Random slope in MMRM analysis

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Random slope in MMRM analysis

Frank Furter
Hi,

I have a MMRM analysis that basically looks like this:

MIXED Change BY Treatment Center Age Sex Visit WITH Baseline
  /FIXED=Treatment Center Age Sex Visit Baseline
         Treatment*Visit | SSTYPE(3)
  /METHOD=REML
  /REPEATED=Visit | SUBJECT(SubjID) COVTYPE(un)
  /EMMEANS = TABLES(Treatment*Visit) compare(Treatment).

Baseline is the baseline value of the outcome of interest, and Change is its change from baseline for each of three post-baseline visits.

How can I specify a random slope for this model?

Thank you,
Andreas
Reply | Threaded
Open this post in threaded view
|

Re: Random slope in MMRM analysis

Maguin, Eugene
Curious. Is Center a level 2 variable such that at multiple centers patients are assigned to treatment condition C1 or C2 and assessed at visits V1 to V4? If so, then wouldn't the random effect be the center intercept? The documentation (CSR) used to have an example like this for mixed. It's gone from the CSR for some reason (or no good reason) and has moved to the case studies documentation component.

This isn't what you asked about but I'm curious about your analysis statement. Correct me if I am misunderstanding but isn't your basic equation this:

Change(i) = intercept + V(1)*B1 + b2*age + b3*sex + e where Change(i) = (V(i, i > 1) - V(1))

What is(are) the advantages of this DV construction over the more typical observed values analysis?

Gene Maguin

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Andreas Voelp
Sent: Tuesday, February 28, 2017 11:32 AM
To: [hidden email]
Subject: Random slope in MMRM analysis

Hi,

I have a MMRM analysis that basically looks like this:

MIXED Change BY Treatment Center Age Sex Visit WITH Baseline
  /FIXED=Treatment Center Age Sex Visit Baseline
         Treatment*Visit | SSTYPE(3)
  /METHOD=REML
  /REPEATED=Visit | SUBJECT(SubjID) COVTYPE(un)
  /EMMEANS = TABLES(Treatment*Visit) compare(Treatment).

Baseline is the baseline value of the outcome of interest, and Change is its change from baseline for each of three post-baseline visits.

How can I specify a random slope for this model?

Thank you,
Andreas



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Random-slope-in-MMRM-analysis-tp5733912.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD