describe_draws()extracts structured metadata from rbmi draws objects, including method type, formula, sample count, and (for Bayesian methods) MCMC convergence diagnostics (ESS, Rhat).describe_imputation()extracts imputation metadata including method, number of imputations (M), reference arm mappings, and a missingness breakdown by visit and treatment arm.pool_to_ard()gains ananalysis_objparameter that enriches the ARD with MI diagnostic statistics (FMI, lambda, RIV, Barnard-Rubin adjusted df, relative efficiency) when the pooling method is Rubin's rules.
efficacy_table()gainsfont_family,font_size, androw_paddingparameters for publication-ready table styling.plot_forest()gainsfont_familyandpanel_widthsparameters for customizable typography and panel layout.plot_forest()left panel now uses left-aligned text (hjust=0) for consistent positioning regardless of label length.- Non-Rubin pooling methods now emit an informative
cli::cli_inform()message when MI diagnostics are not applicable, rather than returning NA rows.
- End-to-end pipeline vignette "From rbmi Analysis to Regulatory Tables" with continuous and binary analysis walkthroughs.
- README enhanced with rendered efficacy table and forest plot visual teasers.
- Rendered example output images in
plot_forest()andefficacy_table()help pages. - Inline cross-references to rbmi and beeca documentation in all vignettes.
validate_data()now uses cli-formatted error messages with clearer guidance for malformed interaction terms, empty data, and type mismatches.prepare_data_ice()now errors immediately whenvars$strategyis NULL instead of silently using a default column name.prepare_data_ice()warns when visit column is character with guidance to convert to factor for correct ordering.validate_data()batches all type coercion warnings into a single message and warns on all-NA covariate columns.- Data preparation functions handle edge cases (single subject, single visit, all-NA outcome, all-complete data) gracefully.
tidy_pool_obj()now uses regex-based parameter parsing instead of splitting on_. Output columns (parameter_type,lsm_type,visit) now contain the full visit name rather than truncated fragments.
efficacy_table()creates regulatory-style gt summary tables from pool objects.plot_forest()creates publication-quality three-panel forest plots from pool objects.pool_to_ard()converts pool objects to pharmaverse ARD format via the cards package.print()andsummary()S3 methods for pool and analysis objects.create_impid()converts lists of imputed datasets into stacked data.frames.combine_results()combines tidy results from multiple analyses.format_results()andformat_results_table()for publication-ready formatting.extract_trt_effects()andextract_lsm()convenience filters for tidy results.
- Added
cli(>= 3.6.0) andlifecycle(>= 1.0.4) to Imports for improved error messaging and deprecation support.
- Enhanced input validation in
analyse_mi_data()with clearer error messages usinginherits()instead of internal class helpers. - Enhanced input validation in
prepare_data_ice()to check vars fields. - Cross-references between related functions in documentation.
- Expanded test coverage and integration tests.
reduce_imputed_data()andexpand_imputed_data()for efficient storage of imputed datasets.validate_data()for pre-flight validation beforerbmi::draws().prepare_data_ice()to builddata_icefrom flagged ICE columns.summarise_missingness()for tabulating missing data patterns.format_pvalue(),format_estimate(),format_results_table()formatting utilities.ADEFFandADMIexample datasets.
- Initial documentation via pkgdown.
- Released to CRAN on 2026-01-24.
- Moved tidyr from Suggests to Imports (used in tidy_pool_obj()).
- Added examples to gcomp_responder() documentation.
- Standardized examples to use native pipe operator |>.
- Released to CRAN on 2025-09-13.
- Added additional tests for all utility functions.
- Released to CRAN on 2025-05-23.
- Initial draft package.
- Added a NEWS.md file to track changes to the package.
- Preparation for CRAN submission.
- Added documentation via pkgdown.