ARDs (Analysis Results Datasets) in R

Check in with Barb!

Barb is on a journey from raw data to the TFLs. At various points we will be finding Barb in our data as we transform from Raw to TFL.

Barb has arrived at RESULTS!

ARDS 🤝 Tables: A Dynamic Duo

Objectives

• By the end of this ARD section you will have:
  • Developed an understanding of ARDs
  • Seen examples of univariate and hierarachical calculations
  • Gained an awareness of other possibilities in cards and statistical capabilities in cardx

Play by Play to achieve our Objectives

• 🕙 5 mins: Intro to ARDs and the ARS

• 🕥 20 mins: Basic summaries using {cards}

• 🕚 10 mins: Exercise

• 🕦 10 mins: Intro to stats with {cardx}

What are ARDs?

• Dataset that stores key metadata and raw results from analysis

• Long dataset that is 1 record per result value

• May contained formatted values as well

• The ARD can be used to to subsequently create tables and figures.

• The ARD does not describe the layout of the results

Analysis Results Data (ARD)

• After the initial creation of an ARD, the results can later be re-used again and again for subsequent reporting needs.

A few notes about ARDs

• ARDs give us the opportunity to rethink QC

  • QC can be focused on the raw value, as well as the formatted display

    • You don’t have to waste time trying to match formatting to match QC

    • But…QC will be more sensitive to floating point

• ARDs can be flexibly saved to different file types

  • For example: a dataset (rds, xpt, etc) or json file

Zooming Out: The Analysis Results Standard (ARS)

Objectives include:

• To leverage analysis results metadata to drive the automation of results
• To support storage, access, processing, traceability and reproducibility of results
• Learn more at https://www.cdisc.org/events/webinar/analysis-results-standard-public-review

Proposed Metadata Framework in the ARS

• The ARS provides a metadata-driven infrastructure for analysis

Proposed Metadata Framework in the ARS

• The ARS provides a metadata-driven infrastructure for analysis

• {cards} serves as the engine for the analysis

ARDs using {cards}

{cards}: Introduction

• Part of the Pharmaverse

• Collaboration between Roche, GSK, Novartis, Eli Lilly, Pfizer

• Contains a variety of utilities for making ARDs

• Can be used within the ARS workflow and separately

• 38k downloads per month 🤯

Data used in examples

ADSL from pharmaverseadam

adsl <- pharmaverseadam::adsl |> 
  dplyr::filter(SAFFL=="Y") |> 
  dplyr::mutate(ARM2 = ifelse(startsWith(ARM, "Xanomeline"), "Xanomeline", ARM))

ADAE from pharmaverseadam

adae <- pharmaverseadam::adae |> 
  dplyr::filter(SAFFL=="Y") |> 
  dplyr::mutate(ARM2 = ifelse(startsWith(ARM, "Xanomeline"), "Xanomeline", ARM)) |> 
  dplyr::filter(AESOC %in% unique(AESOC)[1:3]) |> 
  dplyr::group_by(AESOC) |> 
  dplyr::filter(AEDECOD %in% unique(AEDECOD)[1:3]) |> 
  dplyr::ungroup()

{cards}: ard_tabulate()

• includes n, %, N by default
• Any unobserved levels of the variables will be present in the resulting ARD.

library(cards)

adsl |> 
  ard_tabulate(
    variables = AGEGR1
  ) 

{cards} data frame: 6 x 9

  variable variable_level  context stat_name stat_label stat
1   AGEGR1            >64 tabulate         n          n  221
2   AGEGR1            >64 tabulate         N          N  254
3   AGEGR1            >64 tabulate         p          % 0.87
4   AGEGR1          18-64 tabulate         n          n   33
5   AGEGR1          18-64 tabulate         N          N  254
6   AGEGR1          18-64 tabulate         p          % 0.13

ℹ 3 more variables: fmt_fun, warning, error

{cards}: ard_tabulate()

• includes n, %, N by default
• Any unobserved levels of the variables will be present in the resulting ARD.

