- Modified reference line for continuous outcome to be at 0 instead of 1 in the network.forest.plot.
- Updated baseline risk section in the vignette.
- Modified network.run function so that the function returns the variables that were tested for convergence using Gelman and Rubin's diagnostic. Now it returns conv.save which stores the variables tested.
- Modified contrast.network.run to avoid error.
- When calculating treatment effect in the natural scale (i.e. absolute treatment effect coded T in JAGS code), added option to specify covariate/baseline risk that treatment effect should be calculated at.
- Added different treatment comparisons for RR, RD, and NNT
- Added Treat.order variable in ume model
- Revised the function network.leverage.plot to include per arm contributions
- Added a plot that compares posterior mean deviance of inconsistency model and consistency model
- When fitting meta regression on baseline risk, made adjustments when the baseline treatment is not the control treatment in each trial (i.e. by adding a fictitious row).
- Added finding absolute treatment effect in binomial and normal outcomes
- Added RNG.inits in the network.run function for reproducibility
- Added modeling baseline risk and covariate effects when fitting a fixed effects model
- Modified code for network.forest.plot that allows plotting only the comparison with the reference treatment and fixed labeling issue
- Added number needed to treat
- Updated vignette and added description section
- Added an extension to calculate relative risk and risk difference (using odds ratio and estimated placebo event rate) when the outcome is binomial
- Added a
NEWS.mdfile to track changes to the package.