Pressure testing when there are >3000 subjects and >5 arms

[LittleBeannie]

Prompt Goal: Generate R code using the stats and dplyr/tidyr R packages to create two synthetic clinical trial datasets: forestly_adsl and forestly_adae.

Dataset Descriptions:

1. forestly_adsl (Subject-Level Data):
   This dataset contains one record per patient.
   The primary patient identifier is USUBJID.
   The treatment group variable is TRT01A.
   Key Variables: USUBJID, SITEID, TRT01A, AGE, SEX, SAFFL, BMIBL, WEIGHTBL.

2. forestly_adae (Adverse Event Data):
   This dataset contains records of adverse events for patients.
   A single patient (USUBJID) can have multiple records, representing multiple adverse events.
   Key Variables: USUBJID, SITEID, AGE, TRT01A, AEDECOD, AEREL, AESER, AEACN, AEOUT.

Code Generation Instructions:
Please generate R code to perform the following steps:

Part 1: Generate the forestly_adsl data.frame
Create a dataframe named forestly_adsl with 10000 rows and the following columns, generated according to these rules:

• SITEID: Randomly sample from the list ["701", "702", "703", "704", "705", "706", "708", "709", "710"] with equal probability for each site.
• USUBJID: Create a unique subject ID by concatenating the SITEID with a randomly generated 5-digit number (e.g., "701-01234"). Ensure all USUBJID values are unique.
• TRT01A: This is a 6-arm study. Randomly assign each subject to one of the following groups with equal probability: "Group A", "Group B", "Group C", "Group D", "Group E", "Group F".
• AGE: Generate a random integer between 30 and 80 (inclusive).
• SEX: Randomly sample from ["F", "M"] with equal probability.
• RACE: Randomly sample from ["WHITE" , "BLACK OR AFRICAN AMERICAN"] with equal probability.
• SAFFL: (Safety Flag) Randomly sample from ["Y", "N"] with probabilities of 90% for "Y" and 10% for "N".
• BMIBL: (Baseline BMI) Generate a random floating-point number between 18.0 and 40.0.
• WEIGHTBL: (Baseline Weight) Generate a random integer between 40 and 120 (inclusive).

Part 2: Generate the forestly_adae data.frame
Create a second dataframe named forestly_adae. This dataframe should be generated based on the forestly_adsl data.
Determine the Number of Adverse Events per Subject:
For each subject in forestly_adsl, generate a random number of adverse events they will experience. This number should be between 0 and 10 (inclusive).

Create Records for Each Adverse Event:
Iterate through each subject from forestly_adsl and create the specified number of adverse event records for them.
The final forestly_adae will be the collection of all these records.
Define the Columns for forestly_adae:

