Evaluates the transitivity assumption by comparing population characteristics across studies in a network. Transitivity requires that studies are similar enough that indirect comparisons are valid.
Usage
assess_transitivity(
study_characteristics,
char_vars,
treatment_var = "treatment",
continuous_vars = NULL,
threshold_smd = 0.1
)Arguments
- study_characteristics
Data frame with study-level characteristics. Must include 'study_id' and 'treatment' columns, plus characteristics to compare.
- char_vars
Character vector. Names of characteristic variables to assess.
- treatment_var
Character. Name of treatment variable. Default: "treatment"
- continuous_vars
Character vector. Which char_vars are continuous (vs categorical)
- threshold_smd
Numeric. SMD threshold for imbalance. Default: 0.1
Value
A list with components:
- summaries
Summary statistics by treatment
- comparison_tables
Pairwise comparison tables
- imbalance_scores
SMDs for continuous variables
- overall_assessment
Text summary of concerns
- n_treatments
Number of unique treatments
- n_characteristics
Number of characteristics assessed
Details
Transitivity is a fundamental assumption in network meta-analysis. When treatments A and C have not been directly compared in a randomized trial, we may infer their relative effect through a common comparator B (A vs B and B vs C). This indirect comparison is valid only if the studies comparing A vs B and B vs C are sufficiently similar in terms of population characteristics, study design, and clinical setting. Violations of transitivity can bias indirect comparisons and undermine the credibility of network meta-analysis results.
Examples
# Assess transitivity across treatment comparisons
chars <- data.frame(
study_id = c("S1", "S1", "S2", "S2", "S3", "S3"),
treatment = c("A", "B", "B", "C", "A", "C"),
mean_age = c(55, 55, 58, 58, 52, 52),
pct_male = c(60, 60, 65, 65, 55, 55)
)
result <- assess_transitivity(
study_characteristics = chars,
char_vars = c("mean_age", "pct_male"),
continuous_vars = c("mean_age", "pct_male")
)
result$overall_assessment
#> [1] "Transitivity assumption appears reasonable"