Shortening the Screener and Opioid Assessment for Patients with Pain-Revised (SOAPP-R): A Proof-of-Principle Study for Customized Computer-Based Testing

Abstract

Background The Screener and Opioid Assessment for Patients with Pain-Revised (SOAPP-R) is a 24-item self-report instrument that was developed to aid providers in predicting aberrant medication-related behaviors among chronic pain patients. Although the SOAPP-R has garnered widespread use, certain patients may be dissuaded from taking it because of its length. Administrative barriers associated with lengthy questionnaires further limit its utility. Objective To investigate the extent to which two techniques for computer-based administration (curtailment and stochastic curtailment) reduce the average test length of the SOAPP-R without unduly affecting sensitivity and specificity. Design Retrospective study Setting Pain management centers Subjects Four hundred and twenty-eight chronic non-cancer pain patients Methods Subjects had taken the full-length SOAPP-R and been classified by the Aberrant Drug Behavior Index (ADBI) as having engaged or not engaged in aberrant medication-related behavior. Curtailment and stochastic curtailment were applied to the data in post-hoc simulation. Sensitivity and specificity with respect to the ADBI, as well as average test length, were computed for the full-length test, curtailment, and stochastic curtailment. Results The full-length SOAPP-R exhibited a sensitivity of 0.745 and a specificity of 0.671 for predicting the ADBI. Curtailment reduced the average test length by 26% while exhibiting the same sensitivity and specificity as the full-length test. Stochastic curtailment reduced the average test length by as much as 65% while always exhibiting sensitivity and specificity for the ADBI within 0.035 of those of the full-length test. Conclusions Curtailment and stochastic curtailment have potential to improve the SOAPP-R’s efficiency in computer-based administrations.

An investigation of completion times on the Screener and Opioid Assessment for Patients with Pain – revised (SOAPP-R)

“…Recent research on the SOAPP-R has indicated an interest in and the feasibility of employing a computerized version of the screener 42,67. The regular use of a computer-based SOAPP-R would facilitate the collection of further response-time data, which would open the door to new methods of determining risk of ADB.…”

Computer‐based testing and the 12‐item Screener and Opioid Assessment for Patients with Pain‐Revised: A combined approach to improving efficiency

“…In stochastic curtailment, the screener is stopped at any stage of testing where curtailment would stop it, as well as at any stage of testing where the final result of the SOAPP‐R (“positive” or “negative”) is predicted with a high level of probability (such as 95% or 99%). In previous research conducted retrospectively, curtailment lowered the average test length of the SOAPP‐R by 26% without any change in sensitivity or specificity; stochastic curtailment lowered the average test length by up to 65% without changing the sensitivity or specificity by more than 3.5% (Finkelman et al., ).…”

“…Existing validated questionnaires are available to assist clinicians in screening patients for potential aberrant medication‐related behaviors, including the Opioid Risk Tool (Webster & Webster, ) and the Screener and Opioid Assessment for Patients with Pain‐Revised (SOAPP‐R; Butler, Budman, Fernandez, Fanciullo, & Jamison, ). The latter questionnaire has undergone validity testing (Butler, Fernandez, Benoit, Budman, & Jamison, ) and has been extensively researched in subsequent studies (Finkelman et al., , ; Finkelman, Jamison, et al., ; Finkelman, Smits, et al., ; Jones et al., ; Weiner, Horton, Green, & Butler, , ). The full‐length version of the SOAPP‐R consists of 24 self‐report items (whose item scores are summed to obtain a total score), with a single cutoff point to categorize patients as either low risk or high risk for aberrant medication‐related behaviors.…”

“…In light of the above concerns, recent research has investigated whether it is possible to shorten the SOAPP‐R without sacrificing its ability to predict aberrant medication‐related behavior (Black, McCaffrey, Villapiano, Jamison, & Butler, ; Finkelman et al., ; Finkelman, Smits, et al., ). One approach that has been suggested (Finkelman et al., ) is to administer the screener by computer, track a given respondent's score as he/she proceeds through the test, and stop the assessment once an internal algorithm determines that further items are not necessary. Two types of internal algorithms to accomplish this task were examined: curtailment and stochastic curtailment .…”

“…Previous research on shortening the SOAPP‐R has investigated the computer‐based methods (curtailment/stochastic curtailment) and the short form separately from each other (Finkelman et al., ; Finkelman, Jamison, et al., ; Finkelman, Smits, et al., ). However, given that both approaches have shown promise in reducing test lengths, applying only one approach at a time may not take full advantage of their potential; it is possible that test efficiency would be enhanced by applying them jointly.…”