• USUBJID, SITEID, TRT01A, AGE: These values must be copied directly from the corresponding subject's record in the forestly_adsl DataFrame.
• AEDECOD: (Adverse Event Term) Randomly sample from the following list of common adverse event terms: ["HEADACHE", "NAUSEA", "FATIGUE", "DIZZINESS", "DIARRHEA", "RASH", "FEVER"].
• AEBODSYS: (Adverse Event Body Organ) It depends on the value of AEDECOD and follows the mapping rule below:
  AEDECOD AEBODSYS
  1 APPLICATION SITE ERYTHEMA GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  2 APPLICATION SITE PRURITUS GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  3 DIARRHOEA GASTROINTESTINAL DISORDERS
  4 ERYTHEMA SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  6 ATRIOVENTRICULAR BLOCK SECOND DEGREE CARDIAC DISORDERS
  12 HIATUS HERNIA GASTROINTESTINAL DISORDERS
  14 UPPER RESPIRATORY TRACT INFECTION INFECTIONS AND INFESTATIONS
  15 BUNDLE BRANCH BLOCK LEFT CARDIAC DISORDERS
  17 PRURITUS GENERALISED SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  18 APPLICATION SITE VESICLES GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  23 PHARYNGOLARYNGEAL PAIN RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  24 NASAL CONGESTION RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  27 PRURITUS SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  28 LOCALISED INFECTION INFECTIONS AND INFESTATIONS
  31 MICTURITION URGENCY RENAL AND URINARY DISORDERS
  32 ARTHRALGIA MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  33 CELLULITIS INFECTIONS AND INFESTATIONS
  36 APPLICATION SITE IRRITATION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  40 FATIGUE GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  42 APPLICATION SITE DERMATITIS GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  43 COUGH RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  44 URINARY TRACT INFECTION INFECTIONS AND INFESTATIONS
  45 PYREXIA GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  47 EYE SWELLING EYE DISORDERS
  48 EYE ALLERGY EYE DISORDERS
  49 EYE PRURITUS EYE DISORDERS
  77 INCREASED APPETITE METABOLISM AND NUTRITION DISORDERS
  93 URTICARIA SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  95 APPLICATION SITE URTICARIA GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  97 NASAL MUCOSA BIOPSY INVESTIGATIONS
  98 SECRETION DISCHARGE GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  103 PNEUMONIA INFECTIONS AND INFESTATIONS
  104 MUSCULAR WEAKNESS MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  112 EYE LASER SURGERY SURGICAL AND MEDICAL PROCEDURES
  114 JOINT DISLOCATION INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  115 SHOULDER PAIN MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  116 INCONTINENCE RENAL AND URINARY DISORDERS
  118 SKIN LACERATION INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  119 CONFUSIONAL STATE PSYCHIATRIC DISORDERS
  120 DYSPNOEA RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  121 SUDDEN DEATH GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  157 RASH ERYTHEMATOUS SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  181 ONYCHOMYCOSIS INFECTIONS AND INFESTATIONS
  182 CYST GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  186 DEPRESSED MOOD PSYCHIATRIC DISORDERS
  189 DYSPHONIA RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  202 NAUSEA GASTROINTESTINAL DISORDERS
  203 HEADACHE NERVOUS SYSTEM DISORDERS
  206 BACK PAIN MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  221 HYPERHIDROSIS SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  239 WHITE BLOOD CELL COUNT INCREASED INVESTIGATIONS
  240 NEUTROPHIL COUNT INCREASED INVESTIGATIONS
  241 URINE ANALYSIS ABNORMAL INVESTIGATIONS
  245 RECTAL HAEMORRHAGE GASTROINTESTINAL DISORDERS
  248 SKIN IRRITATION SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  250 INSOMNIA PSYCHIATRIC DISORDERS
  261 CHILLS GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  267 CATARACT OPERATION SURGICAL AND MEDICAL PROCEDURES
