Joint models for continuous and discrete longitudinal data
Measurements of both continuous and categorical outcomes appear in many statistical problems. One such example is a toxicology experiment, where a set of outcomes is recorded to study the toxicity of the substance of interest. For example, both the malformation status of a live fetus (typically recorded as a binary or ordinal outcome) and low birth weight (measured on a continuous scale) are important variables in the context of teratogenicity. Also in longitudinal studies, it is common to collect several outcomes from a diﬀerent type. Perhaps the most common situation is that of a continuous, often normally distributed, and a binary or ordinal outcome.