Joint modelling of longitudinal outcome and interval‐censored competing risk dropout in a schizophrenia clinical trial