adsl |> 
  ard_tabulate(
    by = ARM2,
    variables = AGEGR1
  ) 

{cards} data frame: 12 x 11

   group1 group1_level variable variable_level stat_name stat_label  stat
1    ARM2      Placebo   AGEGR1            >64         n          n    72
2    ARM2      Placebo   AGEGR1            >64         N          N    86
3    ARM2      Placebo   AGEGR1            >64         p          % 0.837
4    ARM2      Placebo   AGEGR1          18-64         n          n    14
5    ARM2      Placebo   AGEGR1          18-64         N          N    86
6    ARM2      Placebo   AGEGR1          18-64         p          % 0.163
7    ARM2    Xanomeli…   AGEGR1            >64         n          n   149
8    ARM2    Xanomeli…   AGEGR1            >64         N          N   168
9    ARM2    Xanomeli…   AGEGR1            >64         p          % 0.887
10   ARM2    Xanomeli…   AGEGR1          18-64         n          n    19
11   ARM2    Xanomeli…   AGEGR1          18-64         N          N   168
12   ARM2    Xanomeli…   AGEGR1          18-64         p          % 0.113

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: ard_summary()

# create ARD with default summary statistics
adsl |> 
  ard_summary(
    variables = AGE
  )

{cards} data frame: 8 x 8

  variable context stat_name stat_label   stat fmt_fun
1      AGE summary         N          N    254       0
2      AGE summary      mean       Mean 75.087       1
3      AGE summary        sd         SD  8.246       1
4      AGE summary    median     Median     77       1
5      AGE summary       p25         Q1     70       1
6      AGE summary       p75         Q3     81       1
7      AGE summary       min        Min     51       1
8      AGE summary       max        Max     89       1

ℹ 2 more variables: warning, error

{cards}: ard_summary() by variable

by: summary statistics are calculated by all combinations of the by variables, including unobserved factor levels

adsl |> 
  ard_summary(
    variables = AGE,
    by = ARM2         # stats by treatment arm
  ) 

{cards} data frame: 16 x 10

   group1 group1_level variable stat_name stat_label   stat
1    ARM2      Placebo      AGE         N          N     86
2    ARM2      Placebo      AGE      mean       Mean 75.209
3    ARM2      Placebo      AGE        sd         SD   8.59
4    ARM2      Placebo      AGE    median     Median     76
5    ARM2      Placebo      AGE       p25         Q1     69
6    ARM2      Placebo      AGE       p75         Q3     82
7    ARM2      Placebo      AGE       min        Min     52
8    ARM2      Placebo      AGE       max        Max     89
9    ARM2    Xanomeli…      AGE         N          N    168
10   ARM2    Xanomeli…      AGE      mean       Mean 75.024
11   ARM2    Xanomeli…      AGE        sd         SD   8.09
12   ARM2    Xanomeli…      AGE    median     Median     77
13   ARM2    Xanomeli…      AGE       p25         Q1     71
14   ARM2    Xanomeli…      AGE       p75         Q3     81
15   ARM2    Xanomeli…      AGE       min        Min     51
16   ARM2    Xanomeli…      AGE       max        Max     88

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: ard_summary() statistics

statistic: specify univariate summary statistics. Accepts any function, base R, from a package, or user-defined.

cv <- function(x)  sd(x, na.rm = TRUE)/mean(x, na.rm = TRUE)

adsl |> 
  ard_summary(  
    variables = AGE,
    by = ARM2,
    statistic = ~ list(cv = cv) # customize statistics
  )

{cards} data frame: 2 x 10

  group1 group1_level variable stat_name stat_label  stat
1   ARM2      Placebo      AGE        cv         cv 0.114
2   ARM2    Xanomeli…      AGE        cv         cv 0.108

