Associated Persons Domains Who, What, When, Where, Why, and How? Michael Stackhouse Alyssa Wittle PharmaSUG SDE RTP 2016 October 27, 2016
Associated Persons Domains
Who, What, When, Where, Why, and How?
Michael Stackhouse
Alyssa Wittle
PharmaSUG SDE RTP 2016October 27, 2016
What and Why?
2 Proprietary & Confidential. © 2016 Chiltern
● SDTM Associated Persons Implementation Guide
• Standardized place for non-subject specific data
• This data is not about the study subject, so it has no home in a subject level SDTM
• Intended for human data
● Family Medical History
● Family Substance Abuse History
● Accidental Exposure (non-subject)
● AEs belonging to a non-subject due to
accidental exposure or device malfunction
When?
● Donor Information
● Caretakers
● Sexual Partner Information
● Persons in the Subject’s Environment
3 Proprietary & Confidential. © 2016 Chiltern
● When do Associated Persons domains apply to a study?
Who?
4 Proprietary & Confidential. © 2016 Chiltern
● For whom are these domains intended?
● Who is an Associated Person (AP)?
• The person can be associated with the subject or the study
• Data needs to be about the AP
Where?
5 Proprietary & Confidential. © 2016 Chiltern
● Where do I put all of the information?
• An AP domain can be created based on traditional SDTM domain
• Prefixes of “AP” are added to the SDTM domain name to identify them as Associated Person
• Domains are only used if you have that information about an AP
- i.e. APDM is not required in order to have APMH
• Supplemental Qualifiers for an AP dataset start with a prefix of “SQAP” rather than “SUPP”
- Example: SQAPSU, SQAPFAMH
How?! – General
6 Proprietary & Confidential. © 2016 Chiltern
● Domain labels must start with “Associated Persons”
● Any kind of exposure related variables are generally not used in AP domains
● Use timing variables with caution
• i.e. VISIT, EPOCH, RFXSTDTC
● Any ‘regular’ SDTM variables keep the xx suffix from APxx
• i.e. DMDTC still is DMDTC, not APDMDTC
● DOMAIN variable should match dataset name
How?! – Structure
7 Proprietary & Confidential. © 2016 Chiltern
How?! – APID
8 Proprietary & Confidential. © 2016 Chiltern
● Only needs to be unique within a study, but may be useful to have unique identifiers for
related studies
● When there is no relationship or the relationship is unknown, distinct APIDs should be used
• Incorporate unique elements such as subject IDs, domain codes, visit numbers
• Ex: DM_1, EX01
● APID identifies a single unit
• A unit could be defined as a person or a group
How?! – APID, cont.
9 Proprietary & Confidential. © 2016 Chiltern
How?! – POOLDEF
10 Proprietary & Confidential. © 2016 Chiltern
● Required if a pooled APID is used
● Identifies individual persons within a pooled APID
How?! – POOLDEF, cont.
11 Proprietary & Confidential. © 2016 Chiltern
- Example: testing a variety of blood samples from a blood bank for undesirable elements
- Each donor as an individual APID within the pooled APID, any test results for the pooled blood
would be under the pooled APID
- If there is a positive test then further tests would be needed on the individual blood samples to determine
which sample is positive
How?! – SREL
12 Proprietary & Confidential. © 2016 Chiltern
● Contains the relationship to the subject
● Uses RELSUB code list (extensible)
● “ACCIDENTAL ASSOCIATION” if there is no relationship or an accidental relationship
● Similar values can be used between SREL and EVAL
● SREL can account for multiple relationships
• If values “MULTIPLE” appear in SREL then a APRELSUB dataset is required
How?! – APRELSUB
13 Proprietary & Confidential. © 2016 Chiltern
● Contains definitions of all relationships when “MULTIPLE” is defined in SREL
● One AP can have multiple relationships to a single subject or a single AP can be related to
multiple subjects
● Often use of controlled terminology will require the use of APRELSUB
● Any APID with more than one relationship to a subject, device, or study must be represented
in APRELSUB
● Create APID in such a way to simplify process
How?! – APRELSUB, cont.
14 Proprietary & Confidential. © 2016 Chiltern
How?! – APRELSUB, cont.
15 Proprietary & Confidential. © 2016 Chiltern
How?! – APRELSUB, cont.
16 Proprietary & Confidential. © 2016 Chiltern
How?! – RSUBJID and RDEVID
17 Proprietary & Confidential. © 2016 Chiltern
● Null if AP is associated with the study rather than a subject or device
● RSUBJID: Populated with USUBJID of related subject or POOLID of the related pool (with
the definitions of the pool in POOLDEF)
● RDEVID: Populated with SPDEVID of the related device, only expected for device studies
● If the subject of a study is a device then RSUBJID would be null and RDEVID should be
populated
How?! – AP In Action – APDM
18 Proprietary & Confidential. © 2016 Chiltern
● Only created and submitted if demographics data is collected about APs in the study, not
required
● An APDM record for a pool should be included only if demographics data about the pool is
specifically collected
• Generally the individually identified persons within a pool (listed in POOLDEF) would have demographic data
and that is what should be presented in APDM
How?! – AP In Action – APMH, cont.
