Day 17: ADSL Part 2 - Population Flags & Demographics

Week 3, Day 17: SAFFL, ITTFL, demographic groupings, and baseline measurements

1 Welcome to ADSL Part 2

Today we complete ADSL by adding population flags, demographic groupings, and baseline measurements.

What we’ll add:

Population flags (SAFFL, ITTFL)
Demographics groupings (AGEGR1, SEXN, RACEN)
Baseline measurements (HEIGHT, WEIGHT, BMI)

2 Setup and Load Day 16 ADSL

library(admiral)
library(pharmaversesdtm)
library(dplyr)
library(lubridate)
library(stringr)
library(xportr)

# Load SDTM domains
dm <- pharmaversesdtm::dm
ex <- pharmaversesdtm::ex
ds <- pharmaversesdtm::ds
vs <- pharmaversesdtm::vs

cat("Loaded SDTM domains\n")

Loaded SDTM domains

2.1 Rebuild ADSL Part 1 from Day 16

# Start with DM
adsl <- dm %>%
  select(-DOMAIN)

# Treatment variables
adsl <- adsl %>%
  mutate(
    TRT01P = ARM,
    TRT01A = ACTARM
  )

# Prepare EX with dates
ex_ext <- ex %>%
  derive_vars_dtm(dtc = EXSTDTC, new_vars_prefix = "EXST") %>%
  derive_vars_dtm(dtc = EXENDTC, new_vars_prefix = "EXEN", time_imputation = "last")

# First dose date
adsl <- adsl %>%
  derive_vars_merged(
    dataset_add = ex_ext,
    filter_add = (EXDOSE > 0 | (EXDOSE == 0 & str_detect(EXTRT, "PLACEBO"))) & !is.na(EXSTDTM),
    new_vars = exprs(TRTSDTM = EXSTDTM),
    order = exprs(EXSTDTM, EXSEQ),
    mode = "first",
    by_vars = exprs(STUDYID, USUBJID)
  )

# Last dose date
adsl <- adsl %>%
  derive_vars_merged(
    dataset_add = ex_ext,
    filter_add = (EXDOSE > 0 | (EXDOSE == 0 & str_detect(EXTRT, "PLACEBO"))) & !is.na(EXENDTM),
    new_vars = exprs(TRTEDTM = EXENDTM),
    order = exprs(EXENDTM, EXSEQ),
    mode = "last",
    by_vars = exprs(STUDYID, USUBJID)
  )

# Convert to dates
adsl <- adsl %>%
  derive_vars_dtm_to_dt(source_vars = exprs(TRTSDTM, TRTEDTM))

# Treatment duration
adsl <- adsl %>%
  derive_var_trtdurd()

cat("\nADSL Part 1 rebuilt:", nrow(adsl), "subjects\n")


ADSL Part 1 rebuilt: 306 subjects

3 Step 1: Derive Safety Population Flag (SAFFL)

Safety Population: Subjects who received at least one dose of study treatment.

# Derive SAFFL
adsl <- adsl %>%
  derive_var_merged_exist_flag(
    dataset_add = ex,
    by_vars = exprs(STUDYID, USUBJID),
    new_var = SAFFL,
    condition = (EXDOSE > 0 | (EXDOSE == 0 & str_detect(EXTRT, "PLACEBO")))
  )

cat("\nSafety population flag derived\n")


Safety population flag derived

# Distribution
adsl %>%
  count(SAFFL, TRT01A)

# A tibble: 4 × 3
  SAFFL TRT01A                   n
  <chr> <chr>                <int>
1 Y     Placebo                 86
2 Y     Xanomeline High Dose    72
3 Y     Xanomeline Low Dose     96
4 <NA>  Screen Failure          52

Logic:

SAFFL = “Y” if subject has any EX record with EXDOSE > 0
Used for safety analyses (AEs, labs, vitals)

4 Step 2: Derive ITT Population Flag (ITTFL)

Intent-to-Treat Population: Randomized subjects.

