Skip to contents

Stratified Table 1 Analysis Examples

zztable1

2026-05-02

Source: vignettes/stratified_examples.Rmd
stratified_examples.Rmd

Introduction

Stratified analysis is a crucial component of epidemiological and clinical research. This vignette demonstrates how to create stratified Table 1 analyses using the zztable1 package. Stratified tables allow you to examine patterns within subgroups of your data, revealing important differences that might be obscured in overall analyses.

Clinical Trial Data Example

Let’s create a comprehensive clinical trial dataset that includes multiple potential stratification variables.

# Create a realistic multi-center clinical trial dataset
set.seed(42)
n <- 300

clinical_data <- data.frame(
  # Primary treatment variable
  treatment = factor(
    sample(c("Placebo", "Low Dose", "High Dose"), n, replace = TRUE, 
           prob = c(0.4, 0.3, 0.3)),
    levels = c("Placebo", "Low Dose", "High Dose")
  ),
  
  # Potential stratification variables
  site = factor(sample(paste("Site", LETTERS[1:4]), n, replace = TRUE)),
  sex = factor(sample(c("Male", "Female"), n, replace = TRUE, prob = c(0.55, 0.45))),
  age_group = factor(
    sample(c("18-44", "45-64", "65+"), n, replace = TRUE, prob = c(0.3, 0.4, 0.3)),
    levels = c("18-44", "45-64", "65+")
  ),
  disease_severity = factor(
    sample(c("Mild", "Moderate", "Severe"), n, replace = TRUE, prob = c(0.4, 0.4, 0.2)),
    levels = c("Mild", "Moderate", "Severe")
  ),
  
  # Baseline characteristics
  age = round(rnorm(n, 58, 15)),
  bmi = round(rnorm(n, 26.5, 4.2), 1),
  systolic_bp = round(rnorm(n, 135, 18)),
  
  # Comorbidities
  diabetes = factor(sample(c("No", "Yes"), n, replace = TRUE, prob = c(0.75, 0.25))),
  hypertension = factor(sample(c("No", "Yes"), n, replace = TRUE, prob = c(0.65, 0.35))),
  
  # Lab values
  hemoglobin = round(rnorm(n, 13.2, 1.8), 1),
  creatinine = round(rnorm(n, 1.1, 0.3), 2)
)

# Add some realistic missing values
clinical_data$bmi[sample(1:n, 8)] <- NA
clinical_data$hemoglobin[sample(1:n, 5)] <- NA
clinical_data$creatinine[sample(1:n, 3)] <- NA

# Show dataset structure
str(clinical_data)

‘data.frame’: 300 obs. of 12 variables: $ treatment : Factor w/ 3 levels “Placebo”,“Low Dose”,..: 2 2 1 2 3 3 2 1 3 2 … $ site : Factor w/ 4 levels “Site A”,“Site B”,..: 1 4 2 4 1 1 2 3 1 2 … $ sex : Factor w/ 2 levels “Female”,“Male”: 2 2 1 2 2 2 1 2 2 2 … $ age_group : Factor w/ 3 levels “18-44”,“45-64”,..: 2 2 3 2 3 1 2 2 3 2 … $ disease_severity: Factor w/ 3 levels “Mild”,“Moderate”,..: 1 1 1 2 3 3 1 1 2 1 … $ age : num 34 29 53 61 24 62 64 77 60 61 … $ bmi : num 27.5 28.3 26.5 29.1 28.3 28.8 28.6 33.5 22.1 32.8 … $ systolic_bp : num 128 132 148 126 144 145 117 134 125 143 … $ diabetes : Factor w/ 2 levels “No”,“Yes”: 1 1 1 2 1 1 1 1 1 1 … $ hypertension : Factor w/ 2 levels “No”,“Yes”: 1 1 1 2 1 1 1 2 1 1 … $ hemoglobin : num 13.9 12.4 13.4 12 14.6 10.1 13.5 13.2 15.1 12.8 … $ creatinine : num 0.94 1.54 1.23 0.79 0.71 1.17 1.01 1.07 1.05 1.3 …

head(clinical_data, 10)

