1
$\begingroup$

I would like to use the causal impact algorithm, however, not in the context of marketing, but in medicine. The problem is that the intervention does not take place at the same time, but on an individual date for every participant. Is it possible to model the post and pre period individually?

I have round about 1000 participants. The intervention period is over one year and I could aggregate the intervention time at the month level: the year would not be concrete enough.

Improve this question
$\endgroup$
3
  • $\begingroup$ Are you open to using other approaches to this problem? There are some synthetic cohort packages that allow for multiple treated units that are treated at different times. A difference-in-differences model is another option. $\endgroup$
    – dimitriy
    Commented Jun 13, 2020 at 21:07
  • $\begingroup$ @DimitriyV.Masterov I'm encountering the same issue. Could you pls share the synthetic cohort package? and for difference-in-difference model in this setting, would the predictor be time and DID term be the interaction of time and intervention/dosage? Thanks! $\endgroup$
    – santoku
    Commented Oct 12, 2020 at 14:25
  • 1
    $\begingroup$ @santoku synth_runner in Stata, microsynth in R. DID has unit-specific treated dummy and post dummies, plus their interaction, which is the parameter of interest. $\endgroup$
    – dimitriy
    Commented Oct 15, 2020 at 4:44

1 Answer 1

Reset to default
1
$\begingroup$

No, the CausalImpact R package currently does not support that. One option is to artificially bring all interventions to the same time, i.e. make time stamps relative to the individual intervention time point. However, this only makes sense if there is no seasonality in the data (which might be true for medical data).

Improve this answer
$\endgroup$
1
  • $\begingroup$ any updates on the same $\endgroup$
    – shoonya
    Commented Oct 18, 2022 at 11:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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