19 Proprietary & Confidential. © 2016 Chiltern
How?! – AP In Action – APMH, cont.
20 Proprietary & Confidential. © 2016 Chiltern
How?! – AP In Action – APMH, cont.
21 Proprietary & Confidential. © 2016 Chiltern
● APID can represent an individual or a group of people – even if these people are not
considered a “POOL”
● Note that the SREL value is singular, the number of people in the group is defined in the
SQAPMH as it is collected
How?! – AP in Action – APRP
22 Proprietary & Confidential. © 2016 Chiltern
● If subject is female, this would be in RP but if subject is MALE this would be in APRP
● SREL=“SEXUAL PARTNER”
• There is no CRF field which determines any other relationship so we cannot be more specific
● Apply CT if applicable to RPTEST and RPTESTCD
How?! – AP in Action – APRP, cont.
23 Proprietary & Confidential. © 2016 Chiltern
How?! – AP in Action – APSU
24 Proprietary & Confidential. © 2016 Chiltern
● Can be used to document family history:
How?! – AP in Action – APSU, cont.
25 Proprietary & Confidential. © 2016 Chiltern
● Could be used to document environmental exposure – substance use exposure from
household members
SREL SUOCCUR SUDOSE
SUDOSUSUDOSFRQ
How?! – AP in Action – APEX
26 Proprietary & Confidential. © 2016 Chiltern
● Accidental Exposure – when a study treatment is inadvertently exposed to a person who is
not in the study
How?! – AP in Action – APEX, cont.
27 Proprietary & Confidential. © 2016 Chiltern
● APEX – Accidental Exposure with a relationship to study subject
How?! – AP in Action – APAE
28 Proprietary & Confidential. © 2016 Chiltern
● APAE – Unanticipated serious adverse event for the operator
AETERM
AESTDTC
AEENDTC
AEENRF
AEACNOTHAESEV
AEOUT
AEREL
SQAPAE.QVAL where SQAP.QNAM=RELPROC
AESDTH
AESLIFEAESHOSP
AESDISAB
AESCONG
SREL
How?! – AP in Action – APAE
29 Proprietary & Confidential. © 2016 Chiltern
How?! – AP in Action – APAE, cont.
30 Proprietary & Confidential. © 2016 Chiltern
● AE Example 2:
● AE Example 3:
How?! – AP in Action – Pinnacle 21
31 Proprietary & Confidential. © 2016 Chiltern
● Using most recent version of Pinnacle 21 (v2.2.0 Pinnacle 21 Community):
● Domain name is truncated to 2 letters
● Unrecognized domain
● No checks performed whatsoever
How?! – Compliance Checks
32 Proprietary & Confidential. © 2016 Chiltern
● Check to make sure all of the AP required variables are populated:
• STUDYID, DOMAIN, APID, --SEQ, SREL
● Ensure all of the POOLIDs are represented in POOLDEF
● Ensure RDEVID is included as well as RSUBJID if there are any device related APs
● Ensure all APIDs are unique within the study
How?! – Compliance Checks, cont.
33 Proprietary & Confidential. © 2016 Chiltern
● Do a FREQ of SREL and ensure that the RELSUB code list is applied where applicable
● Ensure SREL formats are correctly created and that any values with derived multiple
relationships are represented
● Ensure that all “MULTIPLE” SRELs have a corresponding record in APRELSUB
● Ensure DOMAIN variable matches the domain name
How?! – Compliance Checks, cont.
34 Proprietary & Confidential. © 2016 Chiltern
● Ensure SQAP is used rather than SUPP
● Ensure the dataset label starts with Associated Persons
● Manually check all regular SDTM checks applicable to domain specific variables
• Specific rules apply for Events, Interventions, and Findings, and Special Purpose domains
Next Steps – ADaM Level AP
35 Proprietary & Confidential. © 2016 Chiltern
● Determine with which ADaM data structure the AP domain aligns the best
● If unique AP ADaM then named ADAPxx
● Always based on the analysis needs – how does this need to be summarized and/or listed?
● Pitfalls/Cautions
● Always be sure to detail your approach in ADaM Reviewer’s Guide
● Consult with the review panel at FDA
Any Questions?
36 Proprietary & Confidential. © 2016 Chiltern
● Michael Stackhouse
Associate Manager, Statistical Programming
Chiltern
Cary, NC
● Alyssa Wittle
Senior Statistical Programmer – 1
Chiltern
King of Prussia, PA
PharmaSUG SDE RTP 2016October 27, 2016