  269 OEDEMA PERIPHERAL GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  280 SOMNOLENCE NERVOUS SYSTEM DISORDERS
  282 DISORIENTATION PSYCHIATRIC DISORDERS
  283 DECREASED APPETITE METABOLISM AND NUTRITION DISORDERS
  295 GLAUCOMA EYE DISORDERS
  296 HYPOTENSION VASCULAR DISORDERS
  297 TACHYCARDIA CARDIAC DISORDERS
  300 MYOCARDIAL INFARCTION CARDIAC DISORDERS
  314 MALAISE GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  315 RASH SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  317 DIZZINESS NERVOUS SYSTEM DISORDERS
  319 CYSTITIS INFECTIONS AND INFESTATIONS
  320 VOMITING GASTROINTESTINAL DISORDERS
  322 EXCORIATION INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  325 CONTUSION INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  334 VERTIGO EAR AND LABYRINTH DISORDERS
  335 PALPITATIONS CARDIAC DISORDERS
  336 ELECTROCARDIOGRAM ST SEGMENT DEPRESSION INVESTIGATIONS
  337 WOLFF-PARKINSON-WHITE SYNDROME CARDIAC DISORDERS
  357 OEDEMA GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  360 INFLUENZA INFECTIONS AND INFESTATIONS
  362 NASOPHARYNGITIS INFECTIONS AND INFESTATIONS
  372 BLOOD GLUCOSE INCREASED INVESTIGATIONS
  391 VIRAL INFECTION INFECTIONS AND INFESTATIONS
  403 POLLAKIURIA RENAL AND URINARY DISORDERS
  404 FLATULENCE GASTROINTESTINAL DISORDERS
  408 PARKINSON'S DISEASE NERVOUS SYSTEM DISORDERS
  409 COMPLETED SUICIDE PSYCHIATRIC DISORDERS
  412 HYPERBILIRUBINAEMIA HEPATOBILIARY DISORDERS
  413 DYSPEPSIA GASTROINTESTINAL DISORDERS
  429 VENTRICULAR HYPERTROPHY CARDIAC DISORDERS
  431 ABDOMINAL PAIN GASTROINTESTINAL DISORDERS
  433 COLON CANCER NEOPLASMS BENIGN, MALIGNANT AND UNSPECIFIED (INCL CYSTS AND POLYPS)
  436 ANXIETY PSYCHIATRIC DISORDERS
  440 IRRITABILITY PSYCHIATRIC DISORDERS
  458 SINUS ARRHYTHMIA CARDIAC DISORDERS
  459 BLOOD CREATINE PHOSPHOKINASE INCREASED INVESTIGATIONS
  460 POSTNASAL DRIP RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  464 CONJUNCTIVITIS EYE DISORDERS
  467 STUPOR NERVOUS SYSTEM DISORDERS
  468 TRANSIENT ISCHAEMIC ATTACK NERVOUS SYSTEM DISORDERS
  469 VISION BLURRED EYE DISORDERS
  476 SINUS BRADYCARDIA CARDIAC DISORDERS
  478 HEART RATE IRREGULAR INVESTIGATIONS
  484 AGITATION PSYCHIATRIC DISORDERS
  503 HEART RATE INCREASED INVESTIGATIONS
  510 MUSCLE SPASMS MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  511 VENTRICULAR EXTRASYSTOLES CARDIAC DISORDERS
  519 HOT FLUSH VASCULAR DISORDERS
  524 CONSTIPATION GASTROINTESTINAL DISORDERS
  525 ATRIAL HYPERTROPHY CARDIAC DISORDERS
  531 PAIN IN EXTREMITY MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  533 APPLICATION SITE SWELLING GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  561 EPISTAXIS RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  564 APPLICATION SITE WARMTH GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  620 ELECTROCARDIOGRAM T WAVE INVERSION INVESTIGATIONS
  637 BLOOD URINE PRESENT INVESTIGATIONS
  638 CYSTOSCOPY INVESTIGATIONS
  639 BENIGN PROSTATIC HYPERPLASIA REPRODUCTIVE SYSTEM AND BREAST DISORDERS
  642 DEHYDRATION METABOLISM AND NUTRITION DISORDERS
  644 ORTHOSTATIC HYPOTENSION VASCULAR DISORDERS
  646 HEMIANOPIA HOMONYMOUS NERVOUS SYSTEM DISORDERS
  652 DIABETES MELLITUS METABOLISM AND NUTRITION DISORDERS
  658 DRUG ERUPTION SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  682 SYNCOPE NERVOUS SYSTEM DISORDERS
  683 APPLICATION SITE BLEEDING GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  688 BLOOD ALKALINE PHOSPHATASE INCREASED INVESTIGATIONS
  707 HAEMOPTYSIS RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  708 RALES RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  718 SKIN EXFOLIATION SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  731 APPLICATION SITE INDURATION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  734 FALL INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  745 ARTHRITIS MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS
  762 COMPLEX PARTIAL SEIZURES NERVOUS SYSTEM DISORDERS
