nif:isString
|
-
We combined two baseline treatment status groups, groups 3 and 4, for one propensity score model, because we hypothesized the propensity of enrollment in the CCP would be influenced by the same potential confounders for these two groups; when we estimated the propensity for enrollment with individual models for groups 3 and 4, the effect estimates from the pooled model did not differ from the effect estimates from the individual models.
|