Abstract
In this final report is decribed the quantitative consideration on the scale or grade of the test for motor recovery in hemiplegia. Crosssectional data of newly involved 161 patients were joined to the previous data, since they seemed alike. Thus the data of 314 subjects, 289 upper and 247 lower limbs, were subjected to the examination for eliminating the redundancy of test. With the aid of scaling me thods including Guttman's, merging the subtest categories and selecting the sub tests among those initially proposed were performed.
For each of the upper and lower limbs, were selected 11 subtests, 2 being redefined as 3categorical and 9 as 2categorical. Of this simplified test, 3 types of scalings which makes 11 or 12 recovery intervals have been elaborated: i) total score, ii) Grade (A) utilizing local total score to reduce the number of error responses, and iii) Grade (B) giving weight to the ordering of subtests. They give virtually the same results, Spearman's rho being nearly 1 among them. A test composed of mainly the remaining subtests and the simplified test mentioned above are found to be quite similar in giving the scale. The reliability and validity of the test composed of 11 subtests are deemed satisfactory for practice.
In 148 subjects, 129 upper and 133 lower limbs, followup data were available. Of these, the shift in the scale of recovery was examined with 3 types of scalings, which appeared practically of the same nature, Spearman's rho being around 0.9 among them. The extent of 8 week shift along recovery process varies depending on the initial, baseline Grade of the subject in both limbs, but the patterns of variation are definitely different between them. Among other factors, the age of the subject in the upper limb and the time lapse after stroke in the lower limb are also influential upon the extent of shift (Step 6).
From various standpoints the test composed of 11 subtests for each limb appears satisfactory for assessment of motor recovery in hemiplegia.
