Multivariate analysis ('forensiometrics') - a new tool in forensic medicine. Differentiation between firearm-related homicides and suicides

A forensiometric model to discriminate between homicidal and suicidal firearm-fatalities is presented. It is based on a survey of all 45 homicidal and 251 suicidal firearm-fatalities that were examined at the Department of Forensic Medicine in Stockholm, Sweden during the period 1983-92. The model used 15 variables to describe 19.8% of the variation in the data and its goodness of prediction, Q(2), was 0.742. The variables ranked in falling order of covariation with homicide (i.e. correlation with or importance in the prediction of homicide) were: presence of other injuries than firearm-wounds, entrance wound in the front part of the chest except precordium, bullet path through clothes, entrance wound in upper extremity, female victim, entrance wound in abdomen, entrance wound in head except temples, central forehead and mouth, and entrance wound in back. The variables most correlated to suicide were: used firearm found close to victim's body, reported suicidal ideation, victim's age, presence of contact wound(s), male victim, presence of farewell letter and entrance wound in mouth. The model was thereafter validated on a test-set of 18 homicides and 84 suicides during 1993-95. All suicides and 16 homicides were classified by the model in agreement with the previous police and forensic medical examinations. Thus, the model's sensitivity to classify homicides is estimated to be 89% and its specificity 100%. (C) 1999 Elseviar Science Ireland Ltd. All rights reserved.