Validation of the Simplified Chinese version of the Core Outcome Measures Index (COMI)

Purpose To translate the Core Outcome Measures Index (COMI) into Simplified Chinese and then validate it for Mainland Chinese patients with low back pain (LBP). Methods A total of 120 consecutive patients with LBP >3 months who visited our outpatient clinic from December 2011 to March 2012 were asked to complete a questionnaire booklet including the following: (1) the Roland Morris disability questionnaire (RMQ) (Fan et al. in Spine 37(10):875–880, 2012 ), (2) the Short Form Health Survey (SF-36) (Zhang et al. in Int J Med Sci 9(7):521–526, 2012 ), (3) the Oswestry Disability Index (ODI) (Liu et al. in Spine 34(11):1211–1216, 2009 ), (4) visual analogue scale (VAS) measure of pain, and (5) COMI. These patients were also asked to complete a second COMI questionnaire and a transition questionnaire (5-point Likert scale: better, a little better, no change, a little worse, worse) and to return the second COMI questionnaire via mail within 1 month. Results The floor effects for the COMI items ranged from 5.8 to 12.5 %. High values (28.3, 27.5, and 25.8 %, respectively) were found for symptom-specific quality of life, social disability, and work disability. Regarding the ceiling effects, the social and work disabilities were relatively high at 17.5 and 24.2 %, respectively. For other items, the values ranged from 0 to 14.2 %. Neither floor nor ceiling effects were found for the COMI summary score. Excellent correlations were found between the COMI pain scores and VAS scores (Rho = 0.89) and between the COMI pain and the SF-36 bodily pain domain (Rho = 0.84). Other individual items and summary scores showed a very good correlation (Rho = 0.54–0.72) with the corresponding questionnaires except for “symptom-specific well-being” (0.31–0.45). One-way repeated measures ANOVA was used to determine the intraclass correlation coefficient (ICC). The ICC for the entire COMI score was 0.91 (95 % CI 0.85–0.94) and 0.81–0.86 for the two pain scores (back and leg). The “minimum detectable change’’ (MDC 95 %) for the COMI summary score was 1.91 points. No significant difference in the mean values was found for the repeated scores of individual items or the summary score. Conclusion The Simplified Chinese version of COMI showed satisfactory reliability and good psychometric properties. This concise questionnaire is suitable for widespread use in Mainland China.