Concordance of clinical estimation of suicidal intent

Objective: The aim of this paper is to assess the level of agreement between clinical estimates of suicidal intent based entirely on information recorded in the Accident and Emergency acute assessment and Beck's Suicide Intent Scale (SIS) scores. Method: As part of the WHO/EURO Multicentre Study of Parasuicide, cases of deliberate self-harm (DSH) in Cork city hospitals were monitored. Over the period 1995-1997, the information recorded in the Accident and Emergency acute assessment was examined by a psychiatrist and, if possible on the available evidence, clinical estimates were made at three levels of suicidal intent: minimal, moderate and definite. Seventy-nine of these cases had fully completed Beck's SIS. Statistical comparison was made between the results of the SIS and the clinical estimate of suicide intent. Results: The agreement (Kappa = 0.146, p = 0.046) and concordance (Lin's concordance coefficient = 0.330, p = 0.001) between the two ratings are statistically significant but both are low. The overlap between those identified as high or low intent is low. Conclusions: These findings indicate low agreement between a clinician's rating of suicide intent based on clinical records and Beck's SIS. This is especially relevant given the increasing reliance on psychometric instruments in assessment in psychiatry. However, further investigation is necessary to clarify which is the more valid method.