Design-level inconsistency diagnostics for network meta-analysis
network_design_inconsistency.RdDesign-level inconsistency diagnostics for network meta-analysis
Value
A list with a design-level Q decomposition, design-specific homogeneity statistics, detach diagnostics, and a design-by-comparison heat matrix.
Examples
# \donttest{
network_design_inconsistency(nma_fit)
#> $Q_decomp
#> component Q df p_value
#> 1 whole_network 0.06798246 2 0.9665800
#> 2 within_designs 0.02500000 1 0.8743671
#> 3 between_designs 0.04298246 1 0.8357583
#>
#> $by_design
#> design n_studies n_effects Q_within df p_value
#> A + B A + B 2 2 0.025 1 0.8743671
#> A + C A + C 1 1 0.000 0 NA
#> B + C B + C 1 1 0.000 0 NA
#>
#> $detach
#> detached_design Q_without_design df p_value delta_Q_between
#> 1 A + B NA NA NA NA
#> 2 A + C 0.04913050 1 0.8245832 0.01885195
#> 3 B + C 0.05290089 1 0.8180903 0.01508156
#>
#> $heat_matrix
#> A vs B A vs C B vs C
#> A + B 0.03404894 0.00000000 0.00000000
#> A + C 0.00000000 0.01885195 0.00000000
#> B + C 0.00000000 0.00000000 0.01508156
#>
# }