FDA Table 4

Patient Disposition, Pooled Analyses

                          A: Drug X    B: Placebo   C: Combination   Risk Difference (%) (95% CI)
                           (N=134)      (N=134)        (N=132)                 (N=268)           
—————————————————————————————————————————————————————————————————————————————————————————————————
Patients randomized       65 (48.5%)   54 (40.3%)     59 (44.7%)          -8.2 (-20.1 - 3.6)     
ITT/mITT population       134 (100%)   134 (100%)     132 (100%)           0.0 (0.0 - 0.0)       
Safety population         134 (100%)   134 (100%)     132 (100%)           0.0 (0.0 - 0.0)       
Per-protocol population   65 (48.5%)   54 (40.3%)     59 (44.7%)          -8.2 (-20.1 - 3.6)     
Discontinued study drug   42 (31.3%)   40 (29.9%)     38 (28.8%)          -1.5 (-12.5 - 9.5)     
  Adverse event            3 (2.2%)     6 (4.5%)       5 (3.8%)            2.2 (-2.1 - 6.5)      
  Lack of efficacy         2 (1.5%)     2 (1.5%)       3 (2.3%)            0.0 (-2.9 - 2.9)      
  Protocol deviation          0            0              0                0.0 (0.0 - 0.0)       
  Death                   25 (18.7%)   23 (17.2%)     22 (16.7%)          -1.5 (-10.7 - 7.7)     
  Withdrawal by subject       0            0              0                0.0 (0.0 - 0.0)       
  Other                    3 (2.2%)     4 (3.0%)       3 (2.3%)            0.7 (-3.1 - 4.6)      
Discontinued study        42 (31.3%)   40 (29.9%)     38 (28.8%)          -1.5 (-12.5 - 9.5)     
  Death                   25 (18.7%)   23 (17.2%)     22 (16.7%)          -1.5 (-10.7 - 7.7)     
  Lost to follow-up           0            0              0                0.0 (0.0 - 0.0)       
  Withdrawal by subject       0            0              0                0.0 (0.0 - 0.0)       
  Physician decision          0            0              0                0.0 (0.0 - 0.0)       
  Protocol deviation       5 (3.7%)     3 (2.2%)       4 (3.0%)           -1.5 (-5.6 - 2.6)      
  Other                    3 (2.2%)     4 (3.0%)       3 (2.3%)            0.7 (-3.1 - 4.6)      
# Load Libraries & Data
library(cardinal)
library(dplyr)

adsl <- random.cdisc.data::cadsl %>%
  mutate(test = rbinom(400, 1, 0.5)) %>%
  mutate(
    RANDFL = ifelse(test == 0, "N", "Y"),
    PPROTFL = ifelse(test == 0, "N", "Y"),
    DCSREAS = if_else(DCSREAS %in% c(
      "ADVERSE EVENT", "LACK OF EFFICACY", "PROTOCOL VIOLATION",
      "DEATH", "WITHDRAWAL BY PARENT/GUARDIAN"
    ), DCSREAS, "OTHER")
  )

# Output Table
risk_diff <- list(arm_x = "B: Placebo", arm_y = "A: Drug X") # optional
tbl <- make_table_04(
  df = adsl, pop_vars = c("RANDFL", "ITTFL", "SAFFL", "PPROTFL"),
  lbl_pop_vars = c("Patients randomized", "ITT/mITT population", "Safety population", "Per-protocol population"),
  risk_diff = risk_diff
)
tbl

make_table_04()

Required variables:

  • df: USUBJID, EOTSTT, DCSREAS, EOSSTT, DCSREAS and the variables specified by arm_var and pop_var.

  • alt_counts_df (if specified): USUBJID, and the variable specified by arm_var and pop_var.

Argument Description Default
df (data.frame) Dataset (typically ADSL) required to build table. No default
alt_counts_df (character) Alternative dataset used only to calculate column counts. NULL
arm_var (character) Arm variable used to split table into columns. "ARM"
pop_vars (named list of character) List of Population variables to use in Table c("SAFFL")
lbl_pop_vars (named list of character) List of row labels for populations defined by pop_vars c("Safety population")
lbl_overall (character) If specified, an overall column will be added to the table with the given value as the column label. NULL
prune_0 (flag) Whether all-zero rows should be removed from the table. FALSE
annotations (named list of character) List of annotations to add to the table. Valid annotation types are title, subtitles, main_footer, and prov_footer. Each name-value pair should use the annotation type as name and the desired string as value. NULL

Source code for this function is available here.