treatment site sex age_group disease_severity age bmi systolic_bp 1 Low Dose Site A Male 45-64 Mild 34 27.5 128 2 Low Dose Site D Male 45-64 Mild 29 28.3 132 3 Placebo Site B Female 65+ Mild 53 26.5 148 4 Low Dose Site D Male 45-64 Moderate 61 29.1 126 5 High Dose Site A Male 65+ Severe 24 28.3 144 6 High Dose Site A Male 18-44 Severe 62 28.8 145 7 Low Dose Site B Female 45-64 Mild 64 28.6 117 8 Placebo Site C Male 45-64 Mild 77 33.5 134 9 High Dose Site A Male 65+ Moderate 60 22.1 125 10 Low Dose Site B Male 45-64 Mild 61 32.8 143 diabetes hypertension hemoglobin creatinine 1 No No 13.9 0.94 2 No No 12.4 1.54 3 No No 13.4 1.23 4 Yes Yes 12.0 0.79 5 No No 14.6 0.71 6 No No 10.1 1.17 7 No No 13.5 1.01 8 No Yes 13.2 1.07 9 No No 15.1 1.05 10 No No 12.8 1.30

Stratified Analysis Examples

Example 1: Stratified by Study Site

Understanding how baseline characteristics vary across different study sites is crucial for multi-center trials.

create_table(
  treatment ~ age + sex + bmi + diabetes + systolic_bp,
  data = clinical_data,
  strata = "site",
  theme = "nejm",
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Site: Site A
age 57 ± 13 57.2 ± 19.7 61.6 ± 17.1 NaN ± NA 0.822
sex – no. (%) 75 (267.9) 52 (247.6) 45 (160.7) 172 (223.4) 0.558
bmi 26.9 ± 3.6 26.9 ± 4.4 26.4 ± 4.8 NaN ± NA 0.746
diabetes – no. (%) 36 (128.6) 29 (138.1) 25 (89.3) 90 (116.9) 0.987
systolic_bp 138.9 ± 14.8 134 ± 17.3 134.1 ± 14.7 NaN ± NA 0.724
Site: Site D
age 62.5 ± 16.1 60.6 ± 14.9 56.3 ± 12.6 NaN ± NA 0.822
sex – no. (%) 75 (340.9) 52 (371.4) 45 (281.2) 172 (330.8) 0.558
bmi 26.9 ± 3.5 27.6 ± 5.4 25.9 ± 3 NaN ± NA 0.746
diabetes – no. (%) 36 (163.6) 29 (207.1) 25 (156.2) 90 (173.1) 0.987
systolic_bp 131.9 ± 16.7 136 ± 18.9 135.8 ± 12.6 NaN ± NA 0.724
Site: Site B
age 54.7 ± 16.6 54.2 ± 12.7 56 ± 19.3 NaN ± NA 0.822
sex – no. (%) 75 (182.9) 52 (152.9) 45 (214.3) 172 (179.2) 0.558
bmi 27.1 ± 4.2 26.2 ± 4.2 27.2 ± 4.4 NaN ± NA 0.746
diabetes – no. (%) 36 (87.8) 29 (85.3) 25 (119) 90 (93.8) 0.987
systolic_bp 135.6 ± 18.2 131.4 ± 18.7 134.2 ± 21.7 NaN ± NA 0.724
Site: Site C
age 58.2 ± 18.3 58.7 ± 16.4 61 ± 14.3 NaN ± NA 0.822
sex – no. (%) 75 (234.4) 52 (208) 45 (250) 172 (229.3) 0.558
bmi 25.2 ± 4 26.9 ± 3.9 24.9 ± 4.4 NaN ± NA 0.746
diabetes – no. (%) 36 (112.5) 29 (116) 25 (138.9) 90 (120) 0.987
systolic_bp 137.3 ± 15.9 141.2 ± 17.7 129.2 ± 19.9 NaN ± NA 0.724

Key Observations: - Each study site shows as a separate table section - Within-site treatment group comparisons - Helps identify site-specific recruitment patterns - Essential for assessing treatment balance across sites

Example 2: Stratified by Sex

Sex-stratified analysis is important for understanding treatment effects and baseline differences between male and female participants.

