Inter-rater reliability and construct validity of a cross-diagnostic movement quality score for rehabilitation assessment.

Every patient was measured on one day, so the study says nothing about whether scores change as someone recovers; responsiveness was not tested. Patients were taken because they were available on four Japanese rehabilitation wards, not at random, and anyone who could not complete the assessments was left out, so this group may be more able than the patients you see. The two therapists knew each patient's diagnosis and condition while scoring, which can nudge their scores towards agreement. Age, sex and time since onset were not recorded, so it is hard to judge who these results apply to. The internal medical subgroup contained only 17 patients and its reliability estimates were unstable; summed muscle strength reliability in that group was 0.44, which the authors put down to the small numbers. Sixteen patients had no second rating, and one patient's muscle strength data were lost to a recording error.

The authors declare no conflicts of interest. The work was carried out as part of the Japanese Physical Therapy Association's Standardized Assessment Development Project, and the scale being tested was developed by the assessment committee of that same association.

Do not start using this score to track whether a patient is improving. Each patient was assessed only once, so the study never tested whether the score changes when someone gets better; the precision figures tell you how large a change has to be to escape measurement error, not that the scale can detect real improvement.