Perspectives on epilepsy in people with intellectual disabilities: comparison of family carer, staff carer and clinician score profiles on the Glasgow Epilepsy Outcome Scale (GEOS).
Espie CA., Watkins J., Duncan R., Sterrick M., McDonach E., Espie E., McGarvey C.
PURPOSE: Clinical practice with people with intellectual disability relies heavily upon caregiver report. Crucially, the carer's perspective may depend upon his or her relationship to the patient. We investigated similarities and differences within and between family and paid carers in their reports on the Glasgow Epilepsy Outcome Scale (GEOS), an instrument that quantifies concerns about epilepsy in this population [Epilepsia 42 (2001) 1043]. METHODS: GEOS forms were available on 186 patients (108 males; mean age 39 years) across 384 primary respondents (141 staff, 83 family, 160 clinicians) and independently completed secondary respondents (67 staff, 36 family). Data were analysed to consider levels of concern as rated bv staff carers, family members and clinicians, and also to consider inter-rater agreement on the concerns raised. RESULTS: There were significant differences in the magnitude of concern on each sub-scale [concerns about seizures, treatment, caring and social impact; range of F(2,171)=9.5-64.7; all P<0.0001]. Post hoc testing revealed that family members scored all sub-scales more highly than staff carers or clinicians, and that staff carers scored more highly than clinicians on all but one sub-scale. Inter-rater agreement between family members was considerably higher (range of r=0.69-0.91) than between staff carers (r=0.30-0.47) across the GEOS sub-scales. Association between staff and family ratings was also modest (r< or = 0.39). CONCLUSIONS: It is preferable for the same staff member to complete each administration of the GEOS because of inter-staff variability in reporting of concerns. Families provide a consistent, but more extreme, picture and clinicians generally underestimate the concerns of direct caregivers. However, content of concerns varies relatively little across respondents.