Standardization allows for efficient unbiased estimation in observational studies and in indirect treatment comparisons: A comprehensive simulation study
share
We evaluate how inverse probability of treatment weighting (IPTW) and standardization-based approaches compare for obtaining marginal estimates of the odds-ratio and the hazards ratio. Specifically, we consider how the two methods compare in two different scenarios: (1) in a single comparative study (either randomized or non-randomized), and (2) in an anchored indirect treatment comparison of randomized controlled trials. For each scenario, we conduct a simulation study to investigate the relative efficiency of each method and the potential for bias. We conclude that standardization-based procedures may be more efficient at providing unbiased estimates than those based on IPTW.
READ FULL TEXT