TSFVIT01

Mean Change From Baseline in Blood Pressure Over Time

Output

Code
# Program Name:              tsfvit01

# Prep environment:

library(envsetup)
library(tern)
library(dplyr)
library(rtables)
library(junco)

# Define script level parameters:

tblid <- "TSFVIT01"
fileid <- tblid
titles <- get_titles_from_file(input_path = '../../_data/', tblid)
string_map <- default_str_map

popfl <- "SAFFL"
trtvar <- "TRT01A"
ctrl_grp <- "Placebo"


# Note on ancova parameter
# when ancova = TRUE
# ancova model will be used to calculate all mean/mean change columns
# not just those from the Difference column
# model specification
summ_vars <- list(arm = trtvar, covariates = NULL)

# when ancova = FALSE, all mean/mean change columns will be from descriptive stats
# for the difference column descriptive stats will be based upon two-sample t-test
ancova <- FALSE


comp_btw_group <- TRUE

selparamcd <- c("DIABP", "SYSBP")
# to Update the timepoints to what is relevant for your study
# For the trial used for the shell program, the screening timepoint is the timepoint considered as baseline
# see further, an alternative method to identify all non-unscheduled visits based upon data
selvisit <- c("Screening", "Cycle 02", "Cycle 03", "Cycle 04")

### as in dataset, order is important for later processing
### not automated, hard coded approach for ease of reading
### ideally the datasets already contain the appropriate case, to ensure units are in proper case
sel_param <- c(
  "Systolic Blood Pressure (mmHg)",
  "Diastolic Blood Pressure (mmHg)"
)
sel_param_case <- c(
  "Systolic blood pressure (mmHg)",
  "Diastolic blood pressure (mmHg)"
)

# Process Data:

adsl <- pharmaverseadamjnj::adsl %>%
  filter(.data[[popfl]] == "Y") %>%
  select(
    STUDYID,
    USUBJID,
    all_of(c(popfl, trtvar)),
    SEX,
    AGEGR1,
    RACE,
    ETHNIC,
    AGE
  )

adsl$colspan_trt <- factor(
  ifelse(adsl[[trtvar]] == ctrl_grp, " ", "Active Study Agent"),
  levels = c("Active Study Agent", " ")
)

adsl$rrisk_header <- "Difference in Mean Change (95% CI)"
adsl$rrisk_label <- paste(adsl[[trtvar]], paste("vs", ctrl_grp))

colspan_trt_map <- create_colspan_map(
  adsl,
  non_active_grp = ctrl_grp,
  non_active_grp_span_lbl = " ",
  active_grp_span_lbl = "Active Study Agent",
  colspan_var = "colspan_trt",
  trt_var = trtvar
)

advs00 <- pharmaverseadamjnj::advs %>%
  select(
    USUBJID,
    AVISITN,
    AVISIT,
    PARAMCD,
    PARAM,
    AVAL,
    BASE,
    CHG,
    starts_with("ANL"),
    ABLFL,
    APOBLFL,
    VSSTAT
  ) %>%
  inner_join(adsl)

# selection of all non-unscheduled visits from data
visits <- advs00 %>%
  select(AVISIT) %>%
  filter(!grepl("UNSCHEDULED", toupper(AVISIT)))

visits$AVISIT <- droplevels(visits$AVISIT)
selvisit_data <- levels(visits$AVISIT)

### if preferred to get it from data, rather than hardcoded list of visits
# selvisit <- selvisit_data

## retrieve the precision of AVAL on the input dataset
## review outcome and make updates manually if needed
## the precision variable will be used for the parameter-based formats in layout

## decimal = 4 is a cap in this derivation: if decimal precision of variable > decimal, the result will end up as decimal
## eg if AVAL has precision of 6 for parameter x, and decimal = 4, the resulting decimal value for parameter x is 4

## note that precision is on the raw values, as we are presenting mean/ci, and extra digit will be added
## eg precision = 2 will result in mean/ci format xx.xxx (xx.xxx, xx.xxx)

vs_precision <- tidytlg:::make_precision_data(
  df = advs00,
  decimal = 3,
  precisionby = "PARAMCD",
  precisionon = "AVAL"
)

### data preparation

filtered_advs <- advs00 %>%
  filter(PARAMCD %in% selparamcd) %>%
  filter(AVISIT %in% selvisit) %>%
  ### unique record per timepoint:
  filter(ANL02FL == "Y" & (ABLFL == "Y" | APOBLFL == "Y"))

#### perform check on unique record per subject/param/timepoint
check_unique <- filtered_advs %>%
  group_by(USUBJID, PARAMCD, AVISIT) %>%
  mutate(n_recsub = n()) %>%
  filter(n_recsub > 1)

if (nrow(check_unique) > 0) {
  stop(
    "Your input dataset needs extra attention, as some subjects have more than one record per parameter/visit"
  )
  ### you will run into issues with fraction portion in count_denom_fraction, as count > denom, and fraction > 1 if you don't adjust your input dataset