# Derive ITTFL
adsl <- adsl %>%
  derive_var_merged_exist_flag(
    dataset_add = ds,
    by_vars = exprs(STUDYID, USUBJID),
    new_var = ITTFL,
    condition = DSDECOD == "RANDOMIZED"
  )

cat("\nITT population flag derived\n")


ITT population flag derived

# Distribution
adsl %>%
  count(ITTFL, SAFFL)

# A tibble: 2 × 3
  ITTFL SAFFL     n
  <chr> <chr> <int>
1 Y     Y       254
2 <NA>  <NA>     52

Logic:

ITTFL = “Y” if randomization event exists in DS
Used for efficacy analyses

5 Step 3: Create Age Groupings

# Derive age group categories
adsl <- adsl %>%
  mutate(
    AGEGR1 = case_when(
      AGE < 65 ~ "<65",
      AGE >= 65 & AGE < 75 ~ "65-74",
      AGE >= 75 ~ ">=75",
      TRUE ~ NA_character_
    ),
    AGEGR1N = case_when(
      AGE < 65 ~ 1,
      AGE >= 65 & AGE < 75 ~ 2,
      AGE >= 75 ~ 3,
      TRUE ~ NA_real_
    )
  )

cat("\nAge groupings created\n")


Age groupings created

# Distribution
adsl %>%
  count(AGEGR1, AGEGR1N)

# A tibble: 3 × 3
  AGEGR1 AGEGR1N     n
  <chr>    <dbl> <int>
1 65-74        2    85
2 <65          1    42
3 >=75         3   179

Age Groups:

<65: Younger adults (N=1)
65-74: Older adults (N=2)
=75: Elderly (N=3)

6 Step 4: Derive Numeric Demographics

# Numeric versions of categorical variables
adsl <- adsl %>%
  mutate(
    SEXN = case_when(
      SEX == "F" ~ 1,
      SEX == "M" ~ 2,
      TRUE ~ NA_real_
    ),
    RACEN = case_when(
      RACE == "AMERICAN INDIAN OR ALASKA NATIVE" ~ 1,
      RACE == "ASIAN" ~ 2,
      RACE == "BLACK OR AFRICAN AMERICAN" ~ 3,
      RACE == "NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER" ~ 4,
      RACE == "WHITE" ~ 5,
      RACE == "MULTIPLE" ~ 6,
      TRUE ~ NA_real_
    )
  )

cat("\nNumeric demographics derived\n")


Numeric demographics derived

# Check mappings
adsl %>%
  count(SEX, SEXN)

# A tibble: 2 × 3
  SEX    SEXN     n
  <chr> <dbl> <int>
1 F         1   179
2 M         2   127

adsl %>%
  count(RACE, RACEN) %>%
  head(10)

# A tibble: 4 × 3
  RACE                             RACEN     n
  <chr>                            <dbl> <int>
1 AMERICAN INDIAN OR ALASKA NATIVE     1     2
2 ASIAN                                2     2
3 BLACK OR AFRICAN AMERICAN            3    29
4 WHITE                                5   273

Why numeric versions?

Enable table sorting
Required for some stats procedures
Standard across studies

7 Step 5: Prepare VS for Baseline Merging

# Filter VS for baseline
vs_bl <- vs %>%
  filter(VISIT %in% c("BASELINE", "SCREENING") | VISITNUM == 0) %>%
  derive_vars_dt(dtc = VSDTC, new_vars_prefix = "VS")

# Available baseline measurements
vs_bl %>%
  count(VSTESTCD)

# A tibble: 5 × 2
  VSTESTCD     n
  <chr>    <int>
1 DIABP      759
2 PULSE      759
3 SYSBP      759
4 TEMP       253
5 WEIGHT     253

8 Step 6: Derive Baseline Height

# Merge baseline height
adsl <- adsl %>%
  derive_vars_merged(
    dataset_add = vs_bl,
    by_vars = exprs(STUDYID, USUBJID),
    filter_add = VSTESTCD == "HEIGHT" & !is.na(VSSTRESN),
    new_vars = exprs(HEIGHTBL = VSSTRESN),
    order = exprs(VSDT),
    mode = "last"
  )

cat("\nBaseline height merged\n")


Baseline height merged

# Summary
adsl %>%
  filter(!is.na(HEIGHTBL)) %>%
  summarise(
    N = n(),
    Mean = round(mean(HEIGHTBL), 1),
    SD = round(sd(HEIGHTBL), 1),
    Median = median(HEIGHTBL),
    Min = min(HEIGHTBL),
    Max = max(HEIGHTBL)
  )

# A tibble: 1 × 6
      N  Mean    SD Median   Min   Max
  <int> <dbl> <dbl>  <dbl> <dbl> <dbl>
1     0   NaN    NA     NA   Inf  -Inf

