stat_mmrm: Create an mmrm stat object.

In aba: Automated Biomarker Analysis

Description

This function creates an mmrm stat object which can be passed as input to the set_stats() function when building an aba model. This stat performs a MMRM analysis using the gls function from the nlme package. Please note that the default mode is to include an interaction term between the time variable and each predictor - i.e., time*predictor will be in the model formula - but this does not happen for covariates. The data for this model should be in long format with one row per subject-visit.

Usage

stat_mmrm(
  id,
  time,
  treatment = NULL,
  baseline_suffix = "bl",
  std.beta = FALSE,
  complete.cases = TRUE
)

Arguments

id	string. This is the variable in the data which represents the subject id to be used for random intercepts and random slopes.
time	string. This is the time variable in the data which represents the time from baseline that the visit occured. This should be a categorical variable or a continuous variable where the values are shared by all subjects. The fact that time visits should be common across all subjects is a major operational difference from stat_lme, among other differences.
treatment	string. The treatment variable whose effect on the outcome you care about. This is useful for aba_emmeans and other functions.
baseline_suffix	string. The suffix to add to each outcome variable in order to pick up the associated baseline variable. You must adjust for the baseline outcome in mmrm, and there is no other way to specify a different predictor for each outcome. So if the outcomes are e.g. "CDRSB" and "MMSE", then a baseline_suffix of "bl" will mean that each mmrm fit with "CDRSB" as outcome will have "CDRSB_bl" added to the formula and every fit with "MMSE" as outcome will have "MMSE_bl" added. This means that these baseline variables must actually exist in the data. Also, there will always be an interaction between the baseline outcome variable and the time variable.
std.beta	logical. Whether to standardize model predictors and covariates prior to analysis.
complete.cases	logical. Whether to only include the subset of data with no missing data for any of the outcomes, predictors, or covariates. Note that complete cases are considering within each group - outcome combination but across all predictor sets.

Value

An abaStat object with mmrm stat type.

Examples

data <- adnimerge %>%
  dplyr::filter(VISCODE %in% c('bl','m06','m12','m24'))

model <- data %>% aba_model() %>%
  set_groups(
    DX_bl %in% c('MCI', 'AD')
  ) %>%
  set_outcomes(CDRSB, ADAS13) %>%
  set_predictors(
    PLASMA_ABETA_bl,
    PLASMA_PTAU181_bl,
    PLASMA_NFL_bl
  ) %>%
  set_covariates(AGE, GENDER, EDUCATION) %>%
  set_stats(
    stat_mmrm(id = 'RID', time = 'VISCODE')
  ) %>%
  fit()

model_summary <- model %>% aba_summary()

aba documentation built on Dec. 17, 2021, 1:06 a.m.