Treatment preferences in dementia with Lewy bodies: a discrete choice experiment

Abstract Background Treating one symptom of dementia with Lewy bodies (DLB) can exacerbate another, leading to inherent treatment trade-offs. However, there is limited evidence on how people with DLB and their care partners (CPs) value improvements across symptoms, particularly when gains in one area may come at the expense of another. This study aimed to elicit treatment preferences among people with DLB and informal CPs and to quantify the relative importance of treatment attributes. Methods A discrete choice experiment (DCE) was administered within a cross-sectional web survey. In the DCE, respondents chose between hypothetical treatment alternatives defined by six attributes: rate of cognitive and functional decline (C&F), visual hallucinations (VHs), parkinsonism, sleep behaviours, cognitive fluctuations, and risk of brain-related side effects (using amyloid-related imaging abnormalities (ARIA) as a hypothetical example). Preferences were analysed using a random parameters logit (RPL) model applied to pooled responses from patients and CPs. Attribute conditional relative importance (CRI) and exploratory maximum acceptable risk (MAR) were estimated. A simplified orally administered survey was available for people with DLB unable to complete the web survey. Results A total of 144 respondents contributed to the DCE analysis (31 patients; 113 CPs; mean age 61 years; 81% female). Conditional on the attribute levels included in the design, VHs and C&F had the highest CRI and did not differ significantly from one another. Both were associated with larger CRI values than parkinsonism, sleep behaviours, and cognitive fluctuations. VHs also had a significantly larger CRI than risk, whereas C&F did not differ significantly from risk. Exploratory MAR estimates for VHs and C&F exceeded the highest risk level presented (30%), although confidence intervals were wide. Conclusions In this sample, C&F and VHs were associated with the highest relative importance across the attributes included in the DCE. However, given the predominance of CPs, results were weighted towards CP-reported preferences. The findings provide exploratory insights into how symptom improvements may be valued in DLB treatment.