Shift table

[KalaBhairava]

Hi All,

Can any one explain easily understanding  way to do Shift table.

[Kurt_Bremser]

Please be more specific. What do you mean by "shift"?

[KalaBhairava]

Shift table for Physical examination in clinical trails. 

[KalaBhairava]

Shift tables are very important in clinical trial studies. A shift table is a table that displays the number of
subjects who are low, normal or high at baseline and the shift at post-dose.

[Kurt_Bremser]

---

@KalaBhairava wrote:
Shift tables are very important in clinical trial studies. A shift table is a table that displays the number of
subjects who are low, normal or high at baseline and the shift at post-dose.

---

So you need to do a certain transformation of the data in a single table.

If you provide some usable example data (data step with datalines), and the expected outcome, I can tinker around with it.

[Jianmin · Accepted solution]

Hi Kurt, I'm late to replay and I hope by now you have figured it out about the shift table.  Here is a link how others talk about shift table: https://www.lexjansen.com/cgi-bin/saspapers_query.php title search for "shift table". 

Shift table is important only because it was not easy to get for some important parameters in clinical trials.  We have done change from Baseline for continuous variables all the time, but for a very long time people haven't figure it out that we can do the same for categorical variables as we did for continuous variables.  In the case of 4 possible categorical outcomes, a 4 by 4 table (shift table) is used to present the result.  It takes a lot of space, and most of those 16 cells have no values. A little bit more complicated when you compare tables.  A simplified and short-handed shift table is proposed.  See the attached. 

Best Regards

Jianmin Long

PS: the data in the attached is openly available on https://www.cdisc.org/sdtmadam-pilot-project

[RW9]

I assume your following CDISC models?  If so then its just a matter of merging on baseline record to all post-baseline records, then reporting this to a table.  So your data would look like:

USUBJID  ...  ANRIND    BNRIND

xxx                 LOW         HIGH

xxx                 NORMAL  HIGH

Then its a matter of processing that information into the output shell you have.  Unfortunately, exact code and such like is company IP, so you will need to do this yourself, using your company standards, shells, macros.

[KalaBhairava]

Yes thanks for replay and i need clarification on how to start procedures not code because i am new to this shift tables.