  765 ATRIOVENTRICULAR BLOCK FIRST DEGREE CARDIAC DISORDERS
  777 COLD SWEAT SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  787 ATRIAL FIBRILLATION CARDIAC DISORDERS
  788 CARDIAC FAILURE CONGESTIVE CARDIAC DISORDERS
  789 HYPONATRAEMIA METABOLISM AND NUTRITION DISORDERS
  799 BLISTER SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  814 APPLICATION SITE DESQUAMATION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  815 HIP FRACTURE INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  816 DELUSION PSYCHIATRIC DISORDERS
  817 PSYCHOMOTOR HYPERACTIVITY NERVOUS SYSTEM DISORDERS
  819 SUPRAVENTRICULAR TACHYCARDIA CARDIAC DISORDERS
  826 ELECTROCARDIOGRAM T WAVE AMPLITUDE DECREASED INVESTIGATIONS
  839 EAR INFECTION INFECTIONS AND INFESTATIONS
  840 GLOSSITIS GASTROINTESTINAL DISORDERS
  842 TINNITUS EAR AND LABYRINTH DISORDERS
  843 PARAESTHESIA ORAL NERVOUS SYSTEM DISORDERS
  845 SUPRAVENTRICULAR EXTRASYSTOLES CARDIAC DISORDERS
  850 RASH PRURITIC SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  853 CONJUNCTIVAL HAEMORRHAGE EYE DISORDERS
  854 CERUMEN IMPACTION EAR AND LABYRINTH DISORDERS
  871 DYSURIA RENAL AND URINARY DISORDERS
  874 ASTHENIA GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  878 PELVIC PAIN REPRODUCTIVE SYSTEM AND BREAST DISORDERS
  879 GASTROENTERITIS VIRAL INFECTIONS AND INFESTATIONS
  880 GASTROOESOPHAGEAL REFLUX DISEASE GASTROINTESTINAL DISORDERS
  881 VAGINAL MYCOSIS INFECTIONS AND INFESTATIONS
  882 CERVICITIS INFECTIONS AND INFESTATIONS
  902 BUNDLE BRANCH BLOCK RIGHT CARDIAC DISORDERS
  908 BRONCHITIS INFECTIONS AND INFESTATIONS
  934 EAR PAIN EAR AND LABYRINTH DISORDERS
  942 DERMATITIS CONTACT SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  945 NEPHROLITHIASIS RENAL AND URINARY DISORDERS
  959 EMPHYSEMA RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  974 RESTLESSNESS PSYCHIATRIC DISORDERS
  975 LETHARGY NERVOUS SYSTEM DISORDERS
  980 ATRIAL FLUTTER CARDIAC DISORDERS
  982 VENTRICULAR SEPTAL DEFECT CONGENITAL, FAMILIAL AND GENETIC DISORDERS
  990 MALIGNANT FIBROUS HISTIOCYTOMA NEOPLASMS BENIGN, MALIGNANT AND UNSPECIFIED (INCL CYSTS AND POLYPS)
  1001 SWELLING GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  1003 HYPERSENSITIVITY IMMUNE SYSTEM DISORDERS
  1021 ULCER GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  1043 ENURESIS RENAL AND URINARY DISORDERS
  1046 RHINORRHOEA RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS
  1060 WOUND INJURY, POISONING AND PROCEDURAL COMPLICATIONS
  1077 ALOPECIA SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  1098 PAIN GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  1104 INFLAMMATION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  1117 BRADYCARDIA CARDIAC DISORDERS
  1123 FOOD CRAVING METABOLISM AND NUTRITION DISORDERS
  1129 BODY TEMPERATURE INCREASED INVESTIGATIONS
  1130 HYPERTENSION VASCULAR DISORDERS
  1133 BALANCE DISORDER NERVOUS SYSTEM DISORDERS
  1140 COORDINATION ABNORMAL NERVOUS SYSTEM DISORDERS
  1147 APPLICATION SITE DISCOLOURATION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
  1152 DYSPHAGIA GASTROINTESTINAL DISORDERS
  1157 SKIN ULCER SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  1159 DERMATITIS ATOPIC SKIN AND SUBCUTANEOUS TISSUE DISORDERS
  1166 APPLICATION SITE REACTION GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS
• AEREL: (Causality/Relationship to Drug) Randomly sample from ["RELATED", "NOT RELATED"] with probabilities of 70% for "RELATED" and 30% for "NOT RELATED".
• AESER: (Serious Event) Randomly sample from ["Y", "N"] with probabilities of 20% for "Y" and 80% for "N".
• AEOUT: (Outcome of Event) Randomly sample from ["RECOVERED/RESOLVED", "NOT RECOVERED/NOT RESOLVED", "FATAL"] with probabilities of 80%, 10%, and 10% respectively.
• AEACN: (Action Taken with Study Treatment) Randomly sample from ["DOSE NOT CHANGED", "DRUG INTERRUPTED", "DOSE REDUCED", "DRUG WITHDRAWN", "NOT APPLICABLE"] with equal probability.

