Example 5: Vital Signs Table
The fifth example produces a summary table of selected Vital Signs for Placebo vs. Treated groups. The report shows statistics for both baseline and after-treatment time points. This example also demonstrates how to use tidyverse functions for data preparation instead of procs.
Program
Note the following about this example:
- The
%eq%operator from the common package allows comparing of a variable that contains many NA values, without error. - The fmtr package provides capabilities to create a
user-defined format, similar to
proc format. - The reporter package supports spanning headers in the report header.
- Notice that sassy functions integrate nicely with
dplyr. You can put a
datastep()right in the middle of a dplyr pipeline and get the best of both worlds!
library(tidyverse)
library(sassy)
options("logr.autolog" = TRUE,
"logr.notes" = FALSE)
# Get path to temp directory
tmp <- tempdir()
# Get path to sample data
pkg <- system.file("extdata", package = "sassy")
# Open log
lgpth <- log_open(file.path(tmp, "example5.log"))
sep("Prepare Data")
# Create libname for csv data
libname(sdtm, pkg, "csv")
put("Join and prepare data")
prep <- sdtm$DM |>
left_join(sdtm$VS, by = c("USUBJID" = "USUBJID")) |>
select(USUBJID, VSTESTCD, VISIT, VISITNUM, VSSTRESN, ARM, VSBLFL) |>
filter(VSTESTCD %in% c("PULSE", "RESP", "TEMP", "DIABP", "SYSBP"),
!(VISIT == "SCREENING" & VSBLFL != "Y")) |>
arrange(USUBJID, VSTESTCD, VISITNUM) |>
group_by(USUBJID, VSTESTCD) |>
datastep(retain = list(BSTRESN = 0), {
# Combine treatment groups
# And distingish baseline time points
if (ARM == "ARM A") {
if (VSBLFL %eq% "Y") {
GRP <- "A_BASE"
} else {
GRP <- "A_TRT"
}
} else {
if (VSBLFL %eq% "Y") {
GRP <- "O_BASE"
} else {
GRP <- "O_TRT"
}
}
# Populate baseline value
if (first.)
BSTRESN = VSSTRESN
}) |>
ungroup()
put("Get population counts")
pop_A <- prep |> select(USUBJID, GRP) |> filter(GRP == "A_BASE") |>
distinct() |> count() |> deframe() |> put()
pop_O <- prep |> select(USUBJID, GRP) |> filter(GRP == "O_BASE") |>
distinct() |> count() |> deframe() |> put()
put("Prepare final data frame")
final <- prep |>
select(VSTESTCD, GRP, VSSTRESN, BSTRESN) |>
group_by(VSTESTCD, GRP) |>
summarize(Mean = fmt_mean_sd(VSSTRESN),
Median = fmt_median(VSSTRESN),
Quantiles = fmt_quantile_range(VSSTRESN),
Range = fmt_range(VSSTRESN)) |>
ungroup() |>
pivot_longer(cols = c(Mean, Median, Quantiles, Range),
names_to = "stats",
values_to = "values") |>
pivot_wider(names_from = GRP,
values_from = values) |>
put()
sep("Create formats")
# Vital sign lookup format
vs_fmt <- c(PULSE = "Pulse",
TEMP = "Temperature °C",
RESP = "Respirations/min",
SYSBP = "Systolic Blood Pressure",
DIABP = "Diastolic Blood Pressure") |>
put()
# Statistics user-defined format
stat_fmt <- value(condition(x == "Mean", "Mean (SD)"),
condition(x == "Quantiles", "Q1 - Q3")) |>
put()
sep("Create Report")
# Apply sort
final <- final |>
mutate(VSTESTCD = factor(VSTESTCD, levels = names(vs_fmt))) |>
arrange(VSTESTCD)
# Define table object
tbl <- create_table(final, borders = "bottom") |>
spanning_header(A_BASE, A_TRT, "Placebo", n = pop_A) |>
spanning_header(O_BASE, O_TRT, "Treated", n = pop_O) |>
column_defaults(width = 1.25, align = "center") |>
stub(c(VSTESTCD, stats), width = 2.5) |>
define(VSTESTCD, "Vital Sign", format = vs_fmt,
blank_after = TRUE, dedupe = TRUE, label_row = TRUE) |>
