Transcript Trial Design Introduction

Trial Design Introduction
Elke Sennewald, 22 September 2011
Trial Design Domains
• Information about study design
– No subject data
• Describe the overall trial design and plan via data representation
Why Do Trial Design
• Rapidly understanding the design of the study
• Standard and relatively simple data structures
• Relatively small number of rows of data and easy to comprehend
• Useful for both FDA reviewers and internal sponsor use
• Information can be centrally accessible and searchable
Trial Design Datasets
• Trial Arms (TA)
• Trial Elements (TE)
• Trial Visits (TV)
• Trial Inclusion /Exclusion (TI)
• Trial Summary (TS)
• Start thinking about this before you start the other SDTM datasets!
4
Trial Summary (TS) Dataset
• Summary of trial information
• No link to subject-level data in SDTM
• TSGRPID used to group multiple related parameters such as Dose,
Units, Frequency etc
• TSSEQ used as a key for multiple records with the same parameters
• Common questions:
– What need to be included?
– Why are we generating this?
5
Trial Inclusion/Exclusion (TI)
• Not subject-oriented
• Link to IE domain
– STUDYID, IECAT, IETESTCD, IETEST
– Subject IETEST/IETESTDC must match Trial Inclusion/Exclusion
IETEST/IETESTCD
– Best to create TI first, before you tackle IE
• Common questions:
– How to truncate if >200 characters?
– Truncation – potential for duplicate IETEST values
– Protocol amendment: do we need to add to TI only the changed criteria or
all criteria?
– Local amendment
6
TA / TE / TV datasets
• A data representation on the different epochs, arms and visit structure
in the study
• Where to start?
• Is there a systematic approach?
7
Example 1 – Trial Design Schema
Drug A
Follow-up
Drug B
Follow-up
Screen
8
Epoch
Drug A
Follow-up
Drug B
Follow-up
Treatment
Follow-up
Screen
EPOCH
Screening
9
(Treatment
Strategy)
ARM
Arm / Treatment Strategy
Drug A
Follow-up
Drug B
Follow-up
Treatment
Follow-up
Screen
Screening
1
2
10
Arm / Treatment Strategy
Drug A
Follow-up
Drug B
Follow-up
Screening
Treatment
Follow-up
1
Screen
Drug A
Follow-up
2
Screen
Drug B
Follow-up
Screen
11
Trial Design Matrix
Screening
Treatment
Follow-up
A
Screen
Drug A
Follow-up
B
Screen
Drug B
Follow-up
12
TE (Trial Elements)
• What are the elements?
– Unique study cell values (=ELEMENT)
Screen
Drug A
Drug B
Follow-up
13
Trial Arms and Elements Overview
Trial Elements describes the
Elements and the rules for
the start and end of each.
Trial Arms
describes the
Elements in each
Arm, their order
and Epoch, and
any branching or
transition rules.
Screen
Run-in
Drug A
Placebo
Drug B
Placebo
Screen
Run-In
Placebo
Drug A
Screen
Run-In
Drug A
Drug B
Screen
Run-In
Drug B
Epochs are described only in Trial
Arms, and have no separate table.
Trial Visits describes the
planned Visits for each Arm,
and any start and end rules.
Screening
Visit 1
Visit 2
Run-In
Treatment
Visit 3
Visit 4
Visit 5
Trial Design Matrix
Screening
Run-in
Treatment
P
Screen
Run-in
Placebo
A
Screen
Run-in
Drug A
B
Screen
Run-in
Drug B
15
Creating Trial Elements (1)
• Usually the most challenging dataset
• Not a duplication of EX (Exposure)
• Assign an element code (ETCD) to each value, define the
start of each element (TESTRL) and end of each element
(TEENRL or TEDUR)
• Start rules are the most important
– Subject data must exist to support the creation of these
– Start of next element defines end of previous
Creating Trial Elements (2)
Screen
Trial Elements describes the
Elements and the rules for
the start and end of each.
