v0.4.0

• Function comp_clustering:
  • Performs quantitative evaluation of the transitivity assumption using
    inter-trial dissimilarities for various trial-level aggregate participant
    and methodological characteristics that may act as effect modifiers.
• Function dendro_heatmap:
  • Returns the dendrogram with integrated heatmap of the clustered comparisons
    and trials based on hierarchical agglomerative clustering (performed using
    the function comp_clustering). The R packages heatmaply and
    dendextend have been used.
• Function distr_characteristics:
  • It returns violin plots with integrated box plots and dots for quantitative
    characteristics, and stacked barplots for qualitative characteristics across
    the observed treatment comparisons. The function can also be used to
    illustrate the distribution of the characteristics across the clusters
    defined from comp_clustering.
• Function miss_characteristics:
  • It returns various plots to visualise the missing cases in the
    characteristics across trials and treatment comparisons.
• Function gower_distance:
  • It returns the N-by-N matrix on Gower's dissimilarity coefficient for all
    pairs of N trials in a network.
• Function mcmc_diagnostics:
  • returns a bar plot on the ratio of MCMC error to the posterior standard
    deviation and a bar plot on the Gelman-Rubin R diagnostic. Green bars
    indicate ratio less than 0.05 and R less than 1.10; otherwise, the bars are
    red.
• Functions baseline_model, run_metareg, run_model,
  run_nodesplit, run_sensitivity, run_series_meta, and run_ume:
  • The corresponding models are updated until convergence via the autojags
    function of the R package R2jags.
  • The argument inits has been added to allow the user define the initial
    values for the parameters, following the documentation of the jags function
    in the R package R2jags.
• Function describe_network:
  • It reports only the tables with the evidence base information on one
    outcome. The network plot is not reported (see and use netplot, instead).
• Function netplot:
  • Self-created function using the R package igraph. This function creates
    the network plot.

v0.3.0

• Function baseline_model:
  • processes the elements in the argument base_risk for a fixed, random or
    predicted baseline model and passes the output to run_model or run_metareg to
    obtain the absolute risks for all interventions.
  • when a predicted baseline model is conducted, it returns a forest plot with
    the trial-specific and summary probability of an event for the selected
    reference intervention.
• Function forestplot_metareg:
  • upgraded plot that presents two forest plots side-by-side: (i) one on
    estimation and prediction from network meta-analysis and network
    meta-regression for a selected comparator intervention (allows comparison of
    these two analyses), and (ii) one on SUCRA values from both analyses.
    Both forest plots present results from network meta-regression for a selected
    value of the investigated covariate.
• Function league_table_absolute_user:
  • (only for binary outcome) yields the same graph with the league_table_absolute function,
    but the input is not rnmamod object: the user defines the input and it
    includes the summary effect and corresponding (credible or confidence)
    interval for comparisons with a reference intervention. These results stem
    from a network meta-analysis conducted using another R-package or statistical
    software.
• Function robustness_index_user:
  • calculates the robustness index for a sensitivity analysis performed using
    the R-package netmeta or metafor. The user defines the input and the
    function returns the robustness index. This function returns the same output
    with the robustness_index function.
• Function trans_quality:
  • classifies a systematic review with multiple interventions as having low,
    unclear or high quality regarding the transitivity evaluation based on five
    quality criteria.

v0.2.0

1. Typos and links (for functions and packages) in the documentation are corrected.
2. Function run_model:

• allows the user to define the reference intervention of the network via the argument ref;
• (only for binary outcome) estimates the absolute risks for all non-reference interventions using a selected baseline risk for the reference intervention (argument base_risk);
• (only for binary outcome) estimates the relative risks and risk difference as functions of the estimated absolute risks.

3. Function league_table_absolute:

• (only for binary outcome) it presents the absolute risks per 1000 participants in the main diagonal, the odds ratio on the upper off-diagonals, and the risk difference per 1000 participants in the lower off-diagonal;
• allow the user to select the interventions to present via the argument show (ideal for very large networks that make the league table cluttered).

4. Functions league_heatmap and league_heatmap_pred:

• allow the user to select the interventions to present via the argument show (ideal for very large networks that make the league table cluttered);
• allow the user to illustrate the results of two outcomes for the same model (i.e. via run_model or run_metareg) or the results of two models on the same outcome (applicable for: (i) run_model versus run_metareg, and (ii) run_model versus run_series_meta).

5. Functions series_meta_plot and nodesplit_plot:

• present the extent of heterogeneity in the forest plot of tau using different colours for low, reasonable, fairly high, and fairly extreme tau (the classification has been suggested by Spiegelhalter et al., 2004; ISBN 0471499757).