Logic: Take last height at or before baseline

9 Step 7: Derive Baseline Weight

# Merge baseline weight
adsl <- adsl %>%
  derive_vars_merged(
    dataset_add = vs_bl,
    by_vars = exprs(STUDYID, USUBJID),
    filter_add = VSTESTCD == "WEIGHT" & !is.na(VSSTRESN),
    new_vars = exprs(WEIGHTBL = VSSTRESN),
    order = exprs(VSDT),
    mode = "last"
  )

cat("\nBaseline weight merged\n")


Baseline weight merged

# Summary
adsl %>%
  filter(!is.na(WEIGHTBL)) %>%
  summarise(
    N = n(),
    Mean = round(mean(WEIGHTBL), 1),
    SD = round(sd(WEIGHTBL), 1),
    Median = median(WEIGHTBL),
    Min = min(WEIGHTBL),
    Max = max(WEIGHTBL)
  )

# A tibble: 1 × 6
      N  Mean    SD Median   Min   Max
  <int> <dbl> <dbl>  <dbl> <dbl> <dbl>
1   253  66.6  14.1   66.7  34.0  108.

10 Step 8: Calculate Baseline BMI

# Calculate BMI
adsl <- adsl %>%
  mutate(
    BMIBL = case_when(
      !is.na(HEIGHTBL) & !is.na(WEIGHTBL) ~ round(WEIGHTBL / (HEIGHTBL / 100)^2, 1),
      TRUE ~ NA_real_
    )
  )

cat("\nBaseline BMI calculated\n")


Baseline BMI calculated

# Summary
adsl %>%
  filter(!is.na(BMIBL)) %>%
  summarise(
    N = n(),
    Mean = round(mean(BMIBL), 1),
    SD = round(sd(BMIBL), 1),
    Median = median(BMIBL),
    Min = min(BMIBL),
    Max = max(BMIBL)
  )

# A tibble: 1 × 6
      N  Mean    SD Median   Min   Max
  <int> <dbl> <dbl>  <dbl> <dbl> <dbl>
1     0   NaN    NA     NA   Inf  -Inf

Formula: BMI = Weight (kg) / (Height (m))²

11 Step 9: Create BMI Groups

# BMI groupings
adsl <- adsl %>%
  mutate(
    BMIBLGR1 = case_when(
      BMIBL < 18.5 ~ "<18.5",
      BMIBL >= 18.5 & BMIBL < 25 ~ "18.5-<25",
      BMIBL >= 25 & BMIBL < 30 ~ "25-<30",
      BMIBL >= 30 ~ ">=30",
      TRUE ~ NA_character_
    )
  )

# Distribution
adsl %>%
  count(BMIBLGR1)

# A tibble: 1 × 2
  BMIBLGR1     n
  <chr>    <int>
1 <NA>       306

BMI Categories:

<18.5: Underweight
18.5-<25: Normal
25-<30: Overweight
=30: Obese

12 Validation Checks

cat("\n=== ADSL Validation ===\n\n")


=== ADSL Validation ===

# Check 1: SAFFL consistency
check1 <- adsl %>%
  filter(SAFFL == "Y" & is.na(TRTSDT))
cat("Check 1 - SAFFL='Y' but no TRTSDT:", nrow(check1), "\n")

Check 1 - SAFFL='Y' but no TRTSDT: 0

# Check 2: Age groups
check2 <- adsl %>%
  filter(!is.na(AGEGR1) & is.na(AGEGR1N))
cat("Check 2 - AGEGR1 without AGEGR1N:", nrow(check2), "\n")

Check 2 - AGEGR1 without AGEGR1N: 0

# Check 3: BMI calculation
check3 <- adsl %>%
  filter(!is.na(BMIBL)) %>%
  mutate(
    BMI_check = round(WEIGHTBL / (HEIGHTBL / 100)^2, 1),
    Match = abs(BMIBL - BMI_check) < 0.1
  ) %>%
  filter(!Match)
cat("Check 3 - BMI mismatch:", nrow(check3), "\n")

Check 3 - BMI mismatch: 0

# Check 4: Numeric demographics
check4 <- adsl %>%
  filter(!is.na(SEX) & is.na(SEXN))
cat("Check 4 - SEX without SEXN:", nrow(check4), "\n")

