Incorporating carer effects into economic evaluation.

BACKGROUND: Despite great interest in the development of methods used in the economic evaluation of health technologies, the effects of carer costs and quality of life (QOL) on the cost effectiveness of treatments has not been widely explored. Yet carer effects are clearly evident in the literature and relevant to the perspective of many published economic evaluations.
OBJECTIVE: To examine whether patient QOL is associated with carer time and carer QOL.
METHODS: We used two datasets to investigate carer effects. Firstly, we used 40,312 cases from the Health Outcomes Data Repository (HODaR) to assess the relationship between patient utility, using the EuroQoL (EQ)-5D, and the number of days care needed from friends and relatives. The stability of the relationship across patient subgroups was assessed by replicating the analysis in ten disease groups. Secondly, we used 64 cases from a study of patients with Alzheimer's disease and their primary carer. These data allowed us to estimate the relationships between patient and carer utility, using the EQ-5D, and patient utility and carer burden using the Community Dementia Quality of Life Profile (CDQLP).
RESULTS: For carer time, a linear model showed that each 0.1-point reduction in patient utility was associated with a 2.5-day increase in carer time over a 6-week period. A more general model, based on EQ-5D domain scores, was better specified and showed that decreased functioning within each domain is associated with increased carer time. Problems with self-care and usual activities have the greatest impact on carer time. These models do not appear to be stable across different disease groups. For carer utility, the relationships between carer and patient utility have low explanatory power and are poorly specified. A clearer relationship is shown between carer burden and utility, such that when sociodemographic covariates are introduced, the relationship reaches conventional levels of statistical significance (p < 0.05).
CONCLUSIONS: The preliminary work described here shows that improving patient QOL may reduce the need for carer time and improve carer QOL. Incorporating such effects into economic evaluations will change cost-per-QALY estimates, with the size of reduction dependent on the domains of health affected by treatment. Clinical studies need to capture carer data so that we can better understand these effects, and how they impact on economic evaluation.