Authors: Ryan A. Black, PhD, Stacey A. McCaffrey, PhD, Albert J. Villapiano, EdD, Robert N. Jamison, PhD, Stephen F. Butler, PhD
Objective: Although the Screener and Opioid Assessment for Patients with Pain–Revised (SOAPP-R) has become a widely used screener for aberrant opioid-related behavior in adults, the length of the instrument may limit its utility. The purpose of the current study was to develop a short form of the SOAPP-R by retaining as few items as possible while maximizing predictive accuracy.
Methods: Participants (N = 555), recruited from pain clinics, completed the 24-item SOAPP-R and participated in a five-month follow-up visit to evaluate aberrant drug-related behaviors. Opioid aberrant-related behavior was determined through self-report, physician report, and urine toxicology screen. The optimal subset of SOAPP-R items to predict aberrant opioid-related behavior was identified empirically by employing the LASSO selection method and the leave-one-out cross-validation (LOOCV) method offered in the GLMSELECT procedure in SAS 9.4 in conjunction with content expertise.
Results: Eight items were identified before the selection method stopped. The receiver operating characteristic curve generated from the predicted probabilities from the model produced an area under the curve (AUC) value greater than the AUC value produced by the 24-item SOAPP-R total score and yielded a sensitivity of 0.74 and a specificity of 0.66.
Click here to view the full article.