Generalized Linear Mixed Models - MMRM

[kaigu1990]

@statasaurus Is there a plan to compare how SAS and R compute the LS mean in the MMRM? And moving forward, to do more specific scenarios, such as computing the LS mean in the non-inferiority or superiority trials.

[statasaurus]

Hello @kaigu1990 we don't have a specific plan for this, but if you would like to add in the content I imagine it would be very helpful. Thank you 😃

[yonicd]

there are examples of LSMeans calculations in the mmrm package repository documentation

[kaigu1990]

Hi @statasaurus and @yonicd , Thanks for your advice. After learning the mmrm and emmeans related documents, we can compute the least-square means, corresponding contrasts and hypothesis testing using functions in emmeans .

First, we fit a MMRM model.

library(mmrm)
library(emmeans)

fit <- mmrm(
  formula = FEV1 ~ RACE + SEX + ARMCD * AVISIT + us(AVISIT | USUBJID),
  reml = TRUE, method = "Kenward-Roger", vcov = "Kenward-Roger-Linear",
  data = fev_data
)
summary(fit)

And then compute the least-square means with corresponding CI.

ems <- emmeans(fit, ~ ARMCD | AVISIT)
confint(ems)

Suppose we want to answer the question that the null hypothesis is that treatment minus placebo equals zero. The contrast can be created below.

contr <- contrast(ems, adjust = "none", method = "pairwise")
contr

Suppose we would like to test the hypothesis that treatment is superior to placebo with a margin of 2.

test(contr, null = 2, side = ">")

---

This is just a brief summary. I have also compared the above results with the SAS code below. If there is something wrong, please tell me.

proc mixed data=fev_data;
    class ARMCD(ref='PBO') AVISIT RACE SEX USUBJID;
    model FEV1 = RACE SEX ARMCD ARMCD*AVISIT / ddfm=KR;
    repeated AVISIT / subject=USUBJID type=UN r rcorr;
    lsmeans ARMCD*AVISIT / cl alpha=0.05 diff slice=AVISIT;
    lsmeans ARMCD / cl alpha=0.05 diff;
    lsmestimate ARMCD*AVISIT [1,1 4] [-1,2 4] / cl upper alpha=0.025 testvalue=2;
    ods output lsmeans=lsm diffs=diff LSMEstimates=est;
run;

[DrLynTaylor]

Just FYI: @stephenwaugh will be investigating further into MMRMs and updating these pages from Sept 2024 - July 2025.