Identifying Predictors of Violence in Hospitalized Patients

Individuals with SMI and co-occurring substance abuse and/or dependence (i.e. dual diagnosis) are at heightened risk for engaging in self–directed and other-directed violent behavior. The causes and predictors of violence among people with dual diagnosis are not well understood, in part, due to the limitations of clinical assessment strategies and research methodologies. Identification of time-sensitive predictors of violent ideation in people with dual diagnosis may facilitate the development of time-sensitive interventions that may reduce the incidents of violent behavior, or prevent its occurrence altogether.

Contemporary mobile technologies offer us novel solutions to the limitations of retrospective clinic or laboratory-based interview approaches. In the current study we use mobile technologies that combine unobtrusive behavioral sensing (i.e. using smartphone embedded-accelerometers, microphone, light sensors) and Ecological Momentary Assessment (i.e. a mobile self-report data collection technique) to capture internal (i.e., thoughts, feelings, bodily sensations) and behavioral (i.e., movement, engagement in conversation, location) factors that may contribute to the emergence of violent ideation in hospitalized patients with dual diagnosis.

Funded by: Center for Technology and Behavioral Health Pilot Program, Dartmouth Psychiatric Research Center