Check 4 - SEX without SEXN: 0

cat("\n✓ Validation complete\n")


✓ Validation complete

13 Complete ADSL Summary

cat("\n=== Complete ADSL ===\n\n")


=== Complete ADSL ===

cat("Total subjects:", nrow(adsl), "\n")

Total subjects: 306

cat("Variables:", ncol(adsl), "\n\n")

Variables: 42

# Population flags
cat("Population Flags:\n")

Population Flags:

adsl %>%
  summarise(
    SAFFL_Y = sum(SAFFL == "Y", na.rm = TRUE),
    ITTFL_Y = sum(ITTFL == "Y", na.rm = TRUE)
  )

# A tibble: 1 × 2
  SAFFL_Y ITTFL_Y
    <int>   <int>
1     254     254

# Demographics by treatment (Safety population)
cat("\nDemographics by Treatment (Safety Population):\n")


Demographics by Treatment (Safety Population):

adsl %>%
  filter(SAFFL == "Y") %>%
  group_by(TRT01A) %>%
  summarise(
    N = n(),
    Age_Mean = round(mean(AGE), 1),
    Age_SD = round(sd(AGE), 1),
    Female_N = sum(SEX == "F"),
    Female_Pct = round(100 * sum(SEX == "F") / n(), 1),
    BMI_Mean = round(mean(BMIBL, na.rm = TRUE), 1),
    .groups = "drop"
  )

# A tibble: 3 × 7
  TRT01A                   N Age_Mean Age_SD Female_N Female_Pct BMI_Mean
  <chr>                <int>    <dbl>  <dbl>    <int>      <dbl>    <dbl>
1 Placebo                 86     75.2    8.6       53       61.6      NaN
2 Xanomeline High Dose    72     73.8    7.9       35       48.6      NaN
3 Xanomeline Low Dose     96     76      8.1       55       57.3      NaN

14 Export Complete ADSL

# Apply variable labels
attr(adsl$SAFFL, "label") <- "Safety Population Flag"
attr(adsl$ITTFL, "label") <- "Intent-to-Treat Population Flag"
attr(adsl$AGEGR1, "label") <- "Age Group 1"
attr(adsl$AGEGR1N, "label") <- "Age Group 1 (N)"
attr(adsl$SEXN, "label") <- "Sex (N)"
attr(adsl$RACEN, "label") <- "Race (N)"
attr(adsl$HEIGHTBL, "label") <- "Baseline Height (cm)"
attr(adsl$WEIGHTBL, "label") <- "Baseline Weight (kg)"
attr(adsl$BMIBL, "label") <- "Baseline Body Mass Index"
attr(adsl$BMIBLGR1, "label") <- "Baseline BMI Group 1"

# Export as XPT
xportr_write(adsl, path = "adsl.xpt", domain = "ADSL")

cat("\n✓ ADSL exported to: adsl.xpt\n")


✓ ADSL exported to: adsl.xpt

15 Key Takeaways

Population Flags:

SAFFL = received treatment (from EX)
ITTFL = randomized (from DS)
Both derived with derive_var_merged_exist_flag()

Demographics:

Always create numeric versions (AGEGR1N, SEXN, RACEN)
Numeric enables sorting and statistics
Character for display in tables

Baseline Measurements:

Merge from VS using derive_vars_merged()
Take last baseline measurement
Calculate BMI from HEIGHT and WEIGHT

Validation:

Check flag consistency (SAFFL=“Y” must have TRTSDT)
Verify calculations (BMI formula)
Ensure char/numeric pairs match

16 Next Steps

Day 18: ADAE (Adverse Events) - OCCDS structure
Day 19: ADLB (Labs) - BDS with baseline
Day 20: ADVS (Vitals) - BDS with visits
Day 21: ADTTE (Time-to-Event)

ADSL is complete - ready for all other ADaMs!

17 Resources

Admiral:

ADSL Vignette: https://pharmaverse.github.io/admiral/articles/adsl.html
derive_var_merged_exist_flag: https://pharmaverse.github.io/admiral/reference/derive_var_merged_exist_flag.html

CDISC:

ADaM IG v1.3: https://www.cdisc.org/standards/foundational/adam

Pharmaverse:

pharmaverse.org: https://pharmaverse.org

End of Day 17

Tomorrow: ADAE (Adverse Events)