Paper
Thursday, 20 July 2006
This presentation is part of : Developing Surveys and Scales
Predicting Adolescent Suicidal Ideation: The Role of Data Reduction
Thomas Hardie, EdD, RN, CS, NP, School of Nursing, University of Delaware, Newark, DE, USA, Kevin G. Lynch, PhD, Department of Psychiatry, University of Pennslyvania, Philadelphia, PA, USA, and Howard Moss, MD, National Institute on Alcohol Abuse, National Institutes of Health, Bethesda, MD, USA.
Learning Objective #1: Select data reduction strategies to improve predictive modeling.
Learning Objective #2: Identify risk factors associated with suicidal ideation in adolescents.
Adolescent suicidal ideation is strongly linked to suicidal behaviors. Suicide is third among mortality risks for those between 10 and 24 years of age with 4,243 (11.7%) deaths each year. The goal of this research project is the development of indicators to predict suicidal ideation in a large community (ADD Health) sample of adolescents (N= 6360). Specifically, concurrent depressive symptoms, externalized problem behaviors, gender, age and alcohol exposure have been used to construct a model to predict the presence of suicidal ideation. A second model explores problem behaviors associated with alcohol use as well as the predictors listed above in those adolescents exposed to alcohol during the last year. Data reduction strategies offer methods to test complex multivariate relationship and enhance our capacity predictors that can translate in improved assessment in the clinical area. Despite some mathematical complexity their ability to evaluate large blocks of predictors or symptoms makes these procedures important to nursing research. We will outline methods for coping with common challenges in managing missing and nonlinear data, using factor analytic models for the reduction in the number of variables to be analyzed and modeling options in the search for a multivariate predictive model for suicidal ideation. The results of a series of exploratory and confirmatory factor analyses and two backward logistic regressions suggest subsets (factors) of depressive items providing both increased risk and protection, gender, selected externalizing problem behavior (factors) and drinking alcohol increase risk of suicidal ideation (odds ratios between 1.3 and 2.9). In drinkers, a similar pattern was found, among the results alcohol related problems behaviors were not in the final model.

See more of Developing Surveys and Scales
See more of The 17th International Nursing Research Congress Focusing on Evidence-Based Practice (19-22 July 2006)