create_table(
  treatment ~ age + age_group + bmi + diabetes + hypertension + hemoglobin,
  data = clinical_data,
  strata = "sex", 
  theme = "lancet",
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Sex: Male
age 58.2 (17.7) 54.4 (16) 59.6 (15.8) NaN (NA) 0.822
age_group
18-44 21 (28%) 13 (25%) 12 (26.7%) 0 (0%) 0.936
45-64 33 (44%) 25 (48.1%) 18 (40%) 0 (0%)
65+ 21 (28%) 14 (26.9%) 15 (33.3%) 0 (0%)
bmi 26.9 (4.1) 26.9 (4.3) 26.3 (3.3) NaN (NA) 0.746
diabetes
No 50 (66.7%) 36 (69.2%) 32 (71.1%) 0 (0%) 0.987
Yes 25 (33.3%) 16 (30.8%) 13 (28.9%) 0 (0%)
hypertension
No 45 (60%) 31 (59.6%) 33 (73.3%) 0 (0%) 0.336
Yes 30 (40%) 21 (40.4%) 12 (26.7%) 0 (0%)
hemoglobin 13.3 (1.7) 13 (1.8) 13.7 (1.7) NaN (NA) 0.329
Sex: Female
age 56.5 (13.8) 60.2 (14.9) 58.4 (17) NaN (NA) 0.822
age_group
18-44 17 (35.4%) 12 (28.6%) 13 (34.2%) 0 (0%) 0.932
45-64 16 (33.3%) 17 (40.5%) 18 (47.4%) 0 (0%)
65+ 15 (31.2%) 13 (31%) 7 (18.4%) 0 (0%)
bmi 26.1 (3.6) 26.5 (4.4) 26.1 (5.3) NaN (NA) 0.746
diabetes
No 37 (77.1%) 29 (69%) 26 (68.4%) 0 (0%) 0.987
Yes 11 (22.9%) 13 (31%) 12 (31.6%) 0 (0%)
hypertension
No 27 (56.2%) 28 (66.7%) 24 (63.2%) 0 (0%) 0.336
Yes 21 (43.8%) 14 (33.3%) 14 (36.8%) 0 (0%)
hemoglobin 13.3 (2) 13.1 (1.9) 12.4 (1.7) NaN (NA) 0.329

Key Observations: - Separate baseline characteristics for males and females - Age and BMI distributions may differ by sex - Hemoglobin levels typically differ between sexes - Treatment allocation balance within each sex

Example 3: Stratified by Disease Severity

Disease severity stratification helps understand how patient characteristics and treatment allocation vary by baseline disease status.

create_table(
  treatment ~ age + sex + bmi + systolic_bp + diabetes + hypertension + creatinine,
  data = clinical_data,
  strata = "disease_severity",
  theme = "jama",
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Disease_severity: Mild
age 56.9 (15.8) 57.8 (15.5) 58.5 (17.2) NaN (NA) 0.822
sex
Female 23 (43.4%) 22 (50%) 9 (34.6%) 0 (0%) 0.558
Male 30 (56.6%) 22 (50%) 17 (65.4%) 0 (0%)
bmi 26.7 (3.7) 26.8 (4.4) 25.2 (4.7) NaN (NA) 0.746
systolic_bp 137.2 (16.4) 135.9 (19.3) 135.7 (19.5) NaN (NA) 0.724
diabetes
No 43 (81.1%) 32 (72.7%) 15 (57.7%) 0 (0%) 0.987
Yes 10 (18.9%) 12 (27.3%) 11 (42.3%) 0 (0%)
hypertension
No 30 (56.6%) 26 (59.1%) 17 (65.4%) 0 (0%) 0.336
Yes 23 (43.4%) 18 (40.9%) 9 (34.6%) 0 (0%)
creatinine 1.1 (0.3) 1 (0.4) 1.1 (0.3) NaN (NA) 0.777
Disease_severity: Moderate
age 56.4 (16) 53.8 (14.8) 59.2 (16.4) NaN (NA) 0.822
sex
Female 20 (39.2%) 12 (35.3%) 21 (52.5%) 0 (0%) 0.558
Male 31 (60.8%) 22 (64.7%) 19 (47.5%) 0 (0%)
bmi 26.7 (4.3) 26.5 (4.6) 27 (4.1) NaN (NA) 0.746
systolic_bp 135.9 (17.8) 134.2 (16) 133.2 (16.5) NaN (NA) 0.724
diabetes
No 31 (60.8%) 24 (70.6%) 30 (75%) 0 (0%) 0.987
Yes 20 (39.2%) 10 (29.4%) 10 (25%) 0 (0%)
hypertension
No 35 (68.6%) 24 (70.6%) 32 (80%) 0 (0%) 0.336
Yes 16 (31.4%) 10 (29.4%) 8 (20%) 0 (0%)
creatinine 1.1 (0.3) 1.2 (0.3) 1.1 (0.3) NaN (NA) 0.777
Disease_severity: Severe
age 62.1 (18.5) 61.9 (17.5) 59.5 (15.4) NaN (NA) 0.822
sex
Female 5 (26.3%) 8 (50%) 8 (47.1%) 0 (0%) 0.558
Male 14 (73.7%) 8 (50%) 9 (52.9%) 0 (0%)
bmi 26 (3.7) 27.3 (3.8) 25.7 (3.9) NaN (NA) 0.746
systolic_bp 133.8 (14) 136.1 (20.8) 130.5 (16.6) NaN (NA) 0.724
diabetes
No 13 (68.4%) 9 (56.2%) 13 (76.5%) 0 (0%) 0.987
Yes 6 (31.6%) 7 (43.8%) 4 (23.5%) 0 (0%)
hypertension
No 7 (36.8%) 9 (56.2%) 8 (47.1%) 0 (0%) 0.336
Yes 12 (63.2%) 7 (43.8%) 9 (52.9%) 0 (0%)
creatinine 1.3 (0.3) 1.2 (0.3) 1.2 (0.2) NaN (NA) 0.777

