Are the EQ-5D-5L and EQ-HWB-9 valid instruments for measuring informal carer outcomes for use in economic evaluations?

OBJECTIVES

Despite growing interest in incorporating informal carer outcomes into economic evaluations, evidence on the performance of utility measures for informal carers remains limited. This study assesses the psychometric performance of the EQ‑5D‑5L and EQ‑HWB‑9 in capturing informal carer outcomes for economic evaluations.

METHODS

This study utilized cross-sectional EQ-DAPHNIE survey data from English‑speaking countries (Australia, Canada, New Zealand, UK, and US) to evaluate the EQ‑HWB‑9 (UK pilot value set) and EQ‑5D‑5L (country‑specific value sets), assessing distributional (i.e., acceptability, distribution, floor and ceiling effects) and construct validity (i.e., convergent and known-group) properties, with subgroup analyses by age and financial situation.

RESULTS

There were 6,743 informal carers across Australia (n = 1447), Canada (n = 1265), New Zealand (n = 1435), UK (n = 1337), and US (n = 1259). The EQ-5D-5L showed a ceiling effect (19%), but not the EQ-HWB-9 (5%). Most respondents reported problems (level 2-5) on the EQ-5D-5L pain/discomfort (67%) and anxiety/depression (60%), with anxiety/depression showing the highest proportion with worst (level 5) responses (5%). For the EQ-HWB-9, ≥ 50% reported problems (level 2-5) across all domains except mobility, with ≥ 5% reporting the worst (level 5) responses in exhaustion, loneliness, cognition, sad/depressed, anxious, and control. Both instruments moderately correlated with the WHO-5, PROMIS-10 (physical and mental t-scores), ASCOT-SCT4, OPQOL-brief, PHQ-2, and GAD-2, and demonstrated known-group validity across groups relating to self-reported physical and mental disability, chronic conditions, sleep, social support, impact of caring, psychological wellbeing, and self-reported quality-of-life. Results were consistent in subgroup analyses.

CONCLUSION

This study provides evidence supporting the acceptability and construct validity of the EQ-5D-5L and EQ-HWB-9 in measuring informal carer outcomes for use in economic evaluations. Responsiveness evidence remains limited and warrants further research.