RMS Semiparametric Ordinal Longitudinal Model

Yes, averaging out is probably the better term. Do you know of any study that elicited utilities to the NIAID scale (through discussions with patients / doctors) ? We’re currently not investigating this.

FWIW: P(Benefit) in ACTT-2 was also higher when analysing the time to recovery compared to the Markov model, which I also did not expect. Working more extensive simulations and resampling studies currently, which to you disappointment, will be frequentist…