A protocol for sensory-motor assessment of brain damaged patients - II. Evaluation of the reliability and validity of the assessment protocol

A pilot study on the sensory-motor assessment of central nervous system damaged patients has been statistically evaluated. The assessment is important not only as a tool for evaluating the different facets in the quality and pattern of movement, but also for measuring the patient's functional capacity. The pilot study consisted of 17 stroke patients hospitalized for rehabilitation. The mean time from the acute stroke was 44 days (range 11-98, S.D. 23). 9 of the patients had right hemiparesis, 7 had left hemiparesis and 1 had a mixed paresis. Each patient was examined simultaneously be 2 senior physical therapists, while, at the same time, 9 of the patients were assessed by an occupational therapist. Inter-tester reliability was tested as well as the different inter-items relationships. Agreement between 2 'testers' was highly correlated for all the parameters (p < 0.002 and r > 0.75). Validity was tested by comparing the functional section of the test with Barthel activity of daily living (ADL) and patients' manual capacity based on Rancho Los Amigos protocol (RLA). Generally, the correlation between parameters of the presented assessment and RLA was stronger than between them and the functional ability measured by ADL. High correlation was found for the mean of the first 5 parameters of the functional movements (r > 0.76, p < 0.009) and for some of the single parameters. No correlation was found for any single parameters of the tonus; but a connection could be demonstrated between mean of the tonus grades of different joints and rates of RLA, i.e. limbs at sitting (r > 0.68, p < 0.023). No connection could be demonstrated regarding ADL and tonus. RAN and 'duration of the hand movement' were correlated (r > 0.89, p < 0.002). High correlation was found between the hand movements and RAN for the mean of movement grades of the different joints; the lowest r > 0.88, p < 0.003 and the highest r > 0.91, p < 0.001. More sophisticated movements showed a slightly higher ratio. Correlation was found between most of the parameters of stance and gait, and RLA. An example of the strongest one was between RLA and the ability to maintain stability r > 0.94, p < 0.001. Significant correlation was found only between the mean of the first 5 functional activities and all the means of the active hand movements (with p between 0.022 and 0.004); and not between the mean of the last 5 functional activities and all the means of the active hand movements. The pilot study of the protocol indicates a satisfactory level of reliability and validity compared with other protocols currently in use. Hopefully the assessment will be applied to a larger sample and for a longer period of time so that it will become a routine tool for patient follow-up.