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: ard_summary() statistics

Customize the statistics returned for each variable

adsl |> 
  dplyr::mutate(AGE2 = AGE) |> 
  ard_summary(
    variables = c(AGE, AGE2),
    by = ARM2,
    statistic = list(AGE = list(cv = cv),
                     AGE2 = continuous_summary_fns(c("mean","median")))
  )

{cards} data frame: 6 x 10

  group1 group1_level variable stat_name stat_label   stat
1   ARM2      Placebo      AGE        cv         cv  0.114
2   ARM2      Placebo     AGE2      mean       Mean 75.209
3   ARM2      Placebo     AGE2    median     Median     76
4   ARM2    Xanomeli…      AGE        cv         cv  0.108
5   ARM2    Xanomeli…     AGE2      mean       Mean 75.024
6   ARM2    Xanomeli…     AGE2    median     Median     77

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: ard_summary() fmt_fun

• Override the default formatting functions
• Can also update later via update_ard_fmt_fun()

adsl |> 
  ard_summary(
    variables = AGE,
    by = ARM2,                               
    fmt_fun = ~list(mean = 0)                
  ) |> 
  apply_fmt_fun() # add a character column of rounded results

{cards} data frame: 16 x 11

   group1 group1_level variable stat_name stat_label   stat stat_fmt
1    ARM2      Placebo      AGE         N          N     86       86
2    ARM2      Placebo      AGE      mean       Mean 75.209       75
3    ARM2      Placebo      AGE        sd         SD   8.59      8.6
4    ARM2      Placebo      AGE    median     Median     76     76.0
5    ARM2      Placebo      AGE       p25         Q1     69     69.0
6    ARM2      Placebo      AGE       p75         Q3     82     82.0
7    ARM2      Placebo      AGE       min        Min     52     52.0
8    ARM2      Placebo      AGE       max        Max     89     89.0
9    ARM2    Xanomeli…      AGE         N          N    168      168
10   ARM2    Xanomeli…      AGE      mean       Mean 75.024       75
11   ARM2    Xanomeli…      AGE        sd         SD   8.09      8.1
12   ARM2    Xanomeli…      AGE    median     Median     77     77.0
13   ARM2    Xanomeli…      AGE       p25         Q1     71     71.0
14   ARM2    Xanomeli…      AGE       p75         Q3     81     81.0
15   ARM2    Xanomeli…      AGE       min        Min     51     51.0
16   ARM2    Xanomeli…      AGE       max        Max     88     88.0

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: Other Summary Functions

• ard_tabulate_value(): similar to ard_tabulate(), but for dichotomous tabulations

• ard_hierarchical(): similar to ard_tabulate(), but built for nested tabulations, e.g. AE terms within SOC

• ard_mvsummary(): similar to ard_summary(), for multivariate summaries. The function accepts other arguments like the full and subsetted (within the by groups) data sets.

• ard_missing(): tabulates rates of missingness

The results from all these functions are entirely compatible with one another, and can be stacked into a single data frame. 🥞

{cards}: Other Functions

In addition to exporting functions to prepare summaries, {cards} exports many utilities for wrangling ARDs and creating new ARDs.

Constructing: bind_ard(), tidy_as_ard(), nest_for_ard(), check_ard_structure(), and many more

Wrangling: get_ard_statistics(), replace_null_statistic(), etc.

{cards}: Stacking utilities

• data and .by are shared by all ard_* calls

• Additional Options .overall, .missing, .attributes, and .total_n provide even more results

• By default, summaries of the .by variable are included

adsl |> 
  ard_stack( 
    .by = ARM2,      
    ard_summary(variables = AGE, statistic = ~ continuous_summary_fns(c("mean","sd"))), 
    ard_tabulate(variables = AGEGR1, statistic = ~ "p")
  )  

