Tampa Scale for Kinesiophobia in chronic neck pain patients (TSK-neck): structural and construct validity and reliability in a Brazilian population

Background To date, there are no studies in the literature that define the internal structure of the Tampa Scale for Kinesiophobia (TSK) in patients with chronic neck pain based on factorial analysis. As such, we aimed to verify and identify the best structure of the Brazilian version of the TSK in patients with chronic neck pain. Methods We included Brazilian participants aged >= 18 years, both sexes, with self-reported neck pain for more than 3 months and pain intensity >= 3 on the Numerical Pain Rating Scale (NPRS). Dimensionality and number of TSK items were assessed using confirmatory factor analysis (CFA). We tested the following internal structures: structure 1 (1 domain and 17 items), structure 2 (1 domain and 11 items), structure 3 (2 domains and 11 items), and structure 4 (2 domains and 9 items). We used the Pain-Related Catastrophizing Thoughts Scale (PCTS) and the NPRS for construct validity. In addition, we assessed test-retest reliability for the seven-day interval using intraclass correlation coefficient (ICC2,1), Cronbach's alpha to assess internal consistency, and ceiling and floor effects. Results The study sample included of 335 patients. Most were women (77.6%), young adults (similar to 34 years), single (48.4%), with complete primary education (57.3%), physically inactive (66.6%), with a mean pain duration of 46 months and a mean pain intensity of similar to 5 points on the NPRS. Redundancy was found in the following items: item 1 with item 2 (modification indices = 21.419) and item 13 with item 15 (modification indices = 13.641). Subsequently, based on these paired analyses, the items with the lowest factor loadings (items 2 and 15) were excluded. As such, TSK structure 4 was composed of two domains ("somatic focus" and "activity avoidance") and 9 items, which showed adequate fit indices and lower AIC and SABIC values. We observed significant values (p < 0.05) with a correlation magnitude greater than 0.142 to 0.657 between the two domains of the TSK-neck and the other instruments (PCTS and NPRS). We found excellent reliability (ICC2,1 >= 0.96) and adequate internal consistency (Cronbach's alpha >= 0.98) of the TSK-neck. Finally, ceiling and floor effects were not observed. Conclusion The TSK-neck structure with two domains (somatic focus and activity avoidance) and nine items is the most appropriate for patients with chronic neck pain.