Performs ANCOVA to estimate treatment effects adjusted for baseline and other covariates.
Usage
ancova_adjust_continuous(
data,
outcome_var,
trt_var,
baseline_var,
covariates = NULL,
ref_group = NULL,
conf_level = 0.95
)Arguments
- data
Data frame with outcome and covariates
- outcome_var
Character. Post-baseline outcome variable
- trt_var
Character. Treatment variable
- baseline_var
Character. Baseline value variable
- covariates
Character vector. Additional covariates to adjust for
- ref_group
Character. Reference group for contrast. If NULL, the first level of the treatment variable is used as reference.
- conf_level
Numeric. Confidence level (default: 0.95)
Value
List with:
treatment_effects: Data frame with adjusted differences vs reference
model: The fitted lm object
summary: Model summary
anova: ANOVA table
Details
Fits model: outcome ~ baseline + trt + covariates Treatment effects are extracted from model coefficients (trt - ref) and confidence intervals are computed with confint. Missing values are handled by lm() via listwise deletion. Specified variables must exist in the data.
Examples
# Simulated example
set.seed(123)
n <- 100
sim_data <- data.frame(
TRT01P = rep(c("Placebo", "Drug A"), each = n/2),
BASE = rnorm(n, 50, 10),
CHG = rnorm(n, 5, 8),
AGEGR1 = sample(c("<65", ">=65"), n, replace = TRUE),
SEX = sample(c("F", "M"), n, replace = TRUE)
)
result <- ancova_adjust_continuous(
data = sim_data,
outcome_var = "CHG",
trt_var = "TRT01P",
baseline_var = "BASE",
ref_group = "Placebo",
covariates = c("AGEGR1", "SEX")
)
print(result$treatment_effects)
#> Treatment estimate ci_lower ci_upper
#> 1 Drug A 2.312759 -0.8017789 5.427296