Part 3: Final Output
Output the code to generate the above 2 data frame.

[LittleBeannie]

The pressure code is:

library(dplyr)
library(tidyr)
library(forestly)
library(metalite)

# Set seed for reproducibility
set.seed(123)

#-------------------------------------------#
# Part 1: Generate forestly_adsl data.frame
#-------------------------------------------#
n_subjects <- 10000

# Define the possible values
sites <- c("701", "702", "703", "704", "705", "706", "708", "709", "710")
treatments <- c("Group A", "Group B", "Group C", "Group D", "Group E", "Group F")
sex_values <- c("F", "M")
race_values <- c("WHITE", "BLACK OR AFRICAN AMERICAN")
saffl_values <- c("Y", "N")

# Generate the base dataframe
forestly_adsl <- data.frame(
  SITEID = sample(sites, n_subjects, replace = TRUE),
  TRT01A = sample(treatments, n_subjects, replace = TRUE),
  AGE = sample(30:80, n_subjects, replace = TRUE),
  SEX = sample(sex_values, n_subjects, replace = TRUE),
  RACE = sample(race_values, n_subjects, replace = TRUE),
  SAFFL = sample(saffl_values, n_subjects, replace = TRUE, prob = c(0.9, 0.1)),
  BMIBL = runif(n_subjects, min = 18.0, max = 40.0),
  WEIGHTBL = sample(40:120, n_subjects, replace = TRUE)
)

# Generate unique USUBJID - sample without replacement to ensure uniqueness
unique_numbers <- sample(10000:99999, n_subjects, replace = FALSE)
forestly_adsl$USUBJID <- paste0(forestly_adsl$SITEID, "-",
                                sprintf("%05d", unique_numbers))

# Reorder columns to match the specification
forestly_adsl <- forestly_adsl[, c("USUBJID", "SITEID", "TRT01A", "AGE", "SEX", "RACE", "SAFFL", "BMIBL", "WEIGHTBL")]

#-------------------------------------------#
# Part 2: Generate forestly_adae data.frame
#-------------------------------------------#