Key Observations: - Baseline characteristics across mild, moderate, and severe disease - Treatment allocation may vary by severity - Comorbidity prevalence often increases with disease severity - Important for stratified randomization assessment

Example 4: Stratified by Age Group

Age-group stratified analysis reveals how baseline characteristics and treatment allocation vary across different age ranges.

create_table(
  treatment ~ sex + bmi + systolic_bp + diabetes + hypertension + hemoglobin + creatinine,
  data = clinical_data,
  strata = "age_group",
  theme = "nejm",
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Age_group: 45-64
sex – no. (%) 75 (153.1) 52 (123.8) 45 (125) 172 (135.4) 0.558
bmi 26.9 ± 3.9 25.7 ± 4.5 25.5 ± 4.5 NaN ± NA 0.746
systolic_bp 135.7 ± 15.7 134.8 ± 19.9 131.8 ± 18.5 NaN ± NA 0.724
diabetes – no. (%) 36 (73.5) 29 (69) 25 (69.4) 90 (70.9) 0.987
hypertension – no. (%) 51 (104.1) 35 (83.3) 26 (72.2) 112 (88.2) 0.336
hemoglobin 13.6 ± 1.7 13.1 ± 1.5 13.2 ± 1.8 NaN ± NA 0.329
creatinine 1.1 ± 0.3 1.2 ± 0.4 1 ± 0.3 NaN ± NA 0.777
Age_group: 65+
sex – no. (%) 75 (208.3) 52 (192.6) 45 (204.5) 172 (202.4) 0.558
bmi 26.1 ± 4.1 28 ± 4.9 26.4 ± 3.6 NaN ± NA 0.746
systolic_bp 134.2 ± 15.1 133.6 ± 17.3 132.5 ± 17.9 NaN ± NA 0.724
diabetes – no. (%) 36 (100) 29 (107.4) 25 (113.6) 90 (105.9) 0.987
hypertension – no. (%) 51 (141.7) 35 (129.6) 26 (118.2) 112 (131.8) 0.336
hemoglobin 13.1 ± 1.8 13 ± 2.1 13.3 ± 1.7 NaN ± NA 0.329
creatinine 1.2 ± 0.3 1 ± 0.3 1.1 ± 0.3 NaN ± NA 0.777
Age_group: 18-44
sex – no. (%) 75 (197.4) 52 (208) 45 (180) 172 (195.5) 0.558
bmi 26.6 ± 3.8 27.1 ± 3 27 ± 4.6 NaN ± NA 0.746
systolic_bp 138.5 ± 19 138 ± 16.8 136.5 ± 15.4 NaN ± NA 0.724
diabetes – no. (%) 36 (94.7) 29 (116) 25 (100) 90 (102.3) 0.987
hypertension – no. (%) 51 (134.2) 35 (140) 26 (104) 112 (127.3) 0.336
hemoglobin 13.1 ± 1.9 13 ± 2.1 12.7 ± 1.9 NaN ± NA 0.329
creatinine 1.1 ± 0.3 1.1 ± 0.2 1.2 ± 0.3 NaN ± NA 0.777

