The construct validity and responsiveness of measures of the predictability of intermittent knee pain in knee osteoarthritis.

Objective: The Intermittent and Constant Osteoarthritis Pain (ICOAP) questionnaire was developed to assess the osteoarthritis (OA) pain experience, but does not currently incorporate pain predictability, which people with OA considered important. We assessed the construct validity and responsiveness of two supplemental questions assessing the frequency of predictable and unpredictable intermittent knee pain. Design: This was a secondary analysis of a prospective cohort of individuals aged 30 years or older undergoing total knee arthroplasty, TKA, for knee OA. Standardized questionnaires assessed socio-demographics, health status, ICOAP intermittent and constant pain subscales, the frequency of predictable and unpredictable intermittent knee pain (5-point Likert scale from never to very often), KOOS-QOL, and two ICOAP comparator measures (Pain Catastrophizing, and Perceived Arthritis Coping Efficacy), pre- and 12-months post-TKA. Construct validity was assessed by examining the Spearman correlations between pre-TKA predictability scores, ICOAP, KOOS QOL and comparator measures. Responsiveness was assessed with standardized response means (SRMs) for pre-post TKA change in predictability scores. Results: Of 1366 participants (mean age 67.2 years; 60.8 ​% female), 1302 had intermittent knee pain. Predictable and unpredictable pain were reported 'very often' by 35.3 ​% and 18.1 ​%, respectively. Spearman correlations for frequency of predictable intermittent knee pain with KOOS QOL, PCS and Coping Efficacy were 0.39, 0.32 and -0.15, respectively; for unpredictable pain, correlations were 0.46, 0.46 and -0.27, respectively (p ​< ​0.0001 for all). SRMs were 1.23 for predictable and 1.31 for unpredictable pain. Conclusion: Our findings support the construct validity and responsiveness of the two predictability items.