A PRELIMINARY STATISTICAL-MODEL FOR IDENTIFYING REPEATERS OF PARASUICIDE

This paper presents a statistical model constructed using logistic reg ression to identify those at high-risk of repeating parasuicide. The s ubjects in the study are Cork city residents who exhibited parasuicida l behaviour between 1 January and 30 June 1995. Repetition of the beha viour within six months of the index episode distinguishes repeaters f rom nonrepeaters. The model was designed so that it could be used by n on-clinicians and hence does not require information relating to psych iatric diagnosis or use of psychiatric services. The proportion of sub jects correctly classified remained stable across a range of cut-point probabilities (mean = 86%, range: 83.9-87.5%). Using a cut-point of 0 .2, 96% of repeaters and 81% of non-repeaters were correctly classifie d. Using 0.45 led to the correct identification of 81% of repeaters an d 90% of non-repeaters. If these high levels of sensitivity and specif icity are maintained in validation tests on future cohorts in Cork cit y then the model could form the basis of an intervention programme des igned to prevent the repetition of parasuicide.