How do different health states impact preferences for social care related quality of life: An exploration of EQ-5D-5L and ASCOT using a vignette approach.

Objectives: Social care related quality of life instruments, such as the Adult Social Care Outcomes Toolkit (ASCOT) are important for valuing interventions, particularly when they include utility weights. Utility weights may vary by population and health conditions, yet most value sets are developed without considering contexts. This study aims to explore this issue by assessing whether the valuation of ASCOT social states differs in the context of poor physical or mental health versus full health.

Methods: An online discrete choice experiment was conducted with a representative sample of 654 respondents in Australia. Respondents completed choice tasks choosing between ASCOT social care states, while imagining living in a poor health (EQ-5D-5L) state for half of the tasks and in full health for the other half. Respondents were evenly assigned to two study arms (poor physical or mental health). Data were analysed using multinomial and mixed logit models, with interaction terms assessing the impact of context.

Results: In the mental health arm, the poor health context had no statistically significant impact on ASCOT estimates compared to the full health context. In the physical health arm, there were significant differences in the coefficients of "control" (level 4), "safety" (level 2) and "accommodation" (level 4). However, in all cases, "control, "safety" and "cleanliness" accounted for over 50% of the relative attribute importance.

Conclusions: This study found limited influence of health context on valuation of social care-related quality of life dimensions, providing evidence supporting development of ASCOT value sets independent of context.