Small-study asymmetry test for network meta-analysis

Usage

network_bias_test(
  object,
  center = c("network", "direct"),
  comparison_adjusted = TRUE
)

Arguments

object: A fitted nma_mars object.
center: Centering method for comparison adjustment; "network" uses total network estimates and "direct" uses direct estimates.
comparison_adjusted: Logical; if TRUE, subtract the chosen comparison-specific center before testing.

Value

A list containing the regression model and a coefficient table for the intercept asymmetry test.

Examples

# \donttest{
network_bias_test(nma_fit)
#> $method
#> [1] "Comparison-adjusted Egger regression"
#> 
#> $data
#>   study treatment_1 treatment_2 comparison effect variance        se    center
#> 1    S1           A           B     A vs B   0.20     0.04 0.2000000 0.1947368
#> 2    S2           A           C     A vs C   0.35     0.05 0.2236068 0.3192982
#> 3    S3           B           C     B vs C   0.10     0.04 0.2000000 0.1245614
#> 4    S4           A           B     A vs B   0.15     0.06 0.2449490 0.1947368
#>   adjusted_effect precision standard_normal_deviate
#> 1     0.005263158  5.000000              0.02631579
#> 2     0.030701754  4.472136              0.13730242
#> 3    -0.024561404  5.000000             -0.12280702
#> 4    -0.044736842  4.082483             -0.18263739
#> 
#> $model
#> 
#> Call:
#> stats::lm(formula = standard_normal_deviate ~ precision, data = dat)
#> 
#> Coefficients:
#> (Intercept)    precision  
#>     -0.3755       0.0733  
#> 
#> 
#> $coefficients
#>          term    estimate std_error    t_value   p_value
#> 1 (Intercept) -0.37546912 1.0401378 -0.3609802 0.7526783
#> 2   precision  0.07329982 0.2234575  0.3280259 0.7740492
#> 
#> $intercept
#>          term   estimate std_error    t_value   p_value
#> 1 (Intercept) -0.3754691  1.040138 -0.3609802 0.7526783
#> 
# }