  # Possible extra derivation - just to ensure program can run without issues
  ### Study team is responsible for adding this derivation onto ADaM dataset and ensure proper derivation rule for ANL02FL is implemented !!!!!!!!!!
  filtered_advsx <- advs00 %>%
    filter(PARAMCD %in% selparamcd) %>%
    filter(AVISIT %in% selvisit) %>%
    ### unique record per timepoint:
    filter(ANL02FL == "Y" & (ABLFL == "Y" | APOBLFL == "Y")) %>%
    group_by(USUBJID, PARAM, AVISIT) %>%
    mutate(n_sub = n()) %>%
    arrange(USUBJID, PARAM, AVISIT, ADT) %>%
    mutate(i = vctrs::vec_group_id(ADT)) %>%
    mutate(
      ANL02FL = case_when(
        n_sub == 1 ~ "Y",
        i == 1 ~ "Y"
      )
    ) %>%
    select(-c(i, n_sub)) %>%
    ungroup()

  filtered_advs <- filtered_advsx %>%
    filter(PARAMCD %in% selparamcd) %>%
    filter(AVISIT %in% selvisit) %>%
    ### unique record per timepoint:
    filter(ANL02FL == "Y" & (ABLFL == "Y" | APOBLFL == "Y"))

  ## now your data should contain 1 record per subject per parameter
}

### for denominator per timepoint: all records from advs on this timepoint: ignoring anl01fl/anl02fl/param
filtered_advs_timepoints <- unique(
  advs00 %>%
    filter(AVISIT %in% selvisit) %>%
    select(USUBJID, AVISITN, AVISIT)
) %>%
  inner_join(adsl)

filtered_advs$PARAM <- factor(
  as.character(filtered_advs$PARAM),
  levels = sel_param,
  labels = sel_param_case
)

params <- unique(filtered_advs %>% select(PARAMCD, PARAM))

filtered_advs_timepoints <- filtered_advs_timepoints %>%
  mutate(dummy_join = 1) %>%
  full_join(
    params %>% mutate(dummy_join = 1),
    relationship = "many-to-many"
  ) %>%
  select(-dummy_join)

### identify subjects in filtered_advs_timepoints and not in filtered_advs

extra_sub <- anti_join(filtered_advs_timepoints, filtered_advs)

### only add these extra_sub to
### this will ensure we still meet the one record per subject per timepoint
### this will ensure length(x) can be used for the denominator derivation inside summarize_aval_chg_diff function

filtered_advs <- bind_rows(filtered_advs, extra_sub) %>%
  arrange(USUBJID, PARAM, AVISITN)

filtered_advs <- filtered_advs %>%
  inner_join(vs_precision, by = "PARAMCD")

### important: previous actions lost the label of variables
### in order to be able to use obj_label(filtered_advs$PARAM) in layout, need to redefine the label
filtered_advs <- var_relabel_list(filtered_advs, var_labels(advs00, fill = T))

# Define layout and build table:

summ_vars <- list(arm = trtvar, covariates = NULL)
ref_path <- c("colspan_trt", " ", trtvar, ctrl_grp)
multivars <- c("AVAL", "AVAL", "CHG")

extra_args_3col <- list(
  format_na_str = rep("NA", 3),
  d = "decimal",
  variables = summ_vars,
  ref_path = ref_path,
  ancova = ancova,
  comp_btw_group = comp_btw_group,
  multivars = multivars
)

lyt <- basic_table(show_colcounts = FALSE, colcount_format = "N=xx") %>%
  ### first columns
  split_cols_by(
    "colspan_trt",
    split_fun = trim_levels_to_map(map = colspan_trt_map),
    show_colcounts = FALSE
  ) %>%
  split_cols_by(trtvar, show_colcounts = TRUE, colcount_format = "N=xx") %>%
  split_rows_by(
    "PARAM",
    label_pos = "topleft",
    split_label = "Blood Pressure",
    section_div = " ",
    split_fun = drop_split_levels
  ) %>%
  ## note the child_labels = hidden for AVISIT, these labels will be taken care off by
  ## applying function summarize_aval_chg_diff further in the layout
  split_rows_by(
    "AVISIT",
    label_pos = "topleft",
    split_label = "Study Visit",
    split_fun = drop_split_levels,
    child_labels = "hidden"
  ) %>%
  ## set up a 3 column split
  split_cols_by_multivar(
    multivars,
    varlabels = c(
      "n/N (%)",
      "Mean (95% CI)",
      "Mean Change From Baseline (95% CI)"
    )
  ) %>%
  ### restart for the rrisk_header columns - note the nested = FALSE option
  ### also note the child_labels = "hidden" in both PARAM and AVISIT
  split_cols_by("rrisk_header", nested = FALSE) %>%
  split_cols_by(
    trtvar,
    split_fun = remove_split_levels(ctrl_grp),
    labels_var = "rrisk_label",
    show_colcounts = TRUE,
    colcount_format = "N=xx"
  ) %>%
  ### difference columns : just 1 column & analysis needs to be done on change
  split_cols_by_multivar(multivars[3], varlabels = c(" ")) %>%
  ### the variable passed here in analyze is not used (STUDYID), it is a dummy var passing,
  ### the function summarize_aval_chg_diff grabs the required vars from cols_by_multivar calls
  analyze(
    "STUDYID",
    afun = a_summarize_aval_chg_diff_j,
    extra_args = extra_args_3col
  )

result <- build_table(lyt, filtered_advs, alt_counts_df = adsl)