# Define AE-specific values
ae_terms <- c("HEADACHE", "NAUSEA", "FATIGUE", "DIZZINESS", "DIARRHOEA", "RASH", "PYREXIA",
              "COUGH", "PRURITUS", "ARTHRALGIA", "VOMITING", "CONSTIPATION", "DYSPNOEA",
              "INSOMNIA", "HYPERTENSION", "PALPITATIONS", "BACK PAIN", "ABDOMINAL PAIN",
              "NASOPHARYNGITIS", "URINARY TRACT INFECTION", "UPPER RESPIRATORY TRACT INFECTION")
aerel_values <- c("RELATED", "NOT RELATED")
aeser_values <- c("Y", "N")
aeout_values <- c("RECOVERED/RESOLVED", "NOT RECOVERED/NOT RESOLVED", "FATAL")
aeacn_values <- c("DOSE NOT CHANGED", "DRUG INTERRUPTED", "DOSE REDUCED", "DRUG WITHDRAWN", "NOT APPLICABLE")
# Define AEDECOD to AEBODSYS mapping
aedecod_to_aebodsys <- c(
  "HEADACHE" = "NERVOUS SYSTEM DISORDERS",
  "NAUSEA" = "GASTROINTESTINAL DISORDERS",
  "FATIGUE" = "GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS",
  "DIZZINESS" = "NERVOUS SYSTEM DISORDERS",
  "DIARRHOEA" = "GASTROINTESTINAL DISORDERS",
  "RASH" = "SKIN AND SUBCUTANEOUS TISSUE DISORDERS",
  "PYREXIA" = "GENERAL DISORDERS AND ADMINISTRATION SITE CONDITIONS",
  "COUGH" = "RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS",
  "PRURITUS" = "SKIN AND SUBCUTANEOUS TISSUE DISORDERS",
  "ARTHRALGIA" = "MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS",
  "VOMITING" = "GASTROINTESTINAL DISORDERS",
  "CONSTIPATION" = "GASTROINTESTINAL DISORDERS",
  "DYSPNOEA" = "RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS",
  "INSOMNIA" = "PSYCHIATRIC DISORDERS",
  "HYPERTENSION" = "VASCULAR DISORDERS",
  "PALPITATIONS" = "CARDIAC DISORDERS",
  "BACK PAIN" = "MUSCULOSKELETAL AND CONNECTIVE TISSUE DISORDERS",
  "ABDOMINAL PAIN" = "GASTROINTESTINAL DISORDERS",
  "NASOPHARYNGITIS" = "INFECTIONS AND INFESTATIONS",
  "URINARY TRACT INFECTION" = "INFECTIONS AND INFESTATIONS",
  "UPPER RESPIRATORY TRACT INFECTION" = "INFECTIONS AND INFESTATIONS"
)

# Generate number of AEs per subject (0-20)
forestly_adsl$n_aes <- sample(0:20, n_subjects, replace = TRUE)

# Create the AE records by expanding the adsl data
forestly_adae <- forestly_adsl %>%
  filter(n_aes > 0) %>%  # Only keep subjects with at least 1 AE
  slice(rep(1:n(), n_aes)) %>%  # Replicate rows based on n_aes
  select(-n_aes) %>%  # Remove the helper column
  mutate(
    AEDECOD = sample(ae_terms, nrow(.), replace = TRUE),
    AEBODSYS = aedecod_to_aebodsys[AEDECOD],
    AEREL = sample(aerel_values, nrow(.), replace = TRUE, prob = c(0.7, 0.3)),
    AESER = sample(aeser_values, nrow(.), replace = TRUE, prob = c(0.2, 0.8)),
    AEOUT = sample(aeout_values, nrow(.), replace = TRUE, prob = c(0.8, 0.1, 0.1)),
    AEACN = sample(aeacn_values, nrow(.), replace = TRUE)
  )

# Clean up temporary column from adsl
forestly_adsl <- forestly_adsl %>% select(-n_aes)

#-------------------------------------------#
# Part 3: generate interactive AE forest plot using forestly
#-------------------------------------------#
adsl <- forestly_adsl
adae <- forestly_adae

adsl$TRT01A <- factor(forestly_adsl$TRT01A,
                    levels = c("Group A", "Group B", "Group C", "Group D", "Group E", "Group F"),
                    labels = c("Group A", "Group B", "Group C", "Group D", "Group E", "Group F")
)
adae$TRT01A <- factor(forestly_adae$TRT01A,
                    levels = c("Group A", "Group B", "Group C", "Group D", "Group E", "Group F"),
                    labels = c("Group A", "Group B", "Group C", "Group D", "Group E", "Group F")
)

