r - Pooling sandwich variance estimator over multiply imputed datasets -

i running poisson regression on multiply imputed data predict common binary outcome. after running mice, have obtained stacked data frame comprising raw data , 5 imputed datasets. here toy example:

df <- mice::nhanes imp <- mice(df) #impute data com <- complete(imp, "long", true) #creates data frame 

i want to:

1. run regression on each imputed dataset
2. calculate robust standard errors using sandwich variance estimator
3. combine / pool results of both analyses

i can run regression on mids object using with , pool commands:

fit.pois.mids <- with(imp, glm(hyp ~ age + bmi + chl, family = poisson)) summary(pool(fit.pois.mids)) 

i can run regression on each of imputed datasets before combining them:

imp.df <- split(com, com$.imp); names(imp.df) <- c("raw", "imp1", "imp2", "imp3", "imp4", "imp5") #creates list of data frames representing each imputed dataset  fit.pois <- lapply(imp.df, function(x) {   fit <- glm(hyp ~ age + bmi + chl, data = x, family = poisson)   fit })  summary(micombine(fit.pois)) 

similarly, can calculate standard errors each imputed dataset:

sand <- lapply(fit.pois, function(x) {   se <- coeftest(x, vcov = sandwich)   se }) 

unfortunately, micombine not seem return p-values. post suggests using zelig, matter, may use mice. further not appear possible combine estimates of standard errors:

summary(micombine(sand.df)) error in usemethod("vcov") :    no applicable method 'vcov' applied object of class "coeftest" 

for sake of simplicity, seems mice better option pooling results of regression; however, wondering how go updating (i.e., pooling , combining) standard errors. ways addressed?