INTRODUCTION: This study aims to assess the feasibility, acceptability, and validity of EQ-5D instrument administration methods and proxy selection for evaluating health-related quality of life (HRQoL) in dementia populations. EQ-5D is a widely used measure of HRQoL and is recommended by the National Institute for Health and Care Excellence for cost-effectiveness analyses of health interventions. METHODS: Individual-level data from three trials were analyzed separately to evaluate missing data rates, inter-rater agreement, responsiveness, and predictors of EQ-5D (EQ-5D-3L and EQ-5D-5L) dimensions and index values. The study used psychometric analyses, correlations, and multivariate linear regression models to evaluate EQ-5D dimension reports. Reports from both people with dementia (PwD) and proxies were compared to assess reliability across different settings and proxy types. RESULTS: Proxy-reported EQ-5D achieved higher completion rates compared to reports from PwD, with proxies showing greater responsiveness to changes in symptom scores over time. Face-to-face instrument administration for informal proxies was favored over postal methods, and proxy selection was found to be crucial, with informal proxies recommended for community-dwelling PwD and staff proxies for institutionalized populations. Inter-rater agreement was strongest for the "mobility" dimension, with differences in reporting by dimension. Novel guidelines on integrating EQ-5D data reported by PwD and proxies are proposed. Combining self- and proxy-reported data to generate an integrated utility score potentially reflects a more holistic perspective and may enhance the accuracy of HRQoL assessment, compared to relying solely on one respondent's reports. DISCUSSION: The importance of careful administration and proxy selection for HRQoL data collection and application in dementia trials and studies is highlighted. These findings have implications for informing economic evaluations of dementia interventions, emphasizing the potential need for tailoring approaches to HRQoL assessment based on the residential status of the PwD. HIGHLIGHTS: The EQ-5D is a widely used measure in dementia trials, but challenges like missing data and discrepancies in inter-rater agreement highlight complexities in assessing health-related quality of life (HRQoL), especially in advanced stages.This study empirically examines the feasibility, acceptability, and validity of EQ-5D data from people with dementia and proxies, using individual-level data from three trials.Recommendations are provided to improve data collection practices, enhancing the reliability of HRQoL assessments in clinical trials for dementia interventions.Optimized proxy selection criteria and administration methods tailored to residential status can improve HRQoL assessments, supporting more accurate economic evaluations and better-informed care decisions.