{cards} data frame: 14 x 11

   group1 group1_level variable variable_level stat_name stat_label   stat
1    ARM2      Placebo      AGE                     mean       Mean 75.209
2    ARM2      Placebo      AGE                       sd         SD   8.59
3    ARM2      Placebo   AGEGR1            >64         p          %  0.837
4    ARM2      Placebo   AGEGR1          18-64         p          %  0.163
5    ARM2    Xanomeli…      AGE                     mean       Mean 75.024
6    ARM2    Xanomeli…      AGE                       sd         SD   8.09
7    ARM2    Xanomeli…   AGEGR1            >64         p          %  0.887
8    ARM2    Xanomeli…   AGEGR1          18-64         p          %  0.113
9    <NA>                  ARM2        Placebo         n          n     86
10   <NA>                  ARM2        Placebo         N          N    254
11   <NA>                  ARM2        Placebo         p          %  0.339
12   <NA>                  ARM2      Xanomeli…         n          n    168
13   <NA>                  ARM2      Xanomeli…         N          N    254
14   <NA>                  ARM2      Xanomeli…         p          %  0.661

ℹ 4 more variables: context, fmt_fun, warning, error

Quick recap!

• Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2)
• First, we compute the continuous summaries for AGE by ARM2

ard_summary(
  data = adsl,
  by = ,
  variables = 
)

Quick recap!

• Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2)
• First, we compute the continuous summaries for AGE by ARM2

ard_summary(
  data = adsl,
  by = ARM2,
  variables = AGE 
)

{cards} data frame: 16 x 10

   group1 group1_level variable stat_name stat_label   stat
1    ARM2      Placebo      AGE         N          N     86
2    ARM2      Placebo      AGE      mean       Mean 75.209
3    ARM2      Placebo      AGE        sd         SD   8.59
4    ARM2      Placebo      AGE    median     Median     76
5    ARM2      Placebo      AGE       p25         Q1     69
6    ARM2      Placebo      AGE       p75         Q3     82
7    ARM2      Placebo      AGE       min        Min     52
8    ARM2      Placebo      AGE       max        Max     89
9    ARM2    Xanomeli…      AGE         N          N    168
10   ARM2    Xanomeli…      AGE      mean       Mean 75.024
11   ARM2    Xanomeli…      AGE        sd         SD   8.09
12   ARM2    Xanomeli…      AGE    median     Median     77
13   ARM2    Xanomeli…      AGE       p25         Q1     71
14   ARM2    Xanomeli…      AGE       p75         Q3     81
15   ARM2    Xanomeli…      AGE       min        Min     51
16   ARM2    Xanomeli…      AGE       max        Max     88

ℹ 4 more variables: context, fmt_fun, warning, error

Quick recap!

• Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2)
• Next, we compute the categorical summaries for AGEGR1 and SEX by ARM2

ard_tabulate(
  data = adsl,
  by = ,
  variables = 
)

Quick recap!

• Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2)
• Next, we compute the categorical summaries for AGEGR1 and SEX by ARM2

ard_tabulate(
  data = adsl,
  by = ARM2,
  variables = c(AGEGR1, SEX) 
)

{cards} data frame: 24 x 11

   group1 group1_level variable variable_level stat_name stat_label  stat
1    ARM2      Placebo   AGEGR1            >64         n          n    72
2    ARM2      Placebo   AGEGR1            >64         N          N    86
3    ARM2      Placebo   AGEGR1            >64         p          % 0.837
4    ARM2      Placebo   AGEGR1          18-64         n          n    14
5    ARM2      Placebo   AGEGR1          18-64         N          N    86
6    ARM2      Placebo   AGEGR1          18-64         p          % 0.163
7    ARM2      Placebo      SEX              F         n          n    53
8    ARM2      Placebo      SEX              F         N          N    86
9    ARM2      Placebo      SEX              F         p          % 0.616
10   ARM2      Placebo      SEX              M         n          n    33

ℹ 14 more rows

ℹ Use `print(n = ...)` to see more rows

ℹ 4 more variables: context, fmt_fun, warning, error

Quick recap!

Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2) in a single ard_stack() call, including:

• summaries by ARM2 as performed above

• continuous summaries from part A for AGE

• categorical summaries from part B for AGEGR1 and SEX

ard_stack(
  data = adsl,
  .by = ARM2,
  
  # add ard_* calls here
  
)

Quick recap!

