| Anxiety disorders are among the most prevalent and impactful mental health diagnoses. In an effort to lessen functional impairment, researchers have been studying how to reduce the impact of anxiety disorders by utilizing evidence based treatments such as cognitive-behavioral therapy (CBT). Due to the already high success rate of CBT for anxiety, identifying non-responders by searching for variables that predict treatment outcome a priori has been of recent focus as this identification will facilitate patient-treatment matching and help allocate resources more efficiently. In an effort to expand upon prediction research, this study seeks to create and validate a brief and easy to administer assessment tool, the Prognostic Indicator Rating Form (PIRF), that can be used to predict treatment outcome. Based on the extant research, the PIRF includes nine patient characteristics (i.e., illness severity, illness duration, previous treatment amount, social support perception, Axis I & II comorbidity, SES, treatment expectations, and treatment motivation) rated on a five-point scale. A total of 241 adult patients were utilized in the current study. The Clinical Global Impression-Severity (CGI-S) and the Clinical Global Impression-Improvement (CGI-I) scales were used as outcome measures. These outcome measures were then utilized to classify participants as more or less improved (i.e., NCGI-I and NCGI-S) to more evenly distribute groups, as most patients either improved or remained at baseline after CBT. A reliability analysis revealed that the PIRF possesses adequate, but low internal consistency (alpha = 0.65). As such, a principal components analysis was conducted and revealed two components: a Social Factors Component and a Diagnostic Characteristics Component. High test-retest existed between initial assessment and the first treatment session (r = 0.92, (p < .001). Based on the analyses performed (i.e., correlation, t-test, ROC curve, and regression), perception of social support and the Social Factors Component consistently predicted treatment response using the outcome measures, possibly indicating that characteristics not directly related to the patient's diagnosis best predict outcome, which conflicts with past research. When utilizing the PIRF total score, the CGI-S, NCGI-S, and NCGI-I best predict CBT outcome. Clinical and methodological implications as well as areas for future research are discussed. |
