FAMILY AND CAREGIVER SPILLOVER EFFECTS IN ALZHEIMER’S DISEASE COST-EFFECTIVENESS ANALYSES

Abstract

Dementia and Alzheimer's disease (AD) can exact a significant financial, physical and emotional toll on family members and caregivers. Failing to incorporate these “spillover effects” in AD economic evaluations may underestimate the value of the health care intervention. We reviewed current practices in the literature and investigated how spillover effects may influence the cost-effectiveness of AD interventions. We analyzed cost-utility analyses (CUAs) related to AD or dementia published from 2000–2016 using PubMed and the Tufts Cost-Effectiveness Analysis Registry (www.cearegistry.org). We investigated the frequency in which spillover effects are incorporated into CUAs. We also characterized reported spillover costs and health effects, and examined the impact of their inclusion on the reported incremental cost-effectiveness ratios (ICERs). Finally, we compared characteristics of studies with or without spillover effects using a chi-squared test. Of 57 dementia or AD CUAs meeting inclusion criteria, 39 (68%) incorporated spillover cost or health effects. Two-thirds of the dementia/AD CUAs incorporated spillover costs and 21% incorporated spillover health effects; only 18% included both. Studies that included spillover effects were significantly more likely to be conducted from the societal (vs. payer) perspective (p<0.001), and be published between years 2011–2016 (p=0.015). Inclusion of spillover did not vary significantly by funding source, intervention type, or prevention stage. The most common analytic approach for accounting for spillover costs was to add informal caregiving time costs to patient costs (n=32). Few studies included caregiver out-of-pocket costs in the analysis (n=3). When spillover health effects were considered, most studies added informal caregivers’ QALYs to patient values (n=10). Thirty three pairs of ICERs reported results with and without spillover effects. Incorporating spillover effects generally decreased ICERs (making them more favorable) (n=16; 48%) or kept the intervention cost-saving (n=13; 39%); and in fewer cases increased ICERs (n=2; 6%) or produced no change (n=2; 6%). Most dementia/AD CUAs incorporated spillover costs, often as time costs to a caregiver, but considered caregiver health impacts less often. Including spillover effects in analyses generally led to more favorable ICERs. The broader value of a dementia/AD intervention to society may be underestimated without considering these broader spillover effects. Selection of studies and review process.