Key Observations: - Younger vs middle-aged vs older participants - Comorbidity prevalence increases with age - Lab values may vary by age group - Treatment allocation balance across age groups

Advanced Stratified Analysis

Multiple Variables with Missing Data

Let’s examine how stratified analysis handles missing data and multiple variable types.

create_table(
  treatment ~ age + bmi + hemoglobin + creatinine + diabetes + hypertension,
  data = clinical_data,
  strata = "sex",
  theme = "lancet", 
  missing = TRUE,  # Show missing value patterns
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Sex: Male
age 58.2 (17.7) 54.4 (16) 59.6 (15.8) NaN (NA) 0.822
bmi 26.9 (4.1) 26.9 (4.3) 26.3 (3.3) NaN (NA) 0.746
hemoglobin 13.3 (1.7) 13 (1.8) 13.7 (1.7) NaN (NA) 0.329
creatinine 1.1 (0.3) 1.2 (0.3) 1.1 (0.3) NaN (NA) 0.777
diabetes
No 50 (66.7%) 36 (69.2%) 32 (71.1%) 0 (0%) 0.987
Yes 25 (33.3%) 16 (30.8%) 13 (28.9%) 0 (0%)
hypertension
No 45 (60%) 31 (59.6%) 33 (73.3%) 0 (0%) 0.336
Yes 30 (40%) 21 (40.4%) 12 (26.7%) 0 (0%)
Sex: Female
age 56.5 (13.8) 60.2 (14.9) 58.4 (17) NaN (NA) 0.822
bmi 26.1 (3.6) 26.5 (4.4) 26.1 (5.3) NaN (NA) 0.746
hemoglobin 13.3 (2) 13.1 (1.9) 12.4 (1.7) NaN (NA) 0.329
creatinine 1.2 (0.3) 1 (0.3) 1.1 (0.3) NaN (NA) 0.777
diabetes
No 37 (77.1%) 29 (69%) 26 (68.4%) 0 (0%) 0.987
Yes 11 (22.9%) 13 (31%) 12 (31.6%) 0 (0%)
hypertension
No 27 (56.2%) 28 (66.7%) 24 (63.2%) 0 (0%) 0.336
Yes 21 (43.8%) 14 (33.3%) 14 (36.8%) 0 (0%)

Site and Sex Combined Analysis

For comprehensive analysis, we might want to examine the interaction of multiple stratification factors.

# Create a combined stratification variable for demonstration
clinical_data$site_sex <- interaction(clinical_data$site, clinical_data$sex, sep = " - ")

# Show the distribution
table(clinical_data$site_sex, clinical_data$treatment)
              Placebo Low Dose High Dose

Site A - Female 10 8 13 Site B - Female 10 19 11 Site C - Female 20 10 8 Site D - Female 8 5 6 Site A - Male 18 13 15 Site B - Male 31 15 10 Site C - Male 12 15 10 Site D - Male 14 9 10

