-
Notifications
You must be signed in to change notification settings - Fork 2
This issue was moved to a discussion.
You can continue the conversation there. Go to discussion →
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Dichotomous data? #87
Comments
For the time being, I restricted |
There are two main challenges for dichotomous endpoints:
For (2), the worst-case scenario is that non-continuous non-identity-link conditional means priors are not possible in
``` r
suppressPackageStartupMessages({
library(brms)
library(coda)
library(emmeans)
library(mmrm)
library(posterior)
library(tidyverse)
library(zoo)
})
emm_options(sep = "|")
FEV data from the mmrm package, using LOCF and then LOCF reversedto impute responses.data(fev_data, package = "mmrm") lm + emmeansmodel <- glm( custom marginal means from lm draws using a custom mappingfrom model parameters to marginal means.proportional_factors <- model.matrix( Compare resultssummary <- summary_custom %>%
|
We may want to focus on "unconditional treatment effects" in the sense of the recently updated FDA guideline on Adjusting for Covariates in Randomized Clinical Trials (see section III C. on "Nonlinear Models"). |
If I understand correctly, I think we are mostly using unconditional treatment effects, i.e. regress on |
I wonder, if we allow a non-identity link and dichotomous data, would we have to change the package name to something like |
And would it make sense anymore to regress on baseline? |
If we use a link function for dichotomous data, we need a distributional family for dichotomous data, such as I am favor of excluding non-Gaussian families from the package, or at least waiting until we know exactly which family or families to bring within scope. Transferring this issue thread to a discussion thread. |
This issue was moved to a discussion.
You can continue the conversation there. Go to discussion →
Should we extend
brms.mmrm
to support dichotomous data? I thinkbrms
straightforwardly accepts nonlinear link functions, but I am wondering how this would affect #53 and #40.The text was updated successfully, but these errors were encountered: