
212 Discrete Event Simulation for Health Technology Assessment
In this representation, j indexes the outcome category, and ranges from one
to c−1. The regression allows a different intercept for each category of the
outcome, allowing the probabilities to vary for each category, but assumes
that the effect of each coefcient is the same when assessing the log odds of
category j versus c for any j=1, …, c−1. This is sometimes referred to as a
proportional odds equation, since the log-odds across different values of
a predictor (e.g., age) for different categories of the outcome are parallel on a
plot, with gap sizes determined by the difference ...