create_table(
  treatment ~ age + bmi + diabetes + systolic_bp,
  data = clinical_data,
  strata = "site_sex",
  theme = "jama",
  pvalue = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		P value
Site_sex: Site A - Male
age 58 (12.5) 51.9 (16.4) 58.9 (19.2) 0.822
bmi 25.9 (3.5) 27.2 (4.7) 26.9 (3.6) 0.746
diabetes
No 8 (44.4%) 9 (69.2%) 12 (80%) 0.987
Yes 10 (55.6%) 4 (30.8%) 3 (20%)
systolic_bp 137.5 (16.3) 133.2 (18.5) 129.1 (14.4) 0.724
Site_sex: Site D - Male
age 64.9 (17.6) 59.7 (16.2) 59.5 (13.6) 0.822
bmi 27.7 (3.9) 26.3 (4.8) 26.3 (2.5) 0.746
diabetes
No 10 (71.4%) 6 (66.7%) 8 (80%) 0.987
Yes 4 (28.6%) 3 (33.3%) 2 (20%)
systolic_bp 133.5 (15.5) 134.3 (18.6) 137.7 (15.3) 0.724
Site_sex: Site B - Female
age 52.1 (13.8) 54.3 (11.1) 50.5 (19.4) 0.822
bmi 25.8 (3.8) 25.8 (3.6) 28.5 (4.3) 0.746
diabetes
No 8 (80%) 17 (89.5%) 7 (63.6%) 0.987
Yes 2 (20%) 2 (10.5%) 4 (36.4%)
systolic_bp 135.5 (12.5) 123.7 (18.3) 129.5 (22.6) 0.724
Site_sex: Site C - Male
age 57.4 (24.2) 53.9 (17.4) 58.1 (11.2) 0.822
bmi 25.4 (4.4) 27.3 (3.4) 25.6 (2.8) 0.746
diabetes
No 9 (75%) 9 (60%) 7 (70%) 0.987
Yes 3 (25%) 6 (40%) 3 (30%)
systolic_bp 131.2 (12.3) 140.3 (20.5) 131.3 (15.5) 0.724
Site_sex: Site B - Male
age 55.5 (17.5) 54 (14.9) 62 (18.2) 0.822
bmi 27.6 (4.3) 26.8 (4.9) 25.9 (4.3) 0.746
diabetes
No 23 (74.2%) 12 (80%) 5 (50%) 0.987
Yes 8 (25.8%) 3 (20%) 5 (50%)
systolic_bp 135.7 (19.9) 141.2 (14.7) 139.5 (20.4) 0.724
Site_sex: Site C - Female
age 58.7 (14.4) 65.9 (12.1) 64.6 (17.5) 0.822
bmi 25.1 (3.9) 26.3 (4.7) 24.2 (5.9) 0.746
diabetes
No 15 (75%) 6 (60%) 7 (87.5%) 0.987
Yes 5 (25%) 4 (40%) 1 (12.5%)
systolic_bp 140.9 (16.9) 142.5 (13.2) 126.5 (25.2) 0.724
Site_sex: Site D - Female
age 58.2 (12.9) 62.4 (13.6) 51 (9.5) 0.822
bmi 25.5 (2.3) 29.9 (6.2) 25.2 (3.8) 0.746
diabetes
No 5 (62.5%) 2 (40%) 3 (50%) 0.987
Yes 3 (37.5%) 3 (60%) 3 (50%)
systolic_bp 129 (19.4) 139 (21.2) 132.5 (6.2) 0.724
Site_sex: Site A - Female
age 55.1 (14.3) 65.9 (22.7) 64.7 (14.5) 0.822
bmi 28.7 (3.2) 26.3 (4.3) 25.8 (5.9) 0.746
diabetes
No 9 (90%) 4 (50%) 9 (69.2%) 0.987
Yes 1 (10%) 4 (50%) 4 (30.8%)
systolic_bp 141.4 (12) 135.4 (16.2) 139.8 (13.4) 0.724

Comparative Analysis

Before and After Stratification

Let’s compare an overall analysis with a stratified analysis to see how stratification reveals important patterns.

Overall Analysis (Non-stratified) 

create_table(
  treatment ~ age + sex + bmi + diabetes + hypertension + systolic_bp,
  data = clinical_data,
  theme = "console",
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
age 57.5 (16.3) 57 (15.7) 59 (16.3) 57.8 (16.1) 0.822
sex
Female 48 (39%) 42 (45%) 38 (46%) 128 (43%) 0.558
Male 75 (61%) 52 (55%) 45 (54%) 172 (57%)
bmi 26.6 (3.9) 26.8 (4.3) 26.2 (4.3) 26.5 (4.2) 0.746
diabetes
No 87 (71%) 65 (69%) 58 (70%) 210 (70%) 0.987
Yes 36 (29%) 29 (31%) 25 (30%) 90 (30%)
hypertension
No 72 (59%) 59 (63%) 57 (69%) 188 (63%) 0.336
Yes 51 (41%) 35 (37%) 26 (31%) 112 (37%)
systolic_bp 136.1 (16.6) 135.3 (18.3) 133.4 (17.4) 135.1 (17.3) 0.724

