stc-methodology

STC methodology transports trial results across populations by adjusting IPD-based estimates for covariate differences with aggregate data. It is needed when comparing treatments across trials with different covariate distributions in ITC analyses.

Choosing between anchored and unanchored STC depends on whether a common comparator exists across trials. Anchored STC uses linked treatment networks, while unanchored STC transports effects without a common comparator across populations.

Select effect modifiers based on clinical relevance and statistical significance, then center covariates at external population values to estimate treatment effects. Bayesian STC implementation requires clear model specification and appropriate covariate centering.

STC methodology aligns with NICE DSU guidance by requiring clear model specification, appropriate covariate centering, documentation of all assumptions, and sensitivity analyses. This ensures valid indirect comparisons between IPD and aggregate data.

Both frequentist and Bayesian STC implementations are supported for population-adjusted indirect comparisons. They include model diagnostics and reporting instructions to estimate external-population treatment effects from IPD and aggregate-data covariates.