Dear Clare,
The GEE results are valid only when there is a small and asymptotically diminishing fraction of missing data
or when the data are missing completely at random (MCAR). Correctly specification of
the correlation struture is required to obtain consistent result for the parameter estimates and their variances
when the patttern of the missing data at a particular time point depends on the previous outcomes.
Therefore, the robustness to the choice of the correlation structure does not remain valid in the case of nonrandom missing data.
If the fraction of missingness is quite small , GEE models will provide valide estimates.
Or if your data are normal why not go for Mixed model which assumes MAR?
Regards,
Divine
> Date: Tue, 3 Nov 2009 12:29:23 +0000
> From: [log in to unmask]
> Subject: Question :GEE with missing data
> To: [log in to unmask]
>
> Dear Allstat,
>
>
>
> I have a longitudinal dataset from a 2 treatment group clinical trial with
> baseline and 4 follow up timepoints. I may have missing baseline covariates
> and follow up data.
>
>
>
> As I am interested in the average effect of the treatment groups I wish to
> analyse using a GEE model. As the missing data is likely to be Missing At
> Random (MAR) I have read that GEE models are likely to provide biased
> estimates.
>
>
>
> My thoughts were to multiply impute the missing data and then carry out a
> GEE model.
>
>
>
> I would be interested to hear your views on whether this is the best
> approach (if so how best to impute the missing data?) , or are there
> alternative approaches.
>
>
>
>
>
> Best Wishes
>
>
>
> Clare Rutterford
>
>
_________________________________________________________________
Windows Live: Make it easier for your friends to see what you’re up to on Facebook.
http://www.microsoft.com/middleeast/windows/windowslive/see-it-in-action/social-network-basics.aspx?ocid=PID23461::T:WLMTAGL:ON:WL:en-xm:SI_SB_2:092009
|