Stratified by Disease Severity 

create_table(
  treatment ~ age + sex + bmi + diabetes + hypertension + systolic_bp,
  data = clinical_data,
  strata = "disease_severity",
  theme = "console", 
  pvalue = TRUE,
  totals = TRUE
)
Variable Placebo
(N=123) 		Low Dose
(N=94) 		High Dose
(N=83) 		Total
(N=300) 		P value
Disease_severity: Mild
age 56.9 (15.8) 57.8 (15.5) 58.5 (17.2) NaN (NA) 0.822
sex
Female 23 (43.4%) 22 (50%) 9 (34.6%) 0 (0%) 0.558
Male 30 (56.6%) 22 (50%) 17 (65.4%) 0 (0%)
bmi 26.7 (3.7) 26.8 (4.4) 25.2 (4.7) NaN (NA) 0.746
diabetes
No 43 (81.1%) 32 (72.7%) 15 (57.7%) 0 (0%) 0.987
Yes 10 (18.9%) 12 (27.3%) 11 (42.3%) 0 (0%)
hypertension
No 30 (56.6%) 26 (59.1%) 17 (65.4%) 0 (0%) 0.336
Yes 23 (43.4%) 18 (40.9%) 9 (34.6%) 0 (0%)
systolic_bp 137.2 (16.4) 135.9 (19.3) 135.7 (19.5) NaN (NA) 0.724
Disease_severity: Moderate
age 56.4 (16) 53.8 (14.8) 59.2 (16.4) NaN (NA) 0.822
sex
Female 20 (39.2%) 12 (35.3%) 21 (52.5%) 0 (0%) 0.558
Male 31 (60.8%) 22 (64.7%) 19 (47.5%) 0 (0%)
bmi 26.7 (4.3) 26.5 (4.6) 27 (4.1) NaN (NA) 0.746
diabetes
No 31 (60.8%) 24 (70.6%) 30 (75%) 0 (0%) 0.987
Yes 20 (39.2%) 10 (29.4%) 10 (25%) 0 (0%)
hypertension
No 35 (68.6%) 24 (70.6%) 32 (80%) 0 (0%) 0.336
Yes 16 (31.4%) 10 (29.4%) 8 (20%) 0 (0%)
systolic_bp 135.9 (17.8) 134.2 (16) 133.2 (16.5) NaN (NA) 0.724
Disease_severity: Severe
age 62.1 (18.5) 61.9 (17.5) 59.5 (15.4) NaN (NA) 0.822
sex
Female 5 (26.3%) 8 (50%) 8 (47.1%) 0 (0%) 0.558
Male 14 (73.7%) 8 (50%) 9 (52.9%) 0 (0%)
bmi 26 (3.7) 27.3 (3.8) 25.7 (3.9) NaN (NA) 0.746
diabetes
No 13 (68.4%) 9 (56.2%) 13 (76.5%) 0 (0%) 0.987
Yes 6 (31.6%) 7 (43.8%) 4 (23.5%) 0 (0%)
hypertension
No 7 (36.8%) 9 (56.2%) 8 (47.1%) 0 (0%) 0.336
Yes 12 (63.2%) 7 (43.8%) 9 (52.9%) 0 (0%)
systolic_bp 133.8 (14) 136.1 (20.8) 130.5 (16.6) NaN (NA) 0.724

Key Differences: - Overall analysis may mask important subgroup differences - Stratified analysis reveals severity-specific patterns - P-values may differ when accounting for stratification - Treatment balance assessment within severity levels

Summary and Best Practices

When to Use Stratified Analysis

  1. Multi-center Studies: Always stratify by study site
  2. Sex Differences: Important for most clinical studies
  3. Age Groups: Especially relevant for studies spanning wide age ranges
  4. Disease Severity: Critical for understanding baseline risk
  5. Geographic Regions: For studies spanning different populations

Interpretation Guidelines

  1. Sample Sizes: Check adequate sample sizes within strata
  2. Missing Data: Consider missing data patterns within strata
  3. P-values: Interpret within-strata comparisons carefully
  4. Clinical Relevance: Focus on clinically meaningful differences
  5. Multiple Comparisons: Consider adjustment for multiple testing

Available Stratification Variables in This Dataset

Available Stratification Variables
Variable Description Use Case
site Study site (A, B, C, D) Multi-center trial balance
sex Participant sex (Male, Female) Sex-specific effects
age_group Age groups (18-44, 45-64, 65+) Age-related patterns
disease_severity Disease severity (Mild, Moderate, Severe) Baseline risk stratification
diabetes Diabetes status (No, Yes) Comorbidity analysis
hypertension Hypertension status (No, Yes) Cardiovascular risk factors

The stratified analysis capabilities of zztable1 provide powerful tools for understanding complex clinical trial data. By examining baseline characteristics within meaningful subgroups, researchers can better assess treatment allocation, identify potential confounders, and plan appropriate statistical analyses.