define(stats, indent = .25, format = stat_fmt) |>
define(A_BASE, "Baseline") |>
define(A_TRT, "After Treatment") |>
define(O_BASE, "Baseline") |>
define(O_TRT, "After Treatment")
# Construct output path
pth <- file.path(tmp, "output/t_vs.pdf")
# Define report object
rpt <- create_report(pth, output_type = "PDF", font = "Times",
font_size = 11) |>
page_header("Sponsor: Company", "Study: ABC") |>
titles("Table 4.0", "Selected Vital Signs", bold = TRUE, font_size = 12) |>
add_content(tbl, align = "center") |>
page_footer(Sys.time(), "CONFIDENTIAL", "Page [pg] of [tpg]")
# Write report to file system
write_report(rpt)
# Close log
log_close()
# View files
# file.show(pth)
# file.show(lgpth)Output
Here is the output report:
Log
Here is the log from the above example:
=========================================================================
Log Path: C:/Users/User/AppData/Local/Temp/Rtmpum5T6o/log/example5.log
Working Directory: C:/packages/Testing
User Name: User
R Version: 4.0.3 (2020-10-10)
Machine: DESKTOP-1F27OR8 x86-64
Operating System: Windows 10 x64 build 18363
Log Start Time: 2021-01-05 08:13:13
=========================================================================
=========================================================================
Prepare Data
=========================================================================
# library 'sdtm': 8 items
- attributes: csv not loaded
- path: C:/Users/User/Documents/R/win-library/4.0/sassy/extdata
- items:
Name Extension Rows Cols Size LastModified
1 AE csv 150 27 88.1 Kb 2020-12-27 23:21:55
2 DA csv 3587 18 527.8 Kb 2020-12-27 23:21:55
3 DM csv 87 24 45.2 Kb 2020-12-27 23:21:55
4 DS csv 174 9 33.7 Kb 2020-12-27 23:21:55
5 EX csv 84 11 26 Kb 2020-12-27 23:21:55
6 IE csv 2 14 13 Kb 2020-12-27 23:21:55
7 SV csv 685 10 69.9 Kb 2020-12-27 23:21:55
8 VS csv 3358 17 467 Kb 2020-12-27 23:21:55
Join and prepare data
left_join: added 18 columns (STUDYID.x, DOMAIN.x, STUDYID.y, DOMAIN.y, VSSEQ, …)
> rows only in x 0
> rows only in y ( 0)
> matched rows 3,358 (includes duplicates)
> =======
> rows total 3,358
select: dropped 33 variables (STUDYID.x, DOMAIN.x, SUBJID, RFSTDTC, RFENDTC, …)
filter: removed 590 rows (18%), 2,768 rows remaining
group_by: 2 grouping variables (USUBJID, VSTESTCD)
datastep: columns decreased from 7 to 9
ungroup: no grouping variables
# A tibble: 2,768 x 9
USUBJID VSTESTCD VISIT VISITNUM VSSTRESN ARM VSBLFL BSTRESN GRP
<chr> <chr> <chr> <dbl> <dbl> <chr> <chr> <dbl> <chr>
1 ABC-01-049 DIABP DAY 1 1 76 ARM D Y 76 O_BASE
2 ABC-01-049 DIABP WEEK 2 2 66 ARM D <NA> 76 O_TRT
3 ABC-01-049 DIABP WEEK 4 4 84 ARM D <NA> 76 O_TRT
4 ABC-01-049 DIABP WEEK 6 6 68 ARM D <NA> 76 O_TRT
5 ABC-01-049 DIABP WEEK 8 8 80 ARM D <NA> 76 O_TRT
6 ABC-01-049 DIABP WEEK 12 12 70 ARM D <NA> 76 O_TRT
7 ABC-01-049 DIABP WEEK 16 16 70 ARM D <NA> 76 O_TRT
8 ABC-01-049 PULSE DAY 1 1 84 ARM D Y 84 O_BASE
9 ABC-01-049 PULSE WEEK 2 2 84 ARM D <NA> 84 O_TRT
10 ABC-01-049 PULSE WEEK 4 4 76 ARM D <NA> 84 O_TRT
# ... with 2,758 more rows
Get population counts
select: dropped 7 variables (VSTESTCD, VISIT, VISITNUM, VSSTRESN, ARM, …)
filter: removed 2,669 rows (96%), 99 rows remaining
distinct: removed 79 rows (80%), 20 rows remaining
count: now one row and one column, ungrouped
[1] 20
select: dropped 7 variables (VSTESTCD, VISIT, VISITNUM, VSSTRESN, ARM, …)
filter: removed 2,435 rows (88%), 333 rows remaining
distinct: removed 266 rows (80%), 67 rows remaining
