I have a survival model that requires a flexible baseline hazard (like the Royston-Parmar, for example) and a shared frailty term. Presently, flexsurv cannot handle frailties and most packages that handle frailties cannot handle flexible baseline hazards. However rstanarm can handle both and I have fit a model like:
stan_surv(Surv(time, event) ~ covariate + (1 | shared.frailty), basehaz = 'ms')
where ms specifies an M-spline on the hazard and the shared.frailty variable identifies the clusters in my data. I would like to use this model as input for heemod. Is this possible or are flexsurv models the only way to compute a Healthy-Sick-Dead model in heemod?
I have a survival model that requires a flexible baseline hazard (like the Royston-Parmar, for example) and a shared frailty term. Presently, flexsurv cannot handle frailties and most packages that handle frailties cannot handle flexible baseline hazards. However rstanarm can handle both and I have fit a model like:
stan_surv(Surv(time, event) ~ covariate + (1 | shared.frailty), basehaz = 'ms')where
msspecifies an M-spline on the hazard and theshared.frailtyvariable identifies the clusters in my data. I would like to use this model as input for heemod. Is this possible or are flexsurv models the only way to compute a Healthy-Sick-Dead model in heemod?