Changelog

tern.mmrm 0.3.3

Created additional tests for plot functions.
Added parent_name parameter to internal rtables::split_rows_by() calls to ensure uniqueness of row names, due to upstream rtables enhancements.

CRAN release: 2024-09-23

CRAN release: 2024-06-12

CRAN release: 2023-12-14

Adapt to release 0.3 of the mmrm package.

Any additional arguments can be passed now via ... to mmrm::mmrm when calling fit_mmrm. In particular, the method argument allows to choose Kenward-Roger adjustment of degrees of freedom and coefficients covariance matrix.

The optimizer interface is different now: For choosing automatically the right optimizer, just omit the optimizer argument when calling fit_mmrm.

parallelly is now used internally to handle the determination of available cores.

Completed refactoring of the computations to use the new mmrm package instead of lme4 and lmerTest. This greatly increases convergence and speed. Different covariance structures and optimizers are now available compared to before.

Added function g_covariance() to visualize a covariance matrix, which can be helpful for choosing or visualizing the covariance structure in the MMRM.
Added option averages_emmeans to fit_mmrm() which allows estimation of least square means for averages of visits.
Added weights option for fitting MMRMs.

Added option accept_singular to fit_mmrm() which allows estimation of rank-deficient models (like lm() and gls()) by omitting the columns of singular coefficients from the design matrix.
Added options show_lines and xlab to g_mmrm_lsmeans() which allow the addition of lines connecting the estimates, as well as a custom x-axis label.
Added options table_stats, table_formats, table_labels, table_font_size, and table_rel_height to g_mmrm_lsmeans() which allow the addition of and configure the appearance of an LS means estimates statistics table below the LS means estimates plot.
Added options constant_baseline and n_baseline to g_mmrm_lsmeans() which allow plotting of a constant baseline value and specifying the corresponding number of patients (visible in the optional table) for the LS means plots.