Health related data with discrete ordered dependent variables are often characterised by excessive 'zero' observations that may relate to two distinct data generating processes. For example, there may be a large proportion of 'no visit' in modeling discrete ordered levels of health service uses...