OBJECTIVES: For over a decade, DEMQOL/DEMQOL-Proxy have provided a practical way of reporting health-related quality of life (HRQL) in dementia clinical research. When people with dementia are no longer able to self-report, DEMQOL-Proxy provides an alternative, though this is independent of the person with dementia’s own view. We aimed to provide: a practical way of using DEMQOL-Proxy on a large scale, psychometric justification for linking DEMQOL-Proxy scores to the person with dementia’s own self-report and a way of understanding what a person’s score means.

METHODS: We conducted four analyses (n=1434 dementia patients; n=1030 carers attending a first appointment at Memory Assessment Services): 1) Classical Test Theory to evaluate a self-administered version of DEMQOL-Proxy; 2) Rasch Measurement Methods (RMT) to investigate whether DEMQOL/DEMQOL-Proxy could be placed on a common metric to develop an equated score for DEMQOL-Proxy; 3) anchor and distribution based methods to determine estimates of minimal important difference (MID); and 4) RMT to identify the item descriptions related to this clinical difference.

RESULTS: