Purpose: To examine the relationships among therapist-reported, patient-reported, and objective assessment scores of balance function. Methods: Inpatients with stroke and occupational therapists were recruited. The objective balance scores were measured using the Balance Computerized Adaptive Testing (Balance CAT) system. The therapist and patient-reported scores were evaluated using a visual analogue scale (VAS) and Likert-type scale. Results: Eighty-eight patients and 16 therapists participated. The correlations (r= 0.64 and 0.65; R-squared about 0.42 at baseline and follow-up assessments, respectively) between the therapist-reported VAS scores and the Balance CAT system were larger than those (r = 0.31 and 0.21) between the patient-reported VAS scores and the Balance CAT system. Low correlations (r = 0.27 and 0.26 for VAS and Likert-type scores, respectively) were found between the therapist-reported and patient-reported change scores. Low correlations (r = 0.12-0.17) were found between the change scores of therapist- and patient-reported ratings and those of the Balance CAT system. Conclusions: The therapists' judgments explained
