Proxy responses to ICECAP-A: Exploring variation across multiple proxy assessments of capability well-being for the...

The ICECAP capability measures are increasingly being used to capture the impact of health and social care interventions on well-being. In cases where the recipient of an intervention is highly vulnerable, proxy completion may be necessary. This study adds to the limited existing evidence on proxy completion of ICECAP-A specifically and adopts the novel approach of investigating multiple proxy responses for the same four (hypothetical) individuals.

62 members of the public who were participating in a series of one day deliberative workshops on public health and social care completed ICECAP-A on behalf of four hypothetical individuals, described in vignettes. Quantitative analysis explored the range of proxy responses for each of the four hypothetical individuals, and any possible correlation between participants' own characteristics and their proxy responses. Participants discussed their proxy responses after completing the task; this discussion was audio recorded and analysed using Framework Analysis.

Wide variation in ICECAP-A scores was observed across proxy respondents for each hypothetical individual. Participants' demographic characteristics and own well-being do not appear to have systematically influenced proxy responses. Qualitative analysis suggests two principal approaches (or perspectives) were adopted by participants: Empathetic (adopting the perspective of the 'subject') and factual (a factual assessment of the subject's well-being). Participants also drew on their own experiences to varying degrees. There were differing interpretations of the Independence attribute on ICECAP-A and some evidence that participants' ideas of what constituted achievement were context (including life-stage and condition/health) specific.

The factual versus empathetic approaches identified from qualitative analysis in this study match to the concept of a proxy-proxy versus proxy-patient perspective, previously outlined in the literature. Researchers should consider specifying which perspective proxy raters should adopt. Findings also suggest proxy responses can be influenced by external points of reference and interpretation of measure attributes.