Update the R functions to allow outputting spending time

R/as_gt.R

@@ -271,11 +271,17 @@ as_gt.gs_design_summary <- function(
 # get different default columns to display
 gsd_columns <- function(columns, method, x) {
   # set different default columns to display
-  if (is.null(columns)) columns <- c(
-    "Analysis", "Bound", "Z", "Nominal p",
-    sprintf("%s at bound", switch(method, ahr = "~HR", wlr = "~wHR", rd = "~Risk difference")),
-    "Alternate hypothesis", "Null hypothesis"
-  )
+  if (is.null(columns)){
+    columns <- c(
+      "Analysis", "Bound", "Z", "Nominal p",
+      sprintf("%s at bound", switch(method, ahr = "~HR", wlr = "~wHR", rd = "~Risk difference")),
+      "Alternate hypothesis", "Null hypothesis")
+
+    if ("Spending time" %in% names(x)) {
+      columns <- c(columns[1:3], "Spending time", columns[4:length(columns)])
+    }
+  }
+
   # filter the columns to display as the output: if `Probability` is selected to
   # output, transform it to `c("Alternate hypothesis", "Null hypothesis")`
   if (any(i <- columns == "Probability"))

R/gs_design_ahr.R

@@ -337,6 +337,7 @@ gs_design_ahr <- function(
   allout$event <- allout$event * inflac_fct
   allout$n <- allout$n * inflac_fct
 
+
   # Get bounds to output ----
   bound <- allout |>
     select(all_of(c(
@@ -345,6 +346,42 @@ gs_design_ahr <- function(
     ))) |>
     arrange(analysis, desc(bound))
 
+  # Output spending time to the bounds table
+  info0 <- (allout |> filter(bound == "upper"))$info0
+  info <- (allout |> filter(bound == "upper"))$info
+  info0_final <- (allout |> filter(analysis == n_analysis, bound == "upper"))$info0
+  info_final <- (allout |> filter(analysis == n_analysis, bound == "upper"))$info
+
+  bound$spending_time <- NA
+
+  if (identical(upper, gs_spending_bound)) {
+
+    if (!is.null(upar$timing)) {
+      spending_time_upper <- upar$timing
+    } else {
+      spending_time_upper <- info0 / info0_final
+    }
+
+
+    bound$spending_time[which(bound$bound == "upper")] <- spending_time_upper
+  }
+
+  if (identical(lower, gs_spending_bound)) {
+    if (!is.null(lpar$timing)) {
+      spending_time_lower <- lpar$timing
+    } else if (h1_spending) {
+      spending_time_lower <- info / info_final
+    } else if (!h1_spending) {
+      spending_time_lower <- info0 / info0_final
+    }
+
+    bound$spending_time[which(bound$bound == "lower")] <- spending_time_lower
+  }
+
+  if (all(is.na(bound$spending_time))){
+    bound$spending_time <- NULL
+  }
+
   # Get analysis summary to output ----
   analysis <- allout |>
     select(analysis, time, n, event, ahr, theta, info, info0, info_frac) |>

R/summary.R

@@ -104,6 +104,8 @@ summary.fixed_design <- function(object, ...) {
 #'   vector is named, you only have to specify the number of digits for the
 #'   columns you want to be displayed differently than the defaults.
 #' @param bound_names Names for bounds; default is `c("Efficacy", "Futility")`.
+#' @param display_spending_time A logical value (TRUE/FALSE) indicating if the spending time
+#' is summarized in the table.
 #'
 #' @export
 #'
@@ -256,6 +258,7 @@ summary.gs_design <- function(object,
                               col_vars = NULL,
                               col_decimals = NULL,
                               bound_names = c("Efficacy", "Futility"),
+                              display_spending_time = FALSE,
                               ...) {
   x <- object
   x_bound <- x$bound
@@ -318,13 +321,19 @@ summary.gs_design <- function(object,
   )
 
   # Prepare the columns decimals ----
-  default_decimals <- c(NA, NA, 2, if (method != "combo") 4, 4, 4, 4)
+  default_decimals <- c(NA, NA, 2, if (method != "combo") 4, 4, 4, 4, 4)
   default_vars <- c(
     "analysis", "bound", "z",
     sprintf("~%s at bound", switch(method, ahr = "hr", wlr = "whr", rd = "risk difference")),
     "nominal p", "Alternate hypothesis", "Null hypothesis"
   )
 
+  if (display_spending_time) {
+    default_decimals <- c(default_decimals[1:3], 4, default_decimals[4:length(default_decimals)])
+    default_vars <- c(default_vars[1:3], "spending_time", default_vars[4:length(default_vars)])
+  }
+
+
   # Filter columns and update decimal places
   col_decimals <- get_decimals(col_vars, col_decimals, default_vars, default_decimals)
 
