Therefore, it will lead to bias if we use a marginal hazard ratio to estimate a hazard ratio.
This process was used as we were unaware of a formula that relates the marginal hazard ratio to the conditional hazard ratio for treatment, characteristics of the distribution of the covariates in the population, and the hazard ratios relating the covariates to the hazard of the occurrence of the outcome.
There were greater effects in no intervention control designs (marginal odds ratio = 9.80) than in manualized intervention (marginal odds ratio = 3.47, p =.03) or another active, nonmanualized intervention (marginal odds ratio = 3.38, p =.04) comparison designs in remission studies.
The statistical power to detect a non-null marginal hazards ratio using IPTW in an observational study was compared with the statistical power to detect a non-null hazards ratio in an RCT with the same number of observed events and similar characteristics.
The reason is that the marginal hazard ratio has finite limits as tends to and.
Comparisons were made under TRACER workspace evaluating the marginal likelihood ratio (marginal with respect to the tree prior) of pairs of models (Bayes Factor calculation).
To do so, we modified a previously described data-generating processes for generating data with a specified marginal odds ratio or risk difference [ 18, 19].
The regression coefficient βtreat was set equal to 0.164156, 0.3945684, and 0.721035, to induce a marginal hazard ratio of 1.1, 1.25, and 1.5, respectively.
With a marginal odds ratio of.996, the odds for a male birth in East Germany in years other than 1991 were thus higher than the marginal odds for a male birth in East Germany in 1991 (OR=.989=1.008*.996*.985).
However, we wanted to generate data in which there was a specified marginal hazard ratio (since propensity score methods and RCTs permit estimation of marginal, rather than conditional effects).
