Add support for group-level estimates from coxme

R/extract_random_parameters.R

@@ -122,3 +122,50 @@
 .extract_random_parameters.MixMod <- function(model, ...) {
   NULL
 }
+
+
+.extract_random_parameters.coxme <- function(model, ci = NULL, effects = "random", ...) {
+  insight::check_if_installed("lme4")
+
+  # extract random effects
+  re_grp <- insight::find_random(model, split_nested = TRUE, flatten = TRUE)
+
+  ## TODO: this currently only works for one grouping variable
+  if (length(re_grp) > 1) {
+    re_grp <- re_grp[1]
+  }
+
+  out <- as.data.frame(lme4::ranef(model, condVar = TRUE), stringsAsFactors = FALSE)
+  out$grp <- unique(insight::get_data(model)[[re_grp]])
+  out <- datawizard::reshape_longer(out, select = -"grp")
+  out <- datawizard::data_separate(out, select = "name", new_columns = c("grpvar", "term"))
+  out <- datawizard::data_arrange(out, select = c("grpvar", "term"))
+  out <- out[c("grpvar", "term", "grp", "value")]
+  out$condsd <- NA
+
+  colnames(out) <- c("Group", "Parameter", "Level", "Coefficient", "SE")
+
+  # coerce to character
+  out$Parameter <- as.character(out$Parameter)
+  out$Level <- as.character(out$Level)
+  out$Group <- as.character(out$Group)
+  out$Effects <- "random"
+
+  out$CI_low <- out$CI_high <- NA
+  ci_cols <- c("CI_low", "CI_high")
+
+  stat_column <- gsub("-statistic", "", insight::find_statistic(model), fixed = TRUE)
+
+  # to match rbind
+  out[[stat_column]] <- NA
+  out$df_error <- NA
+  out$p <- NA
+
+  out <- out[c("Parameter", "Level", "Coefficient", "SE", ci_cols, stat_column, "df_error", "p", "Effects", "Group")]
+
+  if (effects == "random") {
+    out[c(stat_column, "df_error", "p")] <- NULL
+  }
+
+  out
+}

R/methods_coxme.R

@@ -1,3 +1,63 @@
+#' @export
+model_parameters.merMod <- function(model,
+                                    ci = 0.95,
+                                    ci_method = NULL,
+                                    bootstrap = FALSE,
+                                    iterations = 1000,
+                                    standardize = NULL,
+                                    effects = "fixed",
+                                    group_level = FALSE,
+                                    exponentiate = FALSE,
+                                    p_adjust = NULL,
+                                    vcov = NULL,
+                                    vcov_args = NULL,
+                                    include_info = getOption("parameters_info", FALSE),
+                                    keep = NULL,
+                                    drop = NULL,
+                                    verbose = TRUE,
+                                    ...) {
+  # which component to return?
+  effects <- insight::validate_argument(
+    effects,
+    c("fixed", "total", "random_total", "all")
+  )
+
+  if (effects == "fixed") {
+    return(model_parameters.default(
+      model,
+      ci = ci,
+      ci_method = ci_method,
+      bootstrap = bootstrap,
+      iterations = iterations,
+      standardize = standardize,
+      exponentiate = exponentiate,
+      p_adjust = p_adjust,
+      vcov = vcov,
+      vcov_args = vcov_args,
+      include_info = include_info,
+      keep = keep,
+      drop = drop,
+      verbose = verbose,
+      ...
+    ))
+  }
+
+  # for coef(), we don't need all the attributes and just stop here
+  if (effects %in% c("total", "random_total")) {
+    params <- .group_level_total(model)
+    params$Effects <- "total"
+    class(params) <- c("parameters_coef", "see_parameters_coef", class(params))
+    return(params)
+  }
+
+  if (effects %in% c("random", "all") && isTRUE(group_level)) {
+    params_random <- .extract_random_parameters(model, ci = ci, effects = effects)
+  }
+
+  params
+}
+
+
 #' @export
 standard_error.coxme <- function(model, ...) {
   beta_coef <- model$coefficients