meta <- meta_adam(population = adsl, observation = adae) |>
  define_plan(plan = plan(
    analysis = "ae_forestly",
    population = "apat",
    observation = "apat",
    parameter = "any;drug-related;serious;serious-drug-related;disc"
  )) |>
  define_analysis(name = "ae_forestly", label = "Interactive Forest Plot") |>
  define_population(
    name = "apat", group = "TRT01A", id = "USUBJID",
    subset = SAFFL == "Y", label = "All Patient as Treated"
  ) |>
  define_observation(
    name = "apat", group = "TRT01A",
    subset = SAFFL == "Y", label = "All Patient as Treated"
  ) |>
  define_parameter(
    name = "any",
    subset = NULL,
    label = "Any AEs",
    var = "AEDECOD", soc = "AEBODSYS"
  ) |>
  define_parameter(
    name = "drug-related",
    subset = AEREL == "RELATED",
    label = "Drug-related AEs",
    var = "AEDECOD", soc = "AEBODSYS"
  ) |>
  define_parameter(
    name = "serious",
    subset = AESER == "Y",
    label = "Serious AEs",
    var = "AEDECOD", soc = "AEBODSYS"
  ) |>
  define_parameter(
    name = "serious-drug-related",
    subset = AESER == "Y" & AEREL == "RELATED",
    label = "Serious Drug-Related AEs",
    var = "AEDECOD", soc = "AEBODSYS"
  ) |>
  define_parameter(
    name = "disc",
    subset = AEACN %in% c("DRUG WITHDRAWN", "DRUG INTERRUPTED"),
    label = "AE leads to drug interruption",
    var = "AEDECOD", soc = "AEBODSYS"
  ) |>
  meta_build()

tictoc::tic()
meta |>
  prepare_ae_forestly(ae_listing_display = c("USUBJID", "SITEID", "SEX", "AGE", "RACE", "BMIBL", "WEIGHTBL")) |>
  format_ae_forestly(color = c("black", "grey60", "grey40", "red", 'green', "blue")) |>
  ae_forestly()
tictoc::toc()

[LittleBeannie]

Emails from Yihui regarding the profiling:

I’m always intrigued by slow code and can’t resist investigating this type of problem. I did a quick profiling (a simple call of base::Rprof()) of the example code Yujie provided and found that, to my surprise, tools::toTitleCase() appeared to be the top bottleneck. Yes, it is a relatively expensive function, but I didn’t expect it to so slow. Anyway, I tried a simple trick of storing converted values in a dictionary and then looking it up instead of doing the conversion in real time every time:

title_case = local({
orig_fun = tools::toTitleCase
dict = list() # a dictionary to look up values already converted
function(text) {
i = !text %in% names(dict) # not converted yet
if (any(i)) {
x = unique(text[i])
dict[x] <<- as.list(orig_fun(x)) # update dictionary
}
unlist(unname(dict[text]), recursive = FALSE) # lookup
}
})

With this tick alone (there are other bottlenecks that are not straightforward to deal with), it brought the running time from 210s to 140s on my poor Merck Windows laptop: it cut out 1/3 of the total time!

Out of strong curiosity, I wondered what toTitleCase() actually worked on, and it turned out that during this expensive computation, it just kept converting these four values to title case:

"M" "F" "black or african american" "white"

Then I thought, converting four values to title case shouldn’t take so long. Upon a closer look, I found we were passing very long vectors consisting of these (repeated) values to toTitleCase(). That solved the mystery! This took 60-70s on my laptop:

system.time(for (i in 1:10) tools::toTitleCase(rep(
c("M", "F", "black or african american", "white"), length = 36000
)))

whereas my version of title_case() finished within a blink of the eye.

This is unbelievable to me since I was expecting certain complicated statistical computing to be slow, and it would cost my a few gray hairs to speed it up, but the first fruit hanged so low :) The rest ones are going to be much higher.

I’ve never used forestly before and I still need to learn more about it before I can say how gglite could help. From what I’ve read so far, it appears to be using a fairly complicated set of packages, and I don’t think I can simplify so many things easily, but HTML/JS/CSS applications are always my favorite things to do!