Presentation (CTI)

OBJECTIVES: Assessing quality of life (QoL) in Huntington’s disease (HD) is challenging due to progressive cognitive decline, which limits patients’ ability to self-report, particularly in later stages. Proxy reporting offers a practical alternative but introduces systematic discrepancies, as caregivers tend to rate patients’ QoL more critically. While parallel patient- and proxy-reported QoL measures have been previously developed and validated, their scores are not directly comparable. This study aimed to (1) establish a common measurement metric enabling comparability between patient and proxy perspectives using Rasch model, and (2) derive an adjustment algorithm to estimate patient-reported outcomes from proxy assessments.
METHODS: Data from a previously conducted multi-country study of 238 HD patients and 150 caregivers were analysed. Rasch modelling was used to link patient and proxy scales onto a common metric using a common-item equating design. Model fit and differential item functioning were evaluated. A cross-walk table was derived to translate scores between proxy and patient scales. Discrepancies between paired patient-proxy measurements were examined, and regression-based methods were applied to develop an algorithm for predicting patient QoL from proxy ratings.
RESULTS: Linking procedures successfully placed both perspectives on a common continuum, enabling direct comparability of scores. Proxy ratings were systematically lower than patient ratings (mean difference 0.36 logits, p<0.05), with moderate correlation (r=0.67), and 19% of paired comparisons showing significant differences. The cross-walk table enabled transformation between proxy and patient scores. Regression-based adjustment reduced systematic bias and improved agreement, supporting the feasibility of estimating patient QoL when self-report is unavailable.
CONCLUSIONS: Rasch-based linking provides a robust framework for integrating patient and proxy QoL assessments in HD, enabling continuity of measurement across disease stages. Although systematic differences exist, proxy data can be aligned with the patient perspective through linking and adjustment methods.