Abstract
Purpose
Japanese Orthopaedic Association (JOA) Score is widely used to measure the neurological functions of patients with cervical spondylotic myelopathy (CSM). However, there is a limitation of JOA score system that JOA score is not determined by patients directly and may be influenced by subjective bias from the observers. This study aims to develop patient-reported JOA score (P-JOA) to address this deficiency.
Methods
This prospective cohort study involved 120 patients with CSM from November 2021 to December 2022 in our hospital. We developed the Chinese version of P-JOA based on the 17 − 2 points system version JOA score, which is most widely used in Asia. The P-JOA was completed twice by patients (P-JOA 1 and P-JOA 2) preoperatively while JOA score for the patients was evaluated by a physician blinded to the result of P-JOA, and the results of other patient-reported outcomes (OROs) were collected as well, including the MOS item short from healthy survey (SF-36) score, and Neck Disability Index (NDI). The above results were analyzed using the paired t-test, Intraclass Correlation Coefficient (ICC), and weighted κ-test to evaluate the re-test reliability and validity of the P-JOA, as well as its consistency with the JOA. Bland-Altman plot was used method for assessing the measurement error or agreement in this study.
Results
The mean age of the study population was 54.6 ± 10.1 years. The mean scores of the two P-JOA were 13.4 ± 2.6 and 13.3 ± 2.8, and the difference was not statistically significant (P = 0.595); the ICC of P-JOA 1 and P-JOA 2 was 0.731, and the Cronbach's coefficient was 0.844, and the results suggested that the re-test reliability of the P-JOA was good; The ICC of the P-JOA 2 and the JOA score was 0.932, and the Cronbach's coefficient was 0.965, the results showed that the consistency between the P-JOA and the JOA was excellent. The patients were divided into the severe, moderate, and mild groups according to the JOA score, and there was no statistically significant difference between the P-JOA and the JOA (severe: 7.3 ± 0.5 vs. 6.2 ± 0.6, P = 0.078; moderate: 11.6 ± 0.2 vs. 11.7 ± 0.2, P = 0.551; mild: 15.1 ± 0.1 vs. 14.9 ± 0.2, P = 0.091). The Pearson correlation coefficient between P-JOA and NDI was − 0.104, which is a poor correlation; P-JOA showed a moderate correlation with the physiological function scores in the SF-36, with a Pearson correlation coefficient of 0.464, and correlation with other scores in the SF -36 with poor correlation with other scoring items in the SF-36. Bland-Altman analyses based on difference between the P-JOA 2 and JOA revealed that limits of agreement were − 2.2 to 1.9. Bland-Altman analyses based on ratio between the P-JOA 2 and JOA revealed that limits of agreement were 0.8 to 1.1.
Conclusion
The P-JOA has good re-test reliability and good agreement with JOA. It has the potential to replace JOA for clinical application to eliminate physicians' subjective bias. However, its clinical validity still needs to be clarified by further prospective follow-up studies.