[Validation of procedures of rehabilitation after cerebrovascular accident].

There are several methodological difficulties in the assessment of the effectiveness of rehabilitation in stroke patients. Obviously, double-blind studies are not possible, and only single-blind procedures can used. It is difficult to control for non specific effects, such as spontaneous recovery, patient's and therapist's motivation, social environment. It is necessary to assess the effect non only at the level of impairments, but also on disability and handicap. However, there are now accumulating data in the literature suggesting that stroke rehabilitation has a significant, although mild, effectiveness. Patients treated in specialized stroke rehabilitation units obtain a better outcome, in terms of independence in daily-life activities, than those treated in general wards. They also have shorter hospitalisation durations and are more frequently able to return home. Treatment effectiveness is related to intensity and duration of rehabilitation, and also to stroke severity. Patients with moderate impairments seem to benefit more from treatment than patients with mild or severe deficits. However, significant improvements can still be obtained in very severe cases, and even late (up to two years) post stroke. Similarly, rehabilitation of cognitive deficits (aphasia and unilateral neglect) has also been found efficient in most studies, even if the beneficial effect is relatively small. Aphasic patients treated by speech therapists improve more than patients treated by non specialized therapists or by family members who received a short training. One limitation of neglect rehabilitation is the inconsistent generalisation of treatment effects to daily-life situations. These data are encouraging but further research is needed to find out what precisely works, and how, in the "black box" of rehabilitation.