Abstract
With the increased need to assess and manage risk in inpatient settings, the Short-Term Assessment of Risk and Treatability (START) was implemented on a civil psychiatric unit. The goal of the present study was to examine the tool's predictive validity when completed by clinical teams as part of routine practice. Data were collected for 34 patients hospitalized for a minimum of 30 days prior to and after a START evaluation. Several challenging behaviors, such as aggression towards others, self-harm, and substance abuse were assessed using the START Outcomes Scale (CitationNicholls et al., 2007). Results from multilevel logistic regression and Receiver Operating Characteristics analyses lend partial support for the predictive validity of the START. A limited set of START items combined was significantly better at predicting the challenging behaviors than the original total Strength and Vulnerability scales. Results are discussed in terms of the clinical use of risk assessment.
Acknowledgments
Dr. Crocker would like to acknowledge the salary award support from Canadian Institutes of Health Research (CIHR). This research was supported by a grant from Canadian Patient Safety Institute.
The authors wish to thank the staff from the Douglas Institute for their collaboration. The authors are very grateful to Marie-Eve Roy, for project coordination as well as Amanda Stevens, Vanessa Leo, Judith Sabetti and Todd Jenkins for data gathering and research-clinical liaison. The authors wish to thank the ongoing collaboration and support of the Douglas Institute Administrators, in particular, Mme Amparo Garcia, Dr. Mimi Israel, Mr. Michel Laverdure as well as those clinicians and coordinators who were the initial motivators for the project, Mr. Yves Carreau and Ms. Yvonne Hindle. Dr. Jean-Pierre Guay also provided invaluable comments in the final stretch of this paper.