The experience of caregiving inventory: further evidence

Background: The aim of this study was to reexamine the construct validity of the Experience of Caregiving Inventory (ECI) using new, independent data from a population of patients and their carers. This involved re-testing the ECI within the stress-coping model, but adding new variables which included independently rated (rather than carer-rated) assessments of the patient's symptoms and disabilities, a rating of social support (this time for the patient rather than the carer) and a measure of a range of service inputs. Method: Data were available on 69 patients and their carers from the PRiSM Psychosis Study. Two regression analyses were performed; the first to establish the extent to which the ECI predicted GHQ scores and the second to examine predictors of the ECI selected from the wider dataset on the basis of their likely relationship to carer appraisal. The second regression analysis was performed in two stages, allowing the effect of service factors to be assessed after controlling for the impact of patient personal characteristics such as illness-related or environmental factors. Results: ECI scores accounted for 27% of the variance of GHQ scores. Over one-third of the ECI negative appraisal can be explained by a combination of patient disability (the Social Behaviour Score Total), extent of patients' social network (Social Network Schedule: Number of Friends)and involvement of a Community Psychiatric Nurse (CPN). At the same level of patient morbidity and informal social network, CPN contact reduced ECI scores. Conclusion: As hypothesised, ECI scores correlated significantly with the other measures in the directions predicted by the stress-coping model; that is, negative caregiving as measured by the ECI predicted carer morbidity, while it in turn was predicted by a combination of stressor variables (patient symptoms and disability) acting to increase it, and mediating variables (social support, service inputs) acting to reduce it. Implications for services are discussed.