I mean, how do I put the dropped out participants' outcome (Missing value) in my analysis? Although I have already known what's ITTA meaning, I still don't know how to do it. Which methods should I use to fill missing values? Which statistical software I can use to perform ITTA & how etc.
If you can give me any explanations or advice, thanks a lot!
Intention-to-treat is less concerned with how to manage missing data than what you do with the data you have. Imagine participants are randomised to medication A or B, and none of them drop out but some take the wrong medication. For ITTA, outcomes from all participants who were randomised to medication A are included in the "group A" results, even if some of those participants actually took medication B.
What to do with missing values is a different question. My understanding is that whether you exclude the missing values or use some kind of modelling to estimate what they would have been, you can still call it an ITTA as long as you have analysed all participants according to their initial grouping.
In fact, the withdrawal of the participants considered as death. The rationale behind this consideration is that the withdrawal (maybe because of side effects or lack of recovery) is a failure for that treatment.
Hi Babak Jamshidi
There are multiple imputation and model-based approaches, such as mixed models and weighted generalized estimating equations (GEEs) for repeatedly measured outcomes, based on all observed data can be valid and unbiased methods for missing not at random data, as long as the models are specified correctly. We should consider using one the following approaches (in R or SAS) that are valid for missing at random data and thought to be more robust than sensitive analysis. Those include multiple imputation, mixed models, inverse probability weighted GEEs, and Bayesian analysis. I hope this clarify your query. Good luck!
thanks a lot, I will try these methods. (∩_∩)O Babak Jamshidi Giovane R Sousa
& Thanks for your reply Claire Styles , so, may I understand that ITTA seems more like a conception, not a method? & if ITTA only used in the RCT study? I have done a community-based intervention, but the participants were not randomized, I'm puzzled should I use the ITTA analysis?
ITTA is only for prospective randomised studies. I'm not sure if there is an equivalent for cohort studies.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software