Physiological Reactivity and Post-traumatic Stress Disorder Diagnosis When Some Data are Missing

• Activity Number: 409
• Type: Contributed
• Date/Time: Thursday, August 12, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Section on Health Policy Statistics
• Author(s): Haekyung Jeon-Slaughter*+ and Phebe Tucker and Carol S. North and Betty Pfefferbaum
• Companies: University of Oklahoma Health Sciences Center and University of Oklahoma Health Sciences Center and Washington University School of Medicine and University of Oklahoma of Health Sciences Center
• Keywords: incomplete data ; missing values ; PTSD ; interview data ; physiological reactivity ; imputation

Abstract

This paper examines the effects of incomplete participation in research on the relationship between chronic post-traumatic stress disorder (PTSD) and physiological reactivity. Seven years after the Oklahoma City bombing, survivors were compared to community controls. All participants completed a structured diagnostic assessment, but 30% declined the physiological assessment (pulse and blood pressure measurements in response to a trauma cue interview). Missing values in physiological assessment may be random and unrelated to PTSD (ignorable), or they may not be random because participants with PTSD may be uncomfortable with the assessment (nonignorable results that may relate to study variables). Thus, excluding incomplete datasets may bias the results. We utilize graphic, maximum likelihood estimation, and simple mean substitute analyses to determine if missing values in this study are ignorable or nonignorable. Logistic regression and discriminant analyses are employed to investigate how well physiological reactivity predicts the research-generated PTSD diagnosis.