Using DS Disposition

1
Using DS- Disposition

• One Key to Good SDTM Adoption
• By J.J. Hantsch

2
DS - Concept

• Disposition the act or power of disposing or
  the state of being disposed as final arrangement,
  transferal.

3
Typically,

• So SDTM has a disposition dataset . . .
• Hmmh, . . . final arrangement.
• Typical reading
• Disposition end of study
• Typical idea
• I have an EOS dataset Ill just convert that.

4
BUT,

• RTFM

5
SDTM IG V3.1.2

• . . . include protocol milestones, such as
  randomization, as well as subjects completion
  status or reason for discontinuation for the
  entire study or each phase or segment of the
  study, including screening and post-treatment
  follow-up. . .

6
Which means?
7

• Start with a schedule of events
• Mark each milestone
• screening
• randomization
• each treatment change
• end of dosing
• any follow-up events
• disposition (EOS)

8
For example, I (a study schedule)

• Screening Visit 1
• Randomization visit Visit 2
• Run-in dosing one Visit 3
• Run-in dosing two Visit 4
• Experimental dose 1A Visit 5
• Experimental dose 1B Visit 6
• Experimental dose 1C Visit 7
• Wash-out Period Visit 8
• Experimental dose 2A Visit 9
• Experimental dose 2B Visit 10
• Experimental dose 2C Visit11
• Follow-Up Visit Visit 12

9
Marking the study schedule

• Screening Visit 1
• Randomization visit Visit 2
• Run-in dosing one Visit 3
• Run-in dosing two Visit 4
• Experimental dose 1A Visit 5
• Experimental dose 1B Visit 6
• Experimental dose 1C Visit 7
• Wash-out Period Visit 8
• Experimental dose 2A Visit 9
• Experimental dose 2B Visit 10
• Experimental dose 2C Visit 11
• Follow-Up Visit Visit 12

10

• Mark each milestone
• screening visit 1
• randomization visit 2
• each treatment change visits 3, 5, 8, 9
  
• disposition (EOS) visit 12, early
  term

11
So 7 records?

• Completing subjects will have seven (7) unique
  records in the DS dataset.
• One record per protocol milestone per subject
• Fewer for Screen Failures, Refused Consenters or
  Early Terminators

12
DS Required Variables

• STUDYID
• DOMAIN
• USUBJID
• DSSEQ
• DSTERM
• DSDECOD
• DSCAT
• DSSTDTC

13
Required Variables (2)

• STUDYID study identifier
• DOMAIN DS
• USUBJID subject identifier
• DSSEQ a sequence number

14
Required Variables (3)

• DSTERM verbatim name of event or protocol
  milestone
• DSDECOD controlled term for event or protocol
  milestone
• DSCAT DISPOSITION EVENT, PROTOCOL MILESTONE, or
  OTHER EVENT
• DSSTDTC Start date/time of event

15
Example

• STUDYID USUBJID DSSEQ DSTERM DSDECOD DSCAT
  DSSTDTC
• BOB-101 B101-001 1 Screening SCREEN
  Protocol Milestone 10-22-07
• BOB-101 B101-002 1 Screening SCREEN
  Protocol Milestone 10-22-07
• BOB-101 B101-003 1 Screening SCREEN
  Protocol Milestone 10-23-07
• BOB-101 B101-004 1 Screening SCREEN
  Protocol Milestone 10-24-07
• BOB-101 B101-005 1 Screening SCREEN
  Protocol Milestone 10-25-07
• BOB-101 B101-001 2 Randomization
  RANDOM Protocol Milestone 10-29-07
• BOB-101 B101-003 2 Randomization
  RANDOM Protocol Milestone 10-30-07
• BOB-101 B101-006 1 Screening
  SCREEN Protocol Milestone 10-31-07
• BOB-101 B101-002 2 Early Termin
  ERLYTRM Disposition Event 10-31-07

16
GREAT!!

• Now what?

17
Would you like this?

• Dear Clinical Operations,
• If I could provide you with a list of the
  status of all of the subjects in each of your
  clinical trials, broken out by updated (changed)
  status and totals for each milestone attained in
  the clinical trial, each time when I receive a
  refreshed source dataset, would you like that?
• Your Programmer,

18
What does that mean?

• A simple report
• Run once each time you receive data
• For each study
• Each Milestone
• Contrast Total and New

19
Study BOB-101
Report Date 25 Dec 2007
Report Prepared by J.J. Hantsch
Status CURRENT
20
Programming Needs

• Sort the new dataset
• Keep the previous versions dataset
• Compare to new dataset
• Create Flag variable for new milestone
• Use Proc Freq or Tabulate
• Tabulate results
• Save most recent as old

21
Embellishments

• Email the report to clinical people
• Put totals on the far left
• For Blackbarristers (Blackberry users)
• Double Report
• Summary
• Cross by treatment
• Throw in Percentages()

22
Dont Forget

• Display the Date PROMINENTLY
• Identify the Study (Studies)
• Think ISS/ ISE
• Identify CURRENT or COMPLETE
• Identify Yourself

23
The Result
24
?