Same horizon for each study?

bcjaeger / melodem-apoe4-het

Workshop and collaborative manuscript

https://bcjaeger.github.io/melodem-apoe4-het/

Creative Commons Attribution Share Alike 4.0 International

3 stars 17 forks source link

Same horizon for each study? #9

Open bcjaeger opened 1 week ago

bcjaeger commented 1 week ago

Defining a treatment effect for the causal random survival forest requires specification of a horizon timepoint.

Should we use the same horizon for all datasets in our main analysis or use data-specific horizons?

magali17 commented 2 days ago

We could consider estimating treatment effects at multiple fixed time intervals so that all cohorts produce estimates for the same time intervals. Some cohorts will have estimates for longer periods of time.

bcjaeger commented 1 day ago

I like this idea a lot! I will write a PR for this.