Among those who present to the emergency department for self-harm, many will repeat. Self-harm repetition is an outcome of interest in both observational and intervention studies. However, few such studies analyse the number of repeat self-harm presentations. Here, hurdle models are introduced as a potentially useful statistical method for these analyses. Emergency department data from the Province of Ontario, Canada, are used to illustrate an example of implementing hurdle models and interpreting their results.
View full text
Research and statistical methods