Overview

Motivation

A table for descriptive statistics is widely used in medical research and typically be the first table (i.e. table1) for a manuscript.

There are R packages and examples to create the table1:

• table1

For use case in clinical trials, Chapter 4 of the R for clinical study reports and submission contains more details.

The metalite.table1 provide an interactive table1 to enhance the communication between statisticians and clinicians.

Example

We use a subject level dataset r2rtf::r2rtf_adsl to create a example interactive table1.

First, we modify the dataset by

• adding a few missing values.
• order the group variables.
• keep relevent variables.

# Prepare analysis ready data
df <- r2rtf_adsl
df$AGE[1:3] <- NA # Create missing value for illustration purpose.
df$ARM <- factor(
  df$ARM,
  c("Placebo", "Xanomeline Low Dose", "Xanomeline High Dose"),
  c("Placebo", "Low Dose", "High Dose")
)
df <- df[, c("USUBJID", "ARM", "AGE", "SEX", "RACE", "BMIBLGR1")]

head(df)
#>       USUBJID       ARM AGE SEX  RACE BMIBLGR1
#> 1 01-701-1015   Placebo  NA   F WHITE   25-<30
#> 2 01-701-1023   Placebo  NA   M WHITE     >=30
#> 3 01-701-1028 High Dose  NA   M WHITE     >=30
#> 4 01-701-1033  Low Dose  74   M WHITE   25-<30
#> 5 01-701-1034 High Dose  77   F WHITE   25-<30
#> 6 01-701-1047   Placebo  85   F WHITE     >=30

The interactive table1 can then be created by the metalite_table1 function.

metalite_table1(
  ~ AGE + SEX + RACE + BMIBLGR1 | ARM, # formula
  data = df, # source data
  id = "USUBJID" # unique subject id
)

The interactive features are illustrated in the GIF below.

In comparison with the table1 R package, we are able to answer ad-hoc questions from clinicians with interactive features such as

• Who are those subjects with missing value.
• Who are those outliers.
  
• What’s the distribution of a variable by treatment group.

Those are common questions for ongoing clinical trials.

Features

Section Title

The section title is defined by the column labels. The metalite::get_labels allows you quickly check the label.

metalite::get_label(df)
#>                       USUBJID                           ARM 
#>   "Unique Subject Identifier"                         "ARM" 
#>                           AGE                           SEX 
#>                         "Age"                         "Sex" 
#>                          RACE                      BMIBLGR1 
#>                        "Race" "Pooled Baseline BMI Group 1"

The metalite::assign_labels can modify labels

df <- metalite::assign_label(
  df,
  var = c("AGE", "BMIBLGR1"),
  label = c("Age (Year)", "BMI Group (kg/m2)")
)

metalite::get_label(df)
#>                     USUBJID                         ARM 
#> "Unique Subject Identifier"                       "ARM" 
#>                         AGE                         SEX 
#>                "Age (Year)"                       "Sex" 
#>                        RACE                    BMIBLGR1 
#>                      "Race"         "BMI Group (kg/m2)"

The record_name argument can modify the term at the first row, for example, “Number of Participants”.

metalite_table1(~ AGE + SEX | ARM,
  data = df,
  id = "USUBJID",
  record_name = "Participants"
)

Total Column

The Total column can be hided.

metalite_table1(~ AGE + SEX | ARM,
  data = df,
  id = "USUBJID",
  total = FALSE
)

Download Data

We can allow user to download raw data, for example we can set download="listing" to enable user download the drill down listing.

metalite_table1(~ AGE + SEX | ARM,
  data = df,
  id = "USUBJID",
  download = "listing"
)

Rmarkdown Render

Sometimes we want to create multiple tables in a for loop, we need to clearly call the metalite_table1_to_html function and output text as raw content by using resutls="asis" (e.g., {r, results="asis"}) in the Rmarkdown code chunk. Create a dummy reactable object outside of for loop is necessary to include all required javascript library. A full Rmarkdown example is as below.

```{r}
library(metalite.table1)
```

```{r, include = FALSE}
reactable::reactable(data.frame(x = 1))
```

```{r, results="asis"}
type <- c("Subjects", "Records")
for (i in 1:2) {
  cat("### Table ", i, "\n")

  tbl <- metalite_table1(~ AGE + SEX | ARM,
    data = r2rtf::r2rtf_adsl,
    id = "USUBJID",
    type = type[i]
  )
  
  metalite_table1_to_html(tbl)
}
```

Table 1

Table 2

RTF file

User can also export a static table to an RTF file by using the metalite_table1_to_rtf function.

metalite_table1(
  ~ AGE + SEX + RACE + BMIBLGR1 | ARM, 
  data = df, 
  id = "USUBJID" 
) |>
  metalite_table1_to_rtf("tmp1.rtf")
#> [1] "tmp1.rtf"