A study of the habituation and acute risk measure as a predictor of treatment decision making and readmission status in an emergency department setting
Williams, Dara Ann
MetadataShow full item record
Internationally, suicide has become an increasingly common phenomenon occurring with greater frequency over the past 50 years. This trend is especially troublesome given that the impacts of suicide are far reaching. For example, the healthcare system has been negatively impacted by this trend, with patient suicide becoming a frequent “occupational hazard” for mental health practitioners and other medical providers charged with providing care for at risk individuals. In an effort to address this issue, clinicians and researchers alike have attempted to identify and implement assessment techniques aimed at the early detection of individuals at increased risk of suicide. Improved methods of risk assessment will help to prevent unnecessary deaths, ensure that others will not lose their loved one to suicide, and that mental health clinicians will not suffer the personal and professional consequences of patient suicide. This study explores the predictive validity of the Habituation and Acute Risk Measure (HARM), a self-report suicide risk assessment questionnaire, in understanding treatment disposition outcomes (e.g. inpatient and outpatient) and readmission status for patients evaluated within an Emergency Department (ED) setting. A measure with encouraging theoretical underpinnings and psychometrics, the HARM was initially administered to patients presenting in one of four ED’s found within a large non-profit healthcare system in the southeastern United States. Data were collected by administrators as part of a system-wide quality improvement initiative mandated by Behavioral Health leadership. Archival data were compiled into a secure database and analyzed at a later date for purposes of this study. Logistic regression and chi square analyses were used to determine the variables most predictive of treatment disposition and readmission status. Correlations amongst all HARM subscales were also explored. Between-group comparisons were conducted via independent t-test and ANOVA analyses to determine whether significant differences existed between demographic groups on any of the HARM subscales. ROC curves, as well as measures of sensitivity and specificity were also utilized to ascertain the predictive properties of the final treatment disposition and readmission models. Several variables were found to significantly predict outcome and readmission including: Agitated Depression, Suicidal Intent, age, ethnicity/race, education level, and relationship status.