pmrm_marginals: Marginal means

In pmrm: Progression Models for Repeated Measures

Marginal means

Description

Report the estimates and standard errors of marginal means at each study arm and visit. The assumed visit times should have been given in the marginals argument of the model-fitting function. Use the type argument to choose marginal means of the outcomes, marginal estimates of change from baseline, and marginal estimates of treatment effects.

Usage

pmrm_marginals(fit, type = c("outcome", "change", "effect"), confidence = 0.95)

Arguments

fit	A pmrm fitted model object returned by a model-fitting function.
type	Character string. "outcome" reports marginal means on the outcome scale, "change" reports estimates of change from baseline, and "effect" reports estimates of treatment effects (change from baseline of each active arm minus that of the control arm.)
confidence	A numeric from 0 to 1 with the confidence level for confidence intervals.

Value

A tibble with one row per marginal mean and columns with the estimate, standard error, 2-sided confidence bounds, and indicator columns. Some estimates, standard errors, and confidence bounds may be NA_real_ if they correspond to the reference level subtracted out in change-from-baseline or treatment effect calculations.

See Also

Other estimates: VarCorr.pmrm_fit(), coef.pmrm_fit(), tidy.pmrm_fit(), vcov.pmrm_fit()

Examples

  set.seed(0L)
  simulation <- pmrm_simulate_decline_proportional(
    visit_times = seq_len(5L) - 1,
    gamma = c(1, 2)
  )
  fit <- pmrm_model_decline_proportional(
    data = simulation,
    outcome = "y",
    time = "t",
    patient = "patient",
    visit = "visit",
    arm = "arm",
    covariates = ~ w_1 + w_2
  )
  pmrm_marginals(fit)

pmrm documentation built on March 12, 2026, 5:07 p.m.