An Assessment of the Longitudinal Construct Validity of the Pain Behavioral Scale (PaBS) in a Saudi Population with Chronic Low Back Pain: A Preliminary Study

Abstract

Background: The Pain Behavioral Scale (PaBS) measures the presence and severity of pain behavior. We examine the longitudinal construct validity of the PaBS using convergent and known-groups approaches on a population of 23 participants with chronic lower back pain (LBP) undergoing routine physiotherapy care and pain neuroscience education. Methods: Participants who satisfied study inclusion and exclusion criteria were recruited from patients who attended two testing sessions at physiotherapy clinics in Saudi Arabia. Participant pain behavior was initially measured using the PaBS scale; participants performed standardized physical tests (e.g., repeated trunk flexion) and provided baseline demographic, clinical data, and self-reported measurements using the Modified Roland and Morris disability questionnaire (MODI), fear-avoidance questionnaire (FABQ), and pain catastrophizing scale (PCS). In subsequent visits, a physiotherapist provided usual care to participants, and weekly sessions were established for online pain-neuroscience education. During week six, participants repeated the same questionnaires and physical performance tests with the PaBS. Paired t-tests are used to compare changes in health characteristics from baseline responses to those in week six. Correlations between changes in PaBS from baseline to week six, with changes in outcome measures (i.e., disability, pain intensity, fear-avoidance beliefs, catastrophizing), were determined. To assess known-group validity, we also used a general linear model. Results: A total of 23 participants completed the PNE and follow-up data collection. The mean change from baseline in the PaBS score was statistically significant, as were changes in MODI, FABQ, and PCS. Almost 70% of participants improved their PaBS scores over the six-week period, with PaBS scores of almost 40% of them improving by three units or more. The change in PaBS score correlated significantly with changes in the PCS-rumination subscale, supporting a proposed approach to estimate convergent validity (r = 0.44, 95% CI = 0.04–0.72, p = 0.035). Conclusions: The mean change from baseline in the PaBS score is statistically significant, as are changes in MODI, FABQ, and PCS, supporting its convergent validity. According to our STarT Back groups, the medium to low-risk group had a lower PaBS score, and high-risk group had a higher PaBS score, indicating that PaBS use in clinical assessment may identify people according to pain-behavior severity, or those at increased risk of developing disability.