The joint impact of cognitive, functional, and behavioral statuses must be measured when exploring the impact of new drugs on Alzheimer's disease (AD) costs. There are very few recent studies of AD costs by all dimensions of disease severity. Our objective was to improve estimation of the relationship between AD severity and costs of AD care by using more comprehensive AD data severity and a large sample size.
Participants were community-dwelling AD patients recruited between 2003 and 2005 and followed annually during a 2-year period in 50 French memory clinics. We used the Resource Use in Dementia questionnaire to estimate costs from a societal perspective. We explored the presence of potential endogeneity bias by using instrumental variable regressions.
Cognitive declines impacted informal costs more than medical and nonmedical costs, while functional declines impacted nonmedical costs more than medical and informal costs. Both cognitive and function declines increased the total costs of care. We found that the endogeneity of these variables led to a large underestimation of their impact of AD severity on costs.
Potential endogeneity should be controlled for to prevent biased estimations of the impact of AD severity measures on costs.