTING PRIORITY  ACTIVE  HOME  HIGH-PRIORITY TUBERCULOSIS CONTACT  NSVs

40 Rule Development for CDISC Standards CDISC has published a draft version of SDTM IG 3.2 Conformance Rules 8 of the 416 rules (1.9%) are linked to AP IG CDISC ID Programmable Variable Condition Rule CG0155 Y DOMAIN AP Core Domain Value length = 4 characters beginning with 'AP' and ending with 2 character SDTM domain CG0156 Y APID APID identifies group of associated persons defined in POOLDEF APID in POOLDEF.POOLID CG0157 Y RSUBJID Relationship described is to a study subject RSUBJID in DM.USUBJID CG0158 Y RSUBJID Relationship described is to a pool RSUBJID in POOLDEF.POOLID CG0159 Y RSUBJID Relationship described is to a study but not to individual study subjects RSUBJID = null CG0160 N SREL RSUBJID ^= null Describes relationship of associated person to study subject or pool CG0161 N SREL RDEVID ^= null Describes relationship of associated person to device identified in RDEVID CG0162 N SREL RSUBJID = null and RDEVID = null Describes relationship of associated person to sudy identified in STUDYID

41 Conclusions AP IG mainly adds a few new variables to support associated persons in relation to any other domain It’s the same domains and concepts, with another prefix, that segregates data for non study subjects Having SDTM-IG AP 1.0 is a positive step in the evolution of the standard. Not having it will only allow for interpretation and confusion in industry between sponsor, CRO and regulatory agencies .

42 Nice to have Industry Goals (other than to start using it) Standardization on these concepts: Use of study-based timing variables If an AP has relationships with multiple study subjects APID naming conventions for data with a mix of both.. single APs group Aps use of APRELSUB to …only include APs with multiple relationships or always to use as a single summary listing for all Aps TA mandates/requirements for AP domains such as ..any studies involving solid organ transplants Infectious disease studies such as TB and HIV for identifying source and contact case investigations Hereditary diseases where family history is heavily involved.

43 Appendix and Extras

44 About the Presenter Mike is an Ursinus College 1999 Grad with a B.S. in Computer Science and Mathematics and a Founding Partner and Technical lead at Pinnacle 21 , LLC.  He has 20 years experience in the IT field of which the last 17 have been supporting all stages of clinical development.    Mike currently works as a   Jumpstart and D ata Fitness analyst at the FDA, while helping them adopt new and existing CDISC standards to the FDA catalog.    He also developed the clinical  trials.gov  miner tool and ADaM rules for Pinnacle 21 Community while participating on the CDISC ADaM Validation Subteam. In the past he was the Data Architect for the FDA Janus CTR  and served at Merck for 11 years as a developer and business analyst. mike.digian@pinnacle21.netmike@opencdisc.orghttp://www.pinnacle21.net/#about http://wiki.cdisc.org/display/~mike