Let’s compute summaries for a demography table that includes age (AGE), age group (AGEGR1), and sex (SEX) by treatment (ARM2) in a single ard_stack() call, including:

• summaries by ARM2 as performed above

• continuous summaries from part A for AGE

• categorical summaries from part B for AGEGR1 and SEX

ard_stack(
  data = adsl,
  .by = ARM2,
  ard_summary(variables = AGE),
  ard_tabulate(variables = c(AGEGR1, SEX))
)

{cards} data frame: 46 x 11

   group1 group1_level variable variable_level stat_name stat_label   stat
1    ARM2      Placebo      AGE                        N          N     86
2    ARM2      Placebo      AGE                     mean       Mean 75.209
3    ARM2      Placebo      AGE                       sd         SD   8.59
4    ARM2      Placebo      AGE                   median     Median     76
5    ARM2      Placebo      AGE                      p25         Q1     69
6    ARM2      Placebo      AGE                      p75         Q3     82
7    ARM2      Placebo      AGE                      min        Min     52
8    ARM2      Placebo      AGE                      max        Max     89
9    ARM2      Placebo   AGEGR1            >64         n          n     72
10   ARM2      Placebo   AGEGR1            >64         N          N     86

ℹ 36 more rows

ℹ Use `print(n = ...)` to see more rows

ℹ 4 more variables: context, fmt_fun, warning, error

Quick recap!

We can also add:

• Overall summaries for all variables
• Total N

ard_stack(
  data = adsl,
  .by = ARM2,
  ard_summary(variables = AGE),
  ard_tabulate(variables = c(AGEGR1, SEX)),
  .overall = TRUE,
  .total_n = TRUE
)

{cards} data frame: 67 x 11

   group1 group1_level variable variable_level stat_name stat_label   stat
1    ARM2      Placebo      AGE                        N          N     86
2    ARM2      Placebo      AGE                     mean       Mean 75.209
3    ARM2      Placebo      AGE                       sd         SD   8.59
4    ARM2      Placebo      AGE                   median     Median     76
5    ARM2      Placebo      AGE                      p25         Q1     69
6    ARM2      Placebo      AGE                      p75         Q3     82
7    ARM2      Placebo      AGE                      min        Min     52
8    ARM2      Placebo      AGE                      max        Max     89
9    ARM2      Placebo   AGEGR1            >64         n          n     72
10   ARM2      Placebo   AGEGR1            >64         N          N     86

ℹ 57 more rows

ℹ Use `print(n = ...)` to see more rows

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: Hierarchical Summary Functions

Following hierarchical summary functions aid in nested tabulations (e.g. AE terms within SOC):

• ard_hierarchical(): calculating nested subject-level rates

• ard_hierarchical_count(): calculating nested event-level counts

{cards}: ard_hierarchical

This function specializes in calculating subject-level rates.

• Rates computed on lowest level variables, nested within others

• id helps to check that no duplicate rows exist within the c(id, variables) columns

• denominator dictates the denominator for the rates

adae |> 
  dplyr::slice_tail(n = 1L, by = c(USUBJID, ARM, AESOC, AEDECOD)) |> 
  ard_hierarchical(
    variables = c(AESOC, AEDECOD),
    by = TRT01A,
    id = USUBJID,
    denominator = adsl
  )

{cards} data frame: 81 x 13

   group1 group1_level group2 group2_level variable variable_level stat_name stat_label  stat
1  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         n          n     9
2  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         N          N    86
3  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         p          % 0.105
4  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         n          n     1
5  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         N          N    86
6  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         p          % 0.012
7  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         n          n     3
8  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         N          N    86
9  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         p          % 0.035
10 TRT01A      Placebo  AESOC    GENERAL …  AEDECOD      APPLICAT…         n          n     3

