Anchoring vignettes for health comparisons: an analysis of response consistency

• Published in: Quality of life research Vol. 23; no. 6; pp. 1721 - 1731
• Main Authors: Au, Nicole, Lorgelly, Paula K.
• Format: Journal Article
• Language: English
• Published: Cham Springer 01.08.2014; Springer International Publishing; Springer Nature B.V
• Subjects: Adolescent; Adult; Aged; Australia; Biomarkers; Design; Feasibility Studies; Female; Follow-Up Studies; Health; Health care delivery; Health Status; Health Surveys; Humans; Interviews as Topic; Legs; Male; Medicine; Medicine & Public Health; Middle Aged; Mortality; Pain; Patient Outcome Assessment; Pilot Projects; Population; Public Health; Quality of Life; Quality of Life Research; Quality-Adjusted Life Years; Questionnaires; Regression Analysis; Reproducibility of Results; Risk Assessment; School surveys; Self Report; Social Class; Sociology; Surveys and Questionnaires - standards; Universities; Walking; Young Adult; Australia; Self-rated health; EQ-5D-5L; Response consistency; Qualitative; Anchoring vignettes
• ISSN: 0962-9343; 1573-2649
• DOI: 10.1007/s11136-013-0615-2

Purpose Self-rated health (SRH) is widely used to measure and compare the health status of different groups of individuals. However, SRH can suffer from heterogeneity in reporting styles, making health comparisons problematic. Anchoring vignettes is a promising technique for improving inter-group comparisons of SRH. A key identifying assumption of the approach is response consistency—that respondents rate themselves using the same underlying response scale that they rate the vignettes. Despite growing research into response consistency, it remains unclear how respondents rate vignettes and why respondents may not assess vignettes and themselves consistently. Method Vignettes for the EQ-5D-5L were developed and included in an online survey. In-depth interviews were conducted with participants following survey completion. Response consistency was examined through qualitative analysis of the interview responses and quantitative coding of participants' thought processes. Results Our analysis showed that anchoring vignettes for the EQ-5D-5L is feasible, but that response consistency may not hold for some participants. Respondents are more likely to rate their own health and vignettes in the same way if presented with overall health state vignettes than single health dimension vignettes, and if they imagined themselves in the health state of the hypothetical individual. Conclusion This research highlights opportunities to improve the design of anchoring vignettes in order to enhance response consistency. It additionally provides new evidence on the feasibility of employing anchoring vignettes for the EQ-5D-5L, which is promising for future work to address reporting heterogeneity in the EQ-5D-5L.