Multiple logistic regression odds ratio on multiply imputed data in R using mice

[mona935]

I am running a hierarchical logistic regression analysis using multiply imputed data in R (using the mice and miceafter packages). I am unable to get the odds ratio and 95% CI per variable adjusted for all other predictor variables.

Here is what I have done so far:

Model1 <- with(imp, glm(complaints ~ Sex + Age + Education, family = binomial(link = 'logit')))

pool.model1 <- summary(pool(Model1), conf.int = TRUE, exponentiate = TRUE)

Now, I manage to get odds ratios, but only separately per combination of predictor and dependent variable:

long <- mice::complete(imp, "long", include = TRUE)

imp_data <- df2milist(long, impvar=".imp")

ra <- with(imp_data, expr=odds_ratio(complaints ~ Sex))

OR <- pool_odds_ratio(ra)

If I add another predictor I get the following error:

Error in odds_ratio(complaints ~ Sex + Age) : Include single independent variable only or use function psfmi_lr
The problem is that I of course want the odds ratios (and 95% CI) while considering the influence of the other variables.

Is there a way to get the pooled odds ratio and corresponding 95% CI per predictor variable in a model of multiple predictors?