@@ -388,7 +397,8 @@ cap_names <- function(x) {
     map, ahr = "AHR", event = "Events", rd = "Risk difference",
     probability = "Alternate hypothesis", probability0 = "Null hypothesis",
     info_frac0 = "Information fraction", info_frac = "Information fraction",
-    event_frac = "Event fraction"
+    event_frac = "Event fraction",
+    spending_time = "Spending time"
   )
   replace_names(x, map)
 }

tests/testthat/_snaps/independent_as_gt.md

@@ -13,7 +13,8 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}Power computed with average hazard ratio method.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} Power computed with average hazard ratio method.\\
     \end{minipage}
     \end{table}
 
@@ -32,7 +33,8 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}Custom footnote.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} Custom footnote.\\
     \end{minipage}
     \end{table}
 
@@ -51,7 +53,8 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}Power for Fleming-Harrington test  FH(0, 0) (logrank) using method of Yung and Liu.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} Power for Fleming-Harrington test  FH(0, 0) (logrank) using method of Yung and Liu.\\
     \end{minipage}
     \end{table}
 
@@ -75,8 +78,9 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
     \end{minipage}
     \end{table}
 
@@ -111,8 +115,9 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
     \end{minipage}
     \end{table}
 
@@ -136,9 +141,10 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
-    \textsuperscript{\textit{3}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
+    \textsuperscript{\textit{3}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -173,10 +179,11 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}this table is generated by gs\_power\_wlr.\\
-    \textsuperscript{\textit{2}}the crossing probability.\\
-    \textsuperscript{\textit{3}}approximate weighted hazard ratio to cross bound.\\
-    \textsuperscript{\textit{4}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} this table is generated by gs\_power\_wlr.\\
+    \textsuperscript{\textit{2}} the crossing probability.\\
+    \textsuperscript{\textit{3}} approximate weighted hazard ratio to cross bound.\\
+    \textsuperscript{\textit{4}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -211,8 +218,9 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}EF is event fraction. AHR is under regular weighted log rank test.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} EF is event fraction. AHR is under regular weighted log rank test.\\
     \end{minipage}
     \end{table}
 
@@ -236,7 +244,8 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
     \end{minipage}
     \end{table}
 
@@ -269,8 +278,9 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Cumulative alpha for final analysis (0.0238) is less than the full alpha (0.025) when the futility bound is non-binding. The smaller value subtracts the probability of crossing a futility bound before crossing an efficacy bound at a later analysis (0.025 - 0.0012 = 0.0238) under the null hypothesis.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Cumulative alpha for final analysis (0.0238) is less than the full alpha (0.025) when the futility bound is non-binding. The smaller value subtracts the probability of crossing a futility bound before crossing an efficacy bound at a later analysis (0.025 - 0.0012 = 0.0238) under the null hypothesis.\\
     \end{minipage}
     \end{table}
 
@@ -305,9 +315,10 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
-    \textsuperscript{\textit{3}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
+    \textsuperscript{\textit{3}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -342,9 +353,10 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
-    \textsuperscript{\textit{3}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
+    \textsuperscript{\textit{3}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -379,10 +391,11 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}this table is generated by gs\_power\_wlr.\\
-    \textsuperscript{\textit{2}}the crossing probability.\\
-    \textsuperscript{\textit{3}}approximate weighted hazard ratio to cross bound.\\
-    \textsuperscript{\textit{4}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} this table is generated by gs\_power\_wlr.\\
+    \textsuperscript{\textit{2}} the crossing probability.\\
+    \textsuperscript{\textit{3}} approximate weighted hazard ratio to cross bound.\\
+    \textsuperscript{\textit{4}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -414,9 +427,10 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}Approximate hazard ratio to cross bound.\\
-    \textsuperscript{\textit{3}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} Approximate hazard ratio to cross bound.\\
+    \textsuperscript{\textit{3}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}
 
@@ -451,8 +465,9 @@
     \bottomrule
     \end{tabular*}
     \begin{minipage}{\linewidth}
-    \textsuperscript{\textit{1}}One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
-    \textsuperscript{\textit{2}}wAHR is the weighted AHR.\\
+    \vspace{.05em}
+    \textsuperscript{\textit{1}} One-sided p-value for experimental vs control treatment. Value < 0.5 favors experimental, > 0.5 favors control.\\
+    \textsuperscript{\textit{2}} wAHR is the weighted AHR.\\
     \end{minipage}
     \end{table}