RESUMEN
PURPOSE: Most economic evaluations of health care programmes do not consider the effects of informal care, while this could lead to suboptimal policy decisions. This study investigates the construct validity of the CarerQol instrument, which measures and values carer effects, in a new population of informal caregivers. METHODS: A questionnaire was distributed by mail (n = 1,100, net response rate = 21%) to regional informal care support centers throughout the Netherlands. Two types of construct validity, i.e., convergent and clinical validity, have been analyzed. Convergent validity was assessed with Spearman's correlation coefficients and multivariate correlation between the burden dimensions (CarerQol-7D) and the valuation component (CarerQol-VAS) of the CarerQol. Additionally, convergent validity was analyzed with Spearman's correlation coefficients between the CarerQol and other measures of subjective caregiver burden (SRB, PU). Clinical validity was evaluated with multivariate correlation between CarerQol-VAS and CarerQol-7D, characteristics of caregivers, care recipients and care situation among the whole sample of caregivers and subgroups. RESULTS: The positive (negative) dimensions of CarerQol-7D were positively (negatively) related to CarerQol-VAS, and almost all had moderate strength of convergent validity. CarerQol-VAS was positively associated with PU and negatively with SRB. The CarerQol-VAS reflects differences in important background characteristics of informal care: type of relationship, age of the care recipient and duration of care giving were associated with higher CarerQol-VAS scores. These results confirmed earlier tests of the construct validity of the CarerQol. Furthermore, the dimensions of CarerQol-7D significantly explained differences in CarerQol-VAS scores among subgroups of carers. CONCLUSION: Notwithstanding the limitations of our study, such as the low response rate, this study shows that the CarerQol provides a valid means to measure carer effects for use in economic evaluations. Future research should derive a valuation set for the CarerQol and further address the instrument's content validity, sensitivity and reliability.