Validity of a telephone administered modified Japanese orthopaedic association score for degenerative cervical myelopathy

Background The modified Japanese Orthopaedic Association (mJOA) score is the gold-standard measure of neurological functional impairment in degenerative cervical myelopathy (DCM). Although designed for face-to-face (F2F) clinical use, telephone administration has become common, particularly following the shift to telemedicine during the COVID-19 pandemic. POLYFIX-DCM is an ongoing randomised controlled trial of laminectomy and laminectomy and fusion for DCM over 2 or more levels. To enable blinded outcome assessment and centralised follow up, a telephone mJOA was conducted by a small group of investigators. The objective was to evaluate the validity of telephone-administered mJOA assessment. Methods Two prespecified cohorts were analysed . (1 ) Multicentre (NIHR POLYFIX-DCM): F2F assessments by site investigators at enrolment / before surgery compared with a centralised telephone assessment. (2) Single Centre (Cambridge) assessment: both assessments performed by the same investigators. Agreement was evaluated using Bland–Altman analysis, intra-class correlation coefficient (ICC), and the proportion of pairs within ± 1 mJOA point (the predefined acceptability threshold). Reliability was evaluated against the prespecified criterion of ICC ≥ 0.70. Sensitivity analyses examined site, age, severity, and timing. Results The multi-centre cohort included 179 paired assessments (F2F and telephone) with mean difference between assessments + 0.24 points (95% CI − 0.02 to 0.50), ICC = 0.69 (95% CI 0.60–0.76), with majority of pairs within the accepted threshold. The single-centre cohort comprised 26 pairs, ICC = 0.97 (95% CI 0.93–0.98), with 96% within ± 1 point. Bland–Altman plots showed no systematic bias. Agreement did not vary by site, age, or time. Conclusions These findings support the validity of telephone-administered mJOA assessment, particularly when undertaken by trained assessors using a standardised approach. It may therefore be a useful method for remote outcome assessment in DCM trials such as POLYFIX-DCM.