- The estimation of the effect sizes now relies on the metaConvert
package.
- The GRADE criteria have been reworked in collaboration with a large
team with extensive experience in evidence synthesis
- Correct some bugs, including an error that could appear in the
Egger’s test if ‘outcome’ instead of ‘outcomes’ was included in the
mutiple_es colmun (github issue).
- Rework importation of the forest function from the meta package
- Redesign of the forest.umbrella function (now a wrapper of the
amazing forest.meta function from the meta package). This also allows to
plot R as effect size measures, and to correct some bugs that could
appear when using “eOR” in v1.0.8.
- The add_evidence function can now stratify the evidence according to
Ioannidis criteria for R and Z measures if users enter the number of
cases
- Fixed an issue that could appear when a meta-analysis presented
multiple outcomes with a correlation as effect size measure
- Added a groove() function to visualize overlap of primary studies
(inspired by the work of Pérez-Bracchiglione et al. 2022)
- Fixed some bugs preventing to use all functions in the app and
jamovi module
- Three new effect size measures are available (SMC, R, and Z)
- Three new ESB tests are available (from Stanley and colleagues,
2021: PSST, TESS, TESSPSST)
- Fixed some bugs when drawing a forest plot with colors
- The summary() function now allows to obtain more information on
heterogeneity (via the ‘het_max’ argument)
- The umbrella() function now allows to request a fixed-effect
meta-analysis (via the ‘method.var’ argument)
- The umbrella() function now allows to choose the excess of
statistical significance test (via the ‘method.esb’ argument)
- The unrestricted weighted least squares weighted average can now be
used to estimate the best approximation of the true effect (via the
‘true_effect’ argument of the umbrella() and esb.test() functions)
- The add.evidence() function now warns users when an issue occurred
when estimating the ESB test (a warning message is thus printed when
using the umbrella() + add.evidence() functions).
- New features are available in the forest.umbrella() function
- Test section: some tests have been updated
- umbrella() function: improvement of the meta-analysis selection
process based on user input
- Test section: several tests have been added
- Test section: made a rounding correction and explicitly calls the
estimator for the amount of heterogeneity (method.tau argument) when
using the functions of the meta package
- Documentation: made some improvements (correction of typos and
clarifications)
- First version released on CRAN