ℹ 71 more rows

ℹ Use `print(n = ...)` to see more rows

ℹ 4 more variables: context, fmt_fun, warning, error

{cards}: ard_hierarchical_count

This function specializes in calculating event-level frequencies.

adae |> 
  ard_hierarchical_count( 
    variables = c(AESOC, AEDECOD),
    by = TRT01A
  )

{cards} data frame: 27 x 13

   group1 group1_level group2 group2_level variable variable_level stat_name stat_label stat
1  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         n          n   10
2  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         n          n    2
3  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         n          n    3
4  TRT01A      Placebo  AESOC    GENERAL …  AEDECOD      APPLICAT…         n          n    3
5  TRT01A      Placebo  AESOC    GENERAL …  AEDECOD      APPLICAT…         n          n   10
6  TRT01A      Placebo  AESOC    GENERAL …  AEDECOD        FATIGUE         n          n    2
7  TRT01A      Placebo  AESOC    SKIN AND…  AEDECOD       ERYTHEMA         n          n   13
8  TRT01A      Placebo  AESOC    SKIN AND…  AEDECOD       PRURITUS         n          n   11
9  TRT01A      Placebo  AESOC    SKIN AND…  AEDECOD      PRURITUS…         n          n    0
10 TRT01A    Xanomeli…  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         n          n    3

ℹ 17 more rows

ℹ Use `print(n = ...)` to see more rows

ℹ 4 more variables: context, fmt_fun, warning, error

Hierarchical summaries many ways

• Displays for hierarchical data typically report on each level of the hierarchy (Any AE Overall, by System Organ Class, by Preferred Term)

• This can mean several calls to the ard_hierarchical_* functions

• Further, subject-level summaries require a different subset of the data each time. For example, to calculate Overall rates, we need to subset to 1 record per subject in ADAE.

• Is there an easier way?

{cards}: Stacking function for ard_hierarchical()

• ard_hierarchical stacking functions simplify this multi-step process into a single step

• The id argument is used to subset the data along the way

adae |> 
  ard_stack_hierarchical(
    variables = c(AESOC, AEDECOD),
    by = TRT01A,
    id = USUBJID,
    denominator = pharmaverseadam::adsl
  )

   group1 group1_level group2 group2_level variable variable_level stat_name stat_label  stat
1  TRT01A      Placebo   <NA>                 AESOC      GASTROIN…         n          n    12
2  TRT01A      Placebo   <NA>                 AESOC      GASTROIN…         N          N    86
3  TRT01A      Placebo   <NA>                 AESOC      GASTROIN…         p          %  0.14
4  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         n          n     9
5  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         N          N    86
6  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         p          % 0.105
7  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         n          n     1
8  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         N          N    86
9  TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         p          % 0.012
10 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         n          n     3
11 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         N          N    86
12 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         p          % 0.035

{cards}: Stacking function for ard_hierarchical_count()

• Below is the stacking function for event-level summaries, aligned with ard_hierarchical_count()

adae |> 
  ard_stack_hierarchical_count(
    variables = c(AESOC, AEDECOD),
    by = TRT01A, 
    denominator = pharmaverseadam::adsl
  )

  group1 group1_level group2 group2_level variable variable_level stat_name stat_label stat
1 TRT01A      Placebo   <NA>                 AESOC      GASTROIN…         n          n   15
2 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      DIARRHOEA         n          n   10
3 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD      HIATUS H…         n          n    2
4 TRT01A      Placebo  AESOC    GASTROIN…  AEDECOD       VOMITING         n          n    3

Exercise 🏃‍➡️

1. Navigate to Posit Cloud script exercises/04-ARD.R

2. Compute the nested AE tabulations as described.

3. Add the “completed” sticky note to your laptop when complete.

10:00

{cardx}

• Extension of the {cards} package, providing additional functions to create Analysis Results Datasets (ARDs)

• The {cardx} package exports many ard_*() function for statistical methods.

