Exploration of combined factors associated with unacceptable knee pain in patients with end-stage knee osteoarthritis awaiting knee arthroplasty: a cross-sectional study using association rule analysis

The aim of this study was to identify biopsychosocial factor combinations associated with an unacceptable Knee injury and Osteoarthritis Outcome Score (KOOS) pain state in patients with knee osteoarthritis (KOA) awaiting knee arthroplasty. This cross-sectional study included consecutive patients with radiographic KOA scheduled for knee arthroplasty at a secondary care hospital in Japan (October 2024 to January 2026). Eligible patients had unilateral KOA (Kellgren–Lawrence grade ≥2) and were ambulatory without a walker (cane permitted). Major exclusions included bilateral KOA, inflammatory arthritis, osteonecrosis-related arthroplasty, and revision surgery. The primary outcome (consequent) was an unacceptable KOOS pain state (KOOS pain <73.6). Other dichotomized biopsychosocial and functional variables were treated as antecedents. Association rule mining (apriori algorithm; R, arules) identified multi-factor patterns. Rules were extracted using prespecified thresholds (support ≥0.10, confidence ≥80%, lift ≥1.10) with 95% bootstrap confidence intervals (1000 resamples) and evaluated using contingency table testing (Fisher’s exact test or χ 2 test). The algorithm generated 6060 candidate rules; 5 single-antecedent and 68 two-antecedent rules met extraction thresholds, and 28 two-antecedent rules differed between applicable and non-applicable groups. Representative high-confidence combined patterns included fear of movement plus depressive symptoms (support 39/101, confidence 100%, lift 1.16; p = 0.001) and unacceptable KOOS symptoms plus fear of movement (support 38/101, confidence 100%, lift 1.16; p = 0.002). In patients with end-stage KOA awaiting arthroplasty, unacceptable pain clustered with greater symptom burden and psychosocial distress. These findings support combined screening to inform risk stratification and integrated care planning. Considering this study is exploratory and cross-sectional, causality cannot be inferred, and generalizability beyond surgical candidates may be limited. • Association rules link biopsychosocial combinations to unacceptable KOOS pain. • As single factors, unacceptable KOOS symptoms and fear of movement were associated. • Fear of movement plus depressive symptoms was linked to unacceptable KOOS pain. • Unacceptable KOOS symptoms plus fear of movement was linked to unacceptable pain. • Combined patterns may support more personalized knee pain management.