3
$\begingroup$

I am running multiple imputation for a set of variables including clinical data. I am wondering if I can use (or should use) outcome variable (follow-up is 99%) to predict missing clinical data. There is about 12 % of non complete cases due to mostly one variable. I intend to use Amelia package (R)?

Some extra information included from comment:

Providing that there are same baseline variables missing do you exclude lost to follow-up patients and then run multiple imputation method or do include patients lost to follow-up and than run multiple imputation method. What is the correct approach?

I would be most grateful for answer.

$\endgroup$
1
$\begingroup$

In principle it is a common practice to use your dependent/outcome variable to impute independent ones. Imputation should keep the data structure (in terms of correlations between variables). Therefore, using the dependent variable to impute others should not be a problem. You might want to later exclude observations with missing values in your dependent variable from analysis.

For more information you can check the article from Paul T. Von Hippel (2007) "REGRESSION WITH MISSING YS: AN IMPROVED STRATEGY FOR ANALYZING MULTIPLY IMPUTED DATA." http://onlinelibrary.wiley.com/doi/10.1111/j.1467-9531.2007.00180.x/abstract

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.