# Post-Processing:

remove_col_count2 <- function(result, string = paste("vs", ctrl_grp)) {
  mcdf <- make_col_df(result, visible_only = FALSE)
  mcdfsel <- mcdf %>%
    filter(stringr::str_detect(toupper(label), toupper(string))) %>%
    pull(path)

  for (i in seq_along(mcdfsel)) {
    facet_colcount(result, mcdfsel[[i]]) <- NA
  }

  return(result)
}

result <- remove_col_count2(result)

# Add titles and footnotes:

result <- set_titles(result, titles)

# Convert to tbl file and output table:

tt_to_tlgrtf(string_map = string_map, tt = 
  result,
  file = fileid,
  nosplitin = list(cols = c(trtvar, "rrisk_header")),
  orientation = "landscape"
)

TSFVIT01: Mean Change From Baseline in Blood Pressure Over Time; Safety Analysis Set (Study jjcs - core)

Active Study Agent

Difference in Mean Change (95% CI)

Xanomeline High Dose

Xanomeline Low Dose

Placebo

Xanomeline High Dose vs Placebo

Xanomeline Low Dose vs Placebo

Blood Pressure

N=53

N=73

N=59

Study Visit

n/N (%)

Mean (95% CI)

Mean Change From Baseline (95% CI)

n/N (%)

Mean (95% CI)

Mean Change From Baseline (95% CI)

n/N (%)

Mean (95% CI)

Mean Change From Baseline (95% CI)

Systolic blood pressure (mmHg)

Screening

53/53 (100.0%)

138.2390 (133.5179, 142.9601)

73/73 (100.0%)

138.7352 (134.5082, 142.9621)

59/59 (100.0%)

135.7627 (131.0746, 140.4508)

Cycle 02

53/53 (100.0%)

131.7170 (127.5681, 135.8659)

-6.5220 (-10.7477, -2.2963)

71/71 (100.0%)

135.3192 (131.2528, 139.3857)

-3.0986 (-6.2424, 0.0452)

58/58 (100.0%)

131.3966 (126.7814, 136.0117)

-4.2241 (-7.5565, -0.8917)

-2.2979 (-7.6221, 3.0264)

1.1255 (-3.4113, 5.6624)

Cycle 03

53/53 (100.0%)

130.7421 (126.5700, 134.9143)

-7.4969 (-11.4782, -3.5155)

53/53 (100.0%)

137.6478 (132.7251, 142.5705)

-0.1321 (-4.2817, 4.0175)

56/56 (100.0%)

131.3512 (126.5164, 136.1860)

-4.0476 (-7.6489, -0.4464)

-3.4492 (-8.7568, 1.8584)

3.9155 (-1.5172, 9.3483)

Cycle 04

51/51 (100.0%)

129.9477 (125.8664, 134.0290)

-8.9869 (-13.2844, -4.6895)

47/47 (100.0%)

135.7872 (130.2234, 141.3510)

-0.6809 (-4.1263, 2.7646)

51/51 (100.0%)

131.6340 (126.4963, 136.7717)

-2.5686 (-6.1802, 1.0430)

-6.4183 (-11.9651, -0.8715)

1.8878 (-3.0400, 6.8156)

Diastolic blood pressure
 (mmHg)

Screening

53/53 (100.0%)

79.3333 (76.6476, 82.0190)

73/73 (100.0%)

76.6256 (74.4932, 78.7579)

59/59 (100.0%)

76.2599 (73.5325, 78.9872)

Cycle 02

53/53 (100.0%)

77.1069 (74.5509, 79.6630)

-2.2264 (-4.9097, 0.4569)

71/71 (100.0%)

76.7324 (74.3843, 79.0804)

0.2207 (-1.4974, 1.9388)

58/58 (100.0%)

73.7816 (71.1451, 76.4181)

-2.4828 (-4.9920, 0.0264)

0.2563 (-3.3763, 3.8889)

2.7034 (-0.3118, 5.7186)

Cycle 03

53/53 (100.0%)

76.0252 (73.5603, 78.4900)

-3.3082 (-6.1264, -0.4899)

53/53 (100.0%)

76.8553 (74.3209, 79.3898)

0.2138 (-1.9788, 2.4065)

56/56 (100.0%)

74.2976 (71.2455, 77.3498)

-1.7619 (-4.2062, 0.6823)

-1.5463 (-5.2349, 2.1424)

1.9757 (-1.2708, 5.2222)

Cycle 04

51/51 (100.0%)

75.3137 (72.6005, 78.0269)

-4.2614 (-6.7224, -1.8005)

47/47 (100.0%)

75.1844 (72.5599, 77.8089)

-0.5177 (-2.3210, 1.2855)

51/51 (100.0%)

74.3399 (71.4934, 77.1864)

-2.1176 (-4.3514, 0.1161)

-2.1438 (-5.4270, 1.1394)

1.5999 (-1.2358, 4.4356)

Download RTF file