count: now one row and one column, ungrouped
[1] 67
Prepare final data frame
select: dropped 5 variables (USUBJID, VISIT, VISITNUM, ARM, VSBLFL)
group_by: 2 grouping variables (VSTESTCD, GRP)
summarize: now 20 rows and 6 columns, one group variable remaining (VSTESTCD)
ungroup: no grouping variables
pivot_longer: reorganized (Mean, Median, Quantiles, Range) into (stats, values) [was 20x6, now 80x4]
pivot_wider: reorganized (GRP, values) into (A_BASE, A_TRT, O_BASE, O_TRT) [was 80x4, now 20x6]
# A tibble: 20 x 6
VSTESTCD stats A_BASE A_TRT O_BASE O_TRT
<chr> <chr> <chr> <chr> <chr> <chr>
1 DIABP Mean 77.2 (10.7) 77.1 (8.1) 77.5 (8.1) 76.9 (9.2)
2 DIABP Median 78.5 78.0 78.0 78.0
3 DIABP Quantiles 70.0 - 82.5 72.0 - 82.0 70.2 - 82.0 70.0 - 83.8
4 DIABP Range 54 - 96 50 - 98 61 - 95 50 - 104
5 PULSE Mean 72.8 (10.1) 74.6 (10.2) 73.6 (9.7) 74.0 (9.9)
6 PULSE Median 72.0 74.0 72.0 72.0
7 PULSE Quantiles 66.0 - 76.0 67.0 - 80.0 67.5 - 79.5 66.0 - 80.0
8 PULSE Range 60 - 103 54 - 102 52 - 100 50 - 109
9 RESP Mean 16.5 (2.5) 15.8 (3.4) 16.0 (3.1) 15.4 (3.4)
10 RESP Median 16.0 16.0 16.0 16.0
11 RESP Quantiles 16.0 - 18.0 16.0 - 18.0 15.5 - 18.0 12.0 - 18.0
12 RESP Range 12 - 20 8 - 24 8 - 22 8 - 24
13 SYSBP Mean 128.2 (16.7) 130.4 (17.6) 126.7 (15.5) 125.9 (15.2)
14 SYSBP Median 125.5 126.0 123.0 124.0
15 SYSBP Quantiles 117.5 - 140.0 118.0 - 140.0 116.0 - 138.0 115.2 - 135.0
16 SYSBP Range 98 - 161 95 - 184 100 - 164 82 - 180
17 TEMP Mean 36.5 (0.4) 36.5 (0.3) 36.4 (0.5) 36.3 (0.4)
18 TEMP Median 36.4 36.5 36.4 36.3
19 TEMP Quantiles 36.2 - 36.9 36.2 - 36.6 36.2 - 36.6 36.1 - 36.5
20 TEMP Range 35.9 - 37.4 35.8 - 37.3 35.3 - 39.8 34.4 - 38.2
=========================================================================
Create formats
=========================================================================
PULSE TEMP RESP SYSBP
"Pulse" "Temperature °C" "Respirations/min" "Systolic Blood Pressure"
DIABP
"Diastolic Blood Pressure"
# A user-defined format: 2 conditions
Name Type Expression Label Order
1 x U x == "Mean" Mean (SD) NA
2 x U x == "Quantiles" Q1 - Q3 NA
=========================================================================
Create Report
=========================================================================
mutate: converted 'VSTESTCD' from character to factor (0 new NA)
# A report specification: 1 pages
- file_path: 'C:\Users\User\AppData\Local\Temp\Rtmpum5T6o/output/t_vs.rtf'
- output_type: RTF
- units: inches
- orientation: landscape
- line size/count: 108/48
- page_header: left=Sponsor: Company right=Study: ABC
- title 1: 'Table 4.0'
- title 2: 'Selected Vital Signs'
- page_footer: left=2021-01-05 08:13:45 center=CONFIDENTIAL right=Page [pg] of [tpg]
- content:
# A table specification:
- data: tibble 'final' 20 rows 6 cols
- show_cols: all
- use_attributes: all
- spanning_header: from='A_BASE' to='A_TRT' 'Placebo' level=1
- spanning_header: from='O_BASE' to='O_TRT' 'Treated' level=1
- stub: VSTESTCD stats width=2.5 align='left'
- define: VSTESTCD 'Vital Sign' dedupe='TRUE'
- define: stats
- define: A_BASE 'Baseline'
- define: A_TRT 'After Treatment'
- define: O_BASE 'Baseline'
- define: O_TRT 'After Treatment'
=========================================================================
Log End Time: 2021-01-05 08:13:49
Log Elapsed Time: 0 00:00:36
=========================================================================