The nodesplit_summary class

The nodesplit_summary class contains posterior summary statistics for node-splitting models, as a result of calling summary() on a nma_nodesplit or nma_nodesplit_df object.

Details

Objects of class nodesplit_summary are tibble data frames, with one row for each node-split comparison and columns:

trt1, trt2

Treatments forming the comparison

summary

A list column containing nma_summary objects with the posterior summaries and draws for each of the node-splitting parameters

p_value

Bayesian p-value for inconsistency

dic

A list column containing nma_dic objects, giving the model fit statistics

The parameters included in summary are:

d_net

Network estimate from the corresponding consistency model, if available

d_dir

Direct estimate from the node-splitting model

d_ind

Indirect estimate from the node-splitting model

omega

Inconsistency factor \(\omega = d_\mathrm{dir} - d_\mathrm{ind}\)

tau

Heterogeneity standard deviation from the node-splitting model, if a random effects model was fitted

tau_consistency

Heterogeneity standard deviation from the corresponding consistency model, if available and if a random effects model was fitted