Association of the Japanese Orthopaedic Association score with the Oswestry Disability Index, Roland-Morris Disability Questionnaire, and short-form 36

Abstract

Study design: Cross-cultural translation and cross-sectional psychometric testing were performed.

Objectives: To cross-culturally translate the Oswestry Disability Index (ODI) and the Roland-Morris Disability Questionnaire (RMDQ) into Japanese, and to compare the Japanese Orthopaedic Association (JOA) score with the ODI and the RMDQ score.

Summary of background data: The two most widely used back-specific measures, the ODI and the RMDQ, have not been translated into Japanese. The JOA score has been used extensively in Japan. However, this score has not been tested in terms of its reliability and validity.

Methods: The ODI and RMDQ were translated into Japanese using the process of forward translation, synthesis of translation, backward translation, expert committee, test of the prefinal version, and submission of the documentation to the developers. The JOA score, ODI, and RMDQ were tested with 97 patients who had degenerative lumbar spinal disorders (average age, 51 years). The correlation among the three disease-specific measures (JOA score, ODI, and RMDQ) and eight subscales of a generic health measure, the Medical Outcomes Survey Short-Form 36 (SF-36), was calculated. The reproducibility of the JOA score also was investigated.

Results: Reliability, as estimated by internal consistency, reached a Cronbach alpha of 0.83 for the ODI and 0.86 for the RMDQ. The calculated test-retest reliability was 0.93 (P < 0.01; n = 20) for the ODI and 0.95 (P < 0.01; n = 20) for the RMDQ. The correlation of the JOA score with the ODI was -0.647 (P < 0.01), and with RMDQ was -0.568 (P < 0.01). There also was a significant correlation between the ODI and the RMDQ (r = 0.785; P < 0.01). There was a significant correlation between the three disease-specific measures (JOA score, ODI, and RMDQ) and all the subscales of the SF-36 (P < 0.01). The calculated reproducibility of the JOA score was as follows: interobserver error (r = 0.92, P < 0.01), test-retest reliability (r = 0.91, P < 0.01).

Conclusions: The Japanese versions of the ODI and the RMDQ were reliable and valid. The use of these translated instruments can be recommended for future clinical trials in Japan. The results also showed the JOA score had acceptable psychometric properties of reliability and construct validity, suggesting that this score is reliable and valid. Further studies are needed to verify the validity and responsiveness of the JOA score.