I am trying to estimate patients lost to follow-up in this dataset of 7 different studies with number of patients in the study available a inclusion(baseline) and a follow-up a 6 weeks, 3 months, 6 months and 12 months:
structure(list(...1 = c("Baseline", "6 Weeks", "3 Months", "6 Months",
"12 Months"), ...2 = c(28, NA, 26, NA, 26), ...3 = c(67, NA,
NA, NA, 62), ...4 = c(86, 77, 75, NA, 80), ...5 = c(146, 112,
121, NA, NA), ...6 = c(75, 69, 67, 65, 64), ...7 = c(95, NA,
NA, NA, 86), ...8 = c(59, NA, NA, NA, 51)), row.names = c(NA,
-5L), class = "data.frame")
for imputation i used the following code:
library(mice)
tempData <- mice(data,m=5,maxit=50,meth='pmm',seed=500)
Imputed_data=complete(tempData,5)
which only calculates 3 of 14 missing values and provides values higher than last observed value, which is not possible since patients cannot join the study again after dropping out.
Is it possible to use imputation for this, or should I just use LOCF?