compute_weights(): Entropy balancing / exponential
tilting for MAIC weight estimation, with support for mean-only and
mean+SD matching. Returns ESS, convergence diagnostics, and pre/post
SMDs.
dr_maic(): Doubly robust MAIC estimator combining
IPW (standard MAIC) with outcome regression (STC / g-computation).
Supports binary, continuous, and time-to-event outcomes. Effect
measures: RD, RR, OR, HR, MD.
maic_diagnostics(): Love plot, weight distribution
plot, covariate balance table, and ESS summary. All plots are ggplot2
objects.
check_assumptions(): Structured assumption checklist
aligned with NICE DSU TSD 18 and Cochrane Handbook Chapter 23. Checks
ESS adequacy, covariate balance, optimiser convergence, and DR
augmentation term.
bootstrap_ci(): Non-parametric bootstrap confidence
intervals (BCa, percentile, normal) for all three estimators (MAIC, STC,
DR-MAIC). Full bootstrap distribution plot included.
sensitivity_analysis(): E-value computation
(VanderWeele & Ding, 2017), weight trimming sensitivity analysis,
and leave-one-variable-out (LOVO) analysis. All results include ggplot2
visualisations.
nice_report(): Structured submission-ready report
covering population characteristics, weight estimation, covariate
balance (NICE TSD 18 format), treatment effect estimates, uncertainty
quantification, sensitivity analysis, assumptions/limitations, and a
citable methods paragraph.
nsclc_ipd: Simulated individual patient data (n = 200)
from a hypothetical single-arm immunotherapy trial in advanced
NSCLC.nsclc_agd: Simulated aggregate data from a hypothetical
comparator trial.