[Development of the behaviour observation scale for psychomotor therapy for elderly people with dementia (BPMT-dem)....

This article describes the development of the Behaviour observation scale for Psychomotor Therapy for elderly people with dementia (BPMT-dem). This scale was developed in the late 1980s in order to evaluate the effect of psychomotor group therapy on the cognitive, social, and emotional functioning of elderly people with dementia within the therapy situation. The currently described research investigates inter-assessor reliability and internal consistency of the ten subscales (such as memory, orientation, contact with others, initiative, anxious behaviour, aggressive behaviour) and three domains, as well as the scale's correlation with other observation scales (concurrent validity) in two different patient groups. The first group consisted of clients receiving psychomotor therapy in a nursing home or a psychiatric hospital (N = 130). Inter-assessor reliability (Cohen's kappa) varied between 0.27 and 1.00 the internal consistency of the subscales (Cronbach's alpha) was calculated between 0.46 and 0.86, and that of the domains between 0.76 and 0.81. Based on the findings of this study 15 of the original 88 items were removed, 5 items were moved to other subscales and the text of 2 items was changed. Next, the subscale internal consistency and concurrent validity of this modified (73-item version) BPMT-dem was investigated in a group of 41 elderly participating in a support programme in meeting centres for elderly people with dementia and their caregivers. Cronbach's alpha for the subscales in this study was calculated between 0.47 and 0.86. To determine concurrent validity, the BPMT-dem was compared with (subscales of) the Assessment Scale for Elderly Patients (ASEP), Behavioural assessment for Intramural Psychogeriatrics (BIP), Brief Cognitive Rating Scale (BCRS) and the Cornell Scale for Depression in Dementia. Correlations with related BPMT-dem subscales varied between 0.38 and 0.75. The inter-assessor reliability and concurrent validity of the 73-item version of the BPMT-dem are satisfactory. The internal consistency of 6 subscales is sufficient, and recommendations are proposed to improve the consistency of the remaining subscales. Further research to test the unidimensionality and scalability of the subscales of the BPMT-dem, as well as the effect of the recommended reformulation and removal of items, is recommended. Before the instrument can be used in actual practice, the psychometric qualities of the BPMT-dem73 need to be studied in a larger and in terms of severity of dementia more heterogeneous study sample, so that statements can also be made on which norms to use in various subgroups of elderly with dementia.