Аннотация:
Background. Effective prediction of the likelihood of relapse of offenses of mentally ill persons and their prevention are important areas of activity to ensure the safety of citizens, urgent tasks of psychiatric service units.
Aim: assessment of the psychometric properties of the predictive model for the risk of repeated illegal behavior in persons with severe mental disorders, its reliability and predictive validity.
Methods. The research sample (n=160) was formed from two groups of respondents with severe mental disorders: the main group – persons with repeated illegal behavior; comparison group – with law-abiding behavior. Quantitative data were described using the median (Me) and the lower and upper quartiles (Q1–Q3). Comparison of two groups in terms of a quantitative indicator, the distribution of which differed from the normal one, was performed using the Mann-Whitney U-test. Percentages were compared using Pearson’s chi-square test. To evaluate the diagnostic model in predicting a certain outcome, the method of analyzing ROC curves was used. The critical level of significance when testing statistical hypotheses was taken equal to p<0.05.
Results. The results of the approbation testify to the information content of the predictive model for determining the risk of repeated socially dangerous behavior. The area under the ROC curve (AUC) was 0.989 ± 0.009 with 95% CI: 0.972 – 1.000 The sensitivity and specificity of the model at a cutoff of 0.74 were 98.8% and 81.2%, respectively.
Conclusion. According to the authors, the use of a prognostic model in combination with a clinical examination, dynamic observation and experimental psychological research will increase the validity and evidence of the assessment of the risk of public danger of persons with mental disorders.
Ключевые слова:
mental disorders, illegal behavior, repeated socially dangerous actions, prevention, risk assessment, prognostic model.