Measuring negative and positive caregiving experiences: a psychometric analysis of the Caregiver Strain Index Expanded

Objective: To compare the psychometric properties of the Caregiver Strain Index Expanded with those of the original Caregiver Strain Index among partners of stroke patients. Design and subjects: Cross-sectional validation study among 173 caregivers of stroke patients six months post-stroke. Main measures:Outcome measure: Caregiver Strain Index Expanded. Reference measures: Hospital Anxiety and Depression Scale, two questions on life satisfaction, Barthel Index and Montreal Cognitive Assessment. Additionally, National Institutes of Health Stroke Scale. Results: Neither the Caregiver Strain Index Expanded nor the original Caregiver Strain Index showed floor or ceiling effects. The sum score of the positive items showed a ceiling effect and was skewed to the right (2.20). Principal component analysis revealed no clear underlying item clustering. Alpha values of the Caregiver Strain Index Expanded and the original Caregiver Strain Index were good (0.82 and 0.83), but the alpha value of the positive subscale of the Caregiver Strain Index Expanded was too low (0.51). Convergent validity was confirmed for the Caregiver Strain Index Expanded, the original Caregiver Strain Index and the positive subscale. The Caregiver Strain Index Expanded and the original Caregiver Strain Index showed nearly identical correlations with the reference measures. Negative caregiving experiences were more strongly related to partners' mood and life satisfaction than positive caregiving experiences. In the regression analyses, the positive subscale showed little added value in predicting partners' mood and life satisfaction. Conclusion: The addition of five positively phrased items does not improve the psychometric properties of the Caregiver Strain Index.