Tidy coefficient summaries for pooled qgcompmulti multiple-imputation fits

Returns a broom-style coefficient table for the pooled marginal structural model (MSM) coefficients stored in a fitted qgcompmulti_mi object. The core estimate and std.error columns stay on the Rubin-pooled fitting scale, while display columns expose the active estimand scale for reporting.

Usage

# S3 method for class 'qgcompmulti_mi'
tidy(x, conf.int = FALSE, conf.level = 0.95, method = NULL, ...)

Arguments

x

A fitted qgcompmulti_mi object.

conf.int

Logical; if TRUE, add confidence interval columns.

conf.level

Confidence level for interval columns when conf.int = TRUE.

method

Optional interval method. Pooled multiple-imputation coefficient reporting supports only "wald" in Version 0.5.0.

...

Not used.

Value

A data frame with one row per pooled MSM coefficient. Columns estimate, std.error, statistic, df, and p.value are on the fitting scale. Columns display.estimate, display.conf.low, and display.conf.high are on the active estimand scale when present.

Details

Pooled multiple-imputation inference is carried out on the fitting scale. For odds-ratio and rate-ratio estimands, the display columns transform only the final pooled estimates and interval limits. Rubin pooling, standard errors, test statistics, and degrees of freedom remain on the fitting scale.

See also

broom::glance(), coef.qgcompmulti_mi(), vcov.qgcompmulti_mi(), confint.qgcompmulti_mi(), summary.qgcompmulti_mi(), qgcomp.glm.multi.mi()