ETCD
Run-in
Drug A
TESTRL
Drug B
STUDYID
DOMAIN
1999001
TE
SCRN
Screen
Informed consent
signed
Start of next
element or date
subject dropped
P2W
1999001
TE
RUNIN
Run-in
First dose of run-in
drug
Start of next
element or date
subject dropped
P1W
1999001
TE
PLAC
Placebo
First dose of placebo
Start of next
element or date
subject dropped
or completed
P2W
Example pseudocode:
DSSTDTC where DSDECOD =
INFORMED CONSENT
ELEMENT
Placebo
TEENRL
Example pseudocode:
EXSTDTC where EXTRT =
RUN-IN DRUG
TEDUR
TE -> SE (Subject Elements)
• Shows the trial progress of each subject
– Whether a subject passes through each element
– Timing of each element
18
Creating Subject Elements
STUDYID DOMAIN USUBJID SESEQ
ETCD
ELEMENT
SESTDTC
SEENDTC
1999001
SE
145-011
1
SCRN
Screen
2003-04-01
2003-04-15
1999001
SE
145-011
2
RUNIN
Run-In
2003-04-15
2003-04-22
1999001
SE
145-011
3
PLAC
Placebo
2003-04-22
2003-05-06
DSSTDTC where DSDECOD =
INFORMED CONSENT
EXSTDTC where EXTRT =
RUN-IN DRUG
SEUPDES
Trial Arms (TA) Dataset
• High level treatment plan
• Composed of Elements from Trial Elements
• Go back to the Trial Design Matrix
• 1 study cell = 1 row of record in TA
• So in our example we expect 9 rows of record
• Planned ARM values in DM correspond to ARM values in Trial Arms
• Names of ARM should reflect the protocol
20
Creating Trial Arms
STUDYID
Placebo
Screen
Run-In
Placebo
Drug A
Screen
Run-In
Drug A
Drug B
Screen
Run-In
Drug B
DOMAIN
ARMCD
ARM
TAETORD
ETCD
TABRANCH
EPOCH
1999001
TA
P
Placebo
1
SCRN
1999001
TA
P
Placebo
2
RUNIN
1999001
TA
P
Placebo
3
PLAC
Treatment
1999001
TA
A
Drug A
1
SCREEN
Screen
1999001
TA
A
Drug A
2
RUNIN
1999001
TA
A
Drug A
3
DRUGA
Treatment
1999001
TA
A
Drug B
1
SCREEN
Screen
1999001
TA
A
Drug B
2
RUNIN
1999001
TA
A
Drug B
3
DRUGB
Screen
Randomized to
Placebo
Randomized to
Drug A
Randomized to
Drug B
Run-In
Run-In
Run-In
Treatment
Trial Visit (TV) Dataset
• Describe the planned visits in a trial
• VISITNUM and TRSTRL is required
• ARMCD expected
• VISIT and VISITDY permissible
• 1 record per planned visit per arm
– A “visit” may span over several days (eg screening visit)
• What is really the start and end of a visit?
• Create Subject Visits dataset from Visit based SDTM datasets
22
TV -> SV (Subject Visits)
• Shows the actual visits of each subject
– Compare against the scheduled/planned visits or assessments in TV
– Include unscheduled visits
• Designation of VISITNUM becomes crucial
– Whole number for planned visits
– Decimals for unscheduled visits in SV – and slot into right place
23
Creating Trial Visits
• Planned schedule of Visits
• Challenge is in defining start and end of a visit
• ARM/ARMCD can be used if schedule varies by Arm
STUDYID
DOMAIN
VISITNUM
VISIT
VISITDY
ARMCD
ARM
TVSTRL
1999001
TV
1
Visit 1
Signing of
informed consent
1999001
TV
2
Visit 2
30 minutes
before receipt of
Run-In drug
1999001
TV
3
Visit 3
30 minutes
before receipt of
blinded treatment
1999001
TV
4
Visit 4
1 week after
receipt of blinded
treatment
1999001
TV
5
Visit 5
2 weeks after
receipt of blinded
treatment
TVENRL
Completion of
lab draw
30 minutes
after receipt of
blinded
treatment
Completion of
final
disposition
page
Summary
• Construction of TA/TE/TV
– Study Schema  Epoch  Arm  Study Cells
– Unique study cells = rows in TE
– All study cells = rows in TA
– If all arms have same visits, then 1 set of visits for all arms. Otherwise 1 set
of visits for each arm.
• Complex study designs
– Systematic approach will make life easier
– Think at protocol/CRF design stage – don’t wait till the end
– Details vs ease of use
25