Abstract
Introduction
The Patient Acceptable Symptom State (PASS) is a widely used outcome threshold for estimating patient satisfaction with patient-reported outcomes (PROs). The aim of this study is to assess the accuracy of threshold-based PASS and to identify risk factors for its misalignment with a patient’s response to the PASS question.
Methods
Patients were included within the study if they had preoperative and 2-year follow-up after hip arthroscopy. PROs evaluated included the modified Harris Hip Score (mHHS), Non-Arthritic Hip Score (NAHS), International Hip Outcome Tool (iHOT-12), Hip Outcome Score-Sports Specific Subscale (HOS-SSS), and the Visual Analogue Scale (VAS). Receiver Operating Characteristic curves were utilized to establish PASS thresholds. Model integrity was evaluated through the Area Under the Curve (AUC). False-negative PASS was defined as responding “yes” to the PASS question but failing the PASS threshold whereas false-positive PASS was defined as responding “no” to the PASS question but achieving the PASS threshold. Univariate and multivariate logistic regression analyses were performed to evaluate risk factors associated with inaccurate PASS.
Results
AUC ranged from 0.85 to 0.89 for each PRO threshold. 31.8% (158/497) of patients were found to have inaccurate PASS on any PRO, with 26.0% exhibiting false-negative PASS and 5.8% exhibiting false-positive PASS. Risk factors for false-negative PASS included lower preoperative PROs, lower preoperative activity levels, the inability to return to preoperative activity levels, and subspine decompression (p<0.05). Risk factors for false-positive PASS included higher preoperative PROs and a successful return to preoperative activity levels (p<0.05).
Conclusion
Inaccurate PASS rates were substantial despite excellent construct validity when defining PASS thresholds. Given the unreliability of the PASS threshold, it is necessary to consider reporting a patient’s response to the anchor question, rather than implementing a threshold-based definition which only provides a poor estimate of that answer, in future research.