{cardx}

• Exports ARD frameworks for statistical analyses from many packages

  - {stats}
  - {car}
  - {effectsize}
  - {emmeans}
  - {geepack}
  - {lme4}
  - {parameters}
  - {smd}
  - {survey}
  - {survival}

• This list is growing (rather quickly) 🌱

{cardx} t-test Example

• We see the results like the mean difference, the confidence interval, and p-value as expected.

• And we also see the function’s inputs, which is incredibly useful for re-use, e.g. we know that we did not use equal variances.

adsl |> 
  cardx::ard_stats_t_test(by = ARM2, variables = AGE)

{cards} data frame: 14 x 9

   group1 variable   context   stat_name stat_label      stat
1    ARM2      AGE stats_t_…    estimate  Mean Dif…     0.185
2    ARM2      AGE stats_t_…   estimate1  Group 1 …    75.209
3    ARM2      AGE stats_t_…   estimate2  Group 2 …    75.024
4    ARM2      AGE stats_t_…   statistic  t Statis…     0.166
5    ARM2      AGE stats_t_…     p.value    p-value     0.868
6    ARM2      AGE stats_t_…   parameter  Degrees …   162.643
7    ARM2      AGE stats_t_…    conf.low  CI Lower…     -2.02
8    ARM2      AGE stats_t_…   conf.high  CI Upper…     2.391
9    ARM2      AGE stats_t_…      method     method Welch Tw…
10   ARM2      AGE stats_t_… alternative  alternat… two.sided
11   ARM2      AGE stats_t_…          mu    H0 Mean         0
12   ARM2      AGE stats_t_…      paired  Paired t…     FALSE
13   ARM2      AGE stats_t_…   var.equal  Equal Va…     FALSE
14   ARM2      AGE stats_t_…  conf.level  CI Confi…      0.95

ℹ 3 more variables: fmt_fun, warning, error

{cardx} t-test Example

• What to do if a method you need is not implemented?

• It’s simple to wrap existing frameworks to customize.

• One-sample t-test example utilizing cards::ard_summary().

adsl |> 
  cards::ard_summary(
    variables = AGE,
    statistic = everything() ~ list(t_test = \(x) t.test(x) |> broom::tidy())
  ) |> 
  dplyr::mutate(context = "t_test_one_sample")

{cards} data frame: 8 x 8

  variable   context   stat_name stat_label      stat fmt_fun
1      AGE t_test_o…    estimate   estimate    75.087       1
2      AGE t_test_o…   statistic  statistic   145.119       1
3      AGE t_test_o…     p.value    p.value         0       1
4      AGE t_test_o…   parameter  parameter       253       1
5      AGE t_test_o…    conf.low   conf.low    74.068       1
6      AGE t_test_o…   conf.high  conf.high    76.106       1
7      AGE t_test_o…      method     method One Samp…    <fn>
8      AGE t_test_o… alternative  alternat… two.sided    <fn>

ℹ 2 more variables: warning, error

{cardx} t-test Example

• How to modify if we need a two-sample test, or more generally accessing other columns in the data frame.

adsl |> 
  cards::ard_mvsummary(
    variables = AGE,
    statistic = 
      ~ list(t_test = \(x, data, ...) t.test(x ~ data$ARM2) |> broom::tidy())
  ) |> 
  dplyr::mutate(group1 = "ARM2", context = "t_test_two_sample") |> 
  cards::tidy_ard_column_order()

{cards} data frame: 10 x 9

   group1 variable   context   stat_name stat_label      stat
1    ARM2      AGE t_test_t…    estimate   estimate     0.185
2    ARM2      AGE t_test_t…   estimate1  estimate1    75.209
3    ARM2      AGE t_test_t…   estimate2  estimate2    75.024
4    ARM2      AGE t_test_t…   statistic  statistic     0.166
5    ARM2      AGE t_test_t…     p.value    p.value     0.868
6    ARM2      AGE t_test_t…   parameter  parameter   162.643
7    ARM2      AGE t_test_t…    conf.low   conf.low     -2.02
8    ARM2      AGE t_test_t…   conf.high  conf.high     2.391
9    ARM2      AGE t_test_t…      method     method Welch Tw…
10   ARM2      AGE t_test_t… alternative  alternat… two.sided

