2

I have a follow-up question on the question of this thread:

Manipulating data for propensity score matching following multiple imputation with mice package

In aforementioned topic we can see the use of multiple imputation and the concomitant regression of the outcome according to the Rubin's Rule (we used multiple imputation to take care of missing values in the dataset).

Example used in the aformentioned topic (usage of MICE in R):

> imp.data <- complete(imp, "long") fit.list <- vector("list",
> nlevels(imp.data$.imp)) for (i in seq_len(nlevels(imp.data2$.imp))) { 
> m.out <- matchit(t ~ v1 + v2 + v3, data = imp.data[imp.data$.imp ==
> i,])   fit.list[[i]] <- glm(y ~ t, data = match.data(m.out)) }
> fit.list.mira <- as.mira(fit.list) #combines into mira object for
> pool() summary(pool(fit.list.mira))

However my question is similar to this one, with a minor chance; can not find it in the literature so far! In this example they perform a glm (generalized linear model). I want to perform a Cox Proportional Hazards Regression across matched datasets acquired from the multiple imputation (so not the whole dataset, but the matched after using MatchIt) and pool them according to the Rubin's Rules, but I can not find an example on how to do it in R.

Does somebody know the literature on this or the package needed in R? I would like to read it!

Kweetvannix
  • 91
  • 3
  • I don't understand. What did you use MI for and why are there multiple datasets? Questions about R are off topic, so it's best to organize your questions into concepts here and turn to stackoverflow.com for implementation. But still better to clarify what the goal and challenges are first. – AdamO May 25 '18 at 12:17
  • Edited! Might post the question at stackoverflow too, if this is not the right forum for these kind of questions! – Kweetvannix May 25 '18 at 12:22
  • I posted [here](https://stats.stackexchange.com/questions/160363/applying-rubins-rule-for-combining-multiply-imputed-datasets/274873#274873) about how Rubin's Rules are derived and ways of multiply imputed standard errors in R. Questions strictly about R are considered off-topic, but most people don't realize RRs are easy enough to just calculate by hand. – AdamO Jun 01 '18 at 20:38
  • does it answer your question? – AdamO Jun 04 '18 at 18:15

0 Answers0