Module 6 — TLF Figures: Exercise Solutions

✅ Module 6 — Solution: Clinical Plots with ggplot2

🧪 Dataset Setup

library(ggplot2)
library(dplyr)

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

library(tidyr)
library(patchwork)

set.seed(2025)
labs <- tibble::tibble(
  USUBJID = rep(paste0("SUBJ", 1:10), each = 3),
  VISIT = rep(c("Day 1", "Week 2", "Week 4"), times = 10),
  ARM = rep(c("Placebo", "Drug"), each = 15),
  HGB = round(rnorm(30, mean = 13, sd = 1.2), 1),
  HCT = round(rnorm(30, mean = 38, sd = 3), 1)
)

labs_long <- labs %>%
  pivot_longer(cols = c(HGB, HCT), names_to = "Test", values_to = "Value")

---

1. Line Plot: HGB & HCT Trajectories per Subject

ggplot(labs_long, aes(x = VISIT, y = Value, group = USUBJID, color = ARM)) +
  geom_line(alpha = 0.7) +
  facet_wrap(~Test, scales = "free_y") +
  theme_minimal() +
  labs(title = "Subject-Level Lab Value Trends", y = "Value")

---

2. Boxplot: HGB by Visit and Treatment

ggplot(labs, aes(x = VISIT, y = HGB, fill = ARM)) +
  geom_boxplot() +
  theme_minimal() +
  labs(title = "HGB Distribution by Visit and Arm", y = "Hemoglobin (g/dL)")

---

3. Heatmap: HGB per Subject and Visit

labs_heat <- labs %>%
  select(USUBJID, VISIT, HGB) %>%
  pivot_wider(names_from = VISIT, values_from = HGB)

labs_heat_long <- labs_heat %>%
  pivot_longer(-USUBJID, names_to = "Visit", values_to = "HGB")

ggplot(labs_heat_long, aes(x = Visit, y = USUBJID, fill = HGB)) +
  geom_tile(color = "white") +
  scale_fill_gradient(low = "#ffffcc", high = "#990000") +
  theme_minimal() +
  labs(title = "Heatmap of HGB Values", fill = "HGB (g/dL)")

---

4. Mean ± SD of HCT by Visit and Arm

summary_hct <- labs %>%
  group_by(VISIT, ARM) %>%
  summarise(mean_val = mean(HCT), sd_val = sd(HCT), .groups = "drop")

ggplot(summary_hct, aes(x = VISIT, y = mean_val, color = ARM, group = ARM)) +
  geom_line(size = 1) +
  geom_point(size = 2) +
  geom_errorbar(aes(ymin = mean_val - sd_val, ymax = mean_val + sd_val), width = 0.1) +
  theme_minimal() +
  labs(title = "Mean ± SD of HCT", y = "Hematocrit (%)")

Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.

---

5. BONUS: Combine Plots with patchwork

# Note: patchwork plotting code temporarily removed due to rendering issues
# p1 <- ggplot(labs, aes(x = VISIT, y = HGB, fill = ARM)) +
#   geom_boxplot() + theme_minimal()
# 
# p2 <- ggplot(summary_hct, aes(x = VISIT, y = mean_val, color = ARM)) +
#   geom_line() + geom_point() + theme_minimal()
# 
# p1 + p2

print("Patchwork plotting functionality will be restored in a future update")

[1] "Patchwork plotting functionality will be restored in a future update"