ℹ 3 more variables: fmt_fun, warning, error

{cardx} Regression

• Includes functionality to summarize nearly every type of regression model in the R ecosystem:

betareg::betareg(), biglm::bigglm(), brms::brm(), cmprsk::crr(), fixest::feglm(), fixest::femlm(), fixest::feNmlm(), fixest::feols(), gam::gam(), geepack::geeglm(), glmmTMB::glmmTMB(), glmtoolbox::glmgee(), lavaan::lavaan(), lfe::felm(), lme4::glmer.nb(), lme4::glmer(), lme4::lmer(), logitr::logitr(), MASS::glm.nb(), MASS::polr(), mgcv::gam(), mice::mira, mmrm::mmrm(), multgee::nomLORgee(), multgee::ordLORgee(), nnet::multinom(), ordinal::clm(), ordinal::clmm(), parsnip::model_fit, plm::plm(), pscl::hurdle(), pscl::zeroinfl(), rstanarm::stan_glm(), stats::aov(), stats::glm(), stats::lm(), stats::nls(), survey::svycoxph(), survey::svyglm(), survey::svyolr(), survival::cch(), survival::clogit(), survival::coxph(), survival::survreg(), svyVGAM::svy_vglm(), tidycmprsk::crr(), VGAM::vgam(), VGAM::vglm() (and more)

{cardx} Regression Example

library(survival)

# build model
mod <- pharmaverseadam::adtte_onco |> 
  dplyr::filter(PARAM %in% "Progression Free Survival") |>
  coxph(ggsurvfit::Surv_CNSR() ~ ARM, data = _)

# put model in a summary table
tbl <- gtsummary::tbl_regression(mod, exponentiate = TRUE) |> 
  gtsummary::add_n(location = c('label', 'level')) |> 
  gtsummary::add_nevent(location = c('label', 'level'))

Characteristic	N	Event N	HR	95% CI	p-value
Description of Planned Arm	254	6
Placebo	86	3	—	—
Xanomeline High Dose	84	2	3.00	0.39, 22.9	0.3
Xanomeline Low Dose	84	1	1.27	0.11, 14.3	0.8
Abbreviations: CI = Confidence Interval, HR = Hazard Ratio

{cardx} Regression Example

The cardx::ard_regression() does a lot for us in the background.

• Identifies the variable from the regression terms (i.e. groups levels of the same variable)
• Identifies reference groups from categorical covariates
• Finds variable labels from the source data frames
• Knows the total N of the model, the number of events, and can do the same for each level of categorical variables
• Contextually aware of slopes, odds ratios, hazard ratios, and incidence rate ratios
• And much much more.

When things go wrong 😱

What happens when statistics are un-calculable?

ard_gone_wrong <- 
  adsl |> 
  ard_summary(
    by = ARM2,
    variable = AGEGR1,
    statistic = ~list(kurtosis = \(x) e1071::kurtosis(x))
  ) |> 
  cards::replace_null_statistic()
ard_gone_wrong

{cards} data frame: 2 x 10

  group1 group1_level variable stat_name stat_label stat   warning     error
1   ARM2      Placebo   AGEGR1  kurtosis   kurtosis   NA argument… non-nume…
2   ARM2    Xanomeli…   AGEGR1  kurtosis   kurtosis   NA argument… non-nume…

ℹ 2 more variables: context, fmt_fun

cards::print_ard_conditions(ard_gone_wrong)

• Where is the statistic? AGEGR1 is character

• Even when there are errors or warnings, we still get the ARD with the expected structure returned.

  • THIS IS BIG! There are MANY circumstances, when you are designing TLGs early in a study when you do not have all the data required to calculate every statistic.

  • This allows you to design everything up-front.

• We can also report these warnings and errors back to users. <!CLICK!>