Skip to contents

MMRM Analysis

Source: R/mmrm.R
mmrm_analysis.Rd

Fits a Mixed Model Repeated Measures (MMRM) for longitudinal data analysis.

Usage

mmrm_analysis(
  data,
  response_var,
  subject_var,
  trt_var,
  time_var,
  covariates = NULL,
  interaction = FALSE,
  cov_covariance = c("us", "cs", "ar1", "ad", "toep", "sp_exp"),
  df_adjustment = c("Kenward-Roger", "Satterthwaite", "Residual"),
  method = c("REML", "ML"),
  control = NULL
)

Arguments

data

A data frame containing the longitudinal data

response_var

Character. Name of the response variable

subject_var

Character. Name of the subject identifier variable

trt_var

Character. Name of the treatment variable

time_var

Character. Name of the time/visit variable

covariates

Character vector of covariate variable names

interaction

Logical. Whether to include treatment by time interaction

cov_covariance

Character. Covariance structure: "us" (unstructured), "cs" (compound symmetry), "ar1" (autoregressive), "ad" (ante-dependence), "toep" (Toeplitz), "sp_exp" (spatial exponential)

df_adjustment

Character. Degrees of freedom adjustment: "Kenward-Roger", "Satterthwaite", "Residual"

method

Character. Estimation method: "REML" (default) or "ML"

control

List of control parameters for optimization, passed to mmrm. Typically created with mmrm_control.

Value

An MMRMResult object

Details

The MMRM model fits: Y_ij = X_ij * beta + epsilon_ij

Where Y_ij is the response for subject i at time j, X_ij contains fixed effects (treatment, time, covariates, interactions), and epsilon_ij ~ N(0, Sigma) where Sigma is the covariance matrix.

Common covariance structures:

  • "us": Unstructured - most flexible, all covariances differ

  • "cs": Compound symmetry - equal variances and covariances

  • "ar1": Autoregressive - covariance decreases with time distance

  • "ad": Ante-dependence - flexible for unequal time intervals

References

Mallinckrodt, C.H. et al. (2013). Choosing the optimal mixed-effects model for repeated measures data. Statistical Methods in Medical Research, 22(2), 113-138.

Examples

if (FALSE) { # \dontrun{
# Basic MMRM with treatment and time effects
result <- mmrm_analysis(
  data = pro_data,
  response_var = "AVAL",
  subject_var = "USUBJID",
  trt_var = "TRT01P",
  time_var = "AVISITN",
  covariates = "BASE"
)

# MMRM with interaction and custom covariance
result <- mmrm_analysis(
  data = pro_data,
  response_var = "AVAL",
  subject_var = "USUBJID",
  trt_var = "TRT01P",
  time_var = "AVISITN",
  covariates = c("BASE", "AGE"),
  interaction = TRUE,
  cov_covariance = "cs"
)
} # }