COMPARING ADOLESCENTS AND ADULT PREFERENCES TO EQ-5D-Y HEALTH STATES USING PROFILE CASE BEST-WORST SCALING

Purpose: Previous studies have reported that, when presented with the same health state, adults and adolescents provide systematically different valuations. This was recently demonstrated using best-worst scaling (BWS) methods and health states defined from the Child Health Utility-9D (CHU-9D) instrument. In this study we tested the hypothesis that preferences between adolescents and adults using a profile case BWS discrete choice experiment and states from the EuroQol Youth version EQ-5D-Y were different. A secondary objective was to test the feasibility of using BWS to obtain valuations from the EQ-5D-Y.

Method(s): An online survey using a full factorial design of EQ-5D-Y health states divided in 20 blocks was administered to two general representative population samples of adolescents and adults in Spain. An empirical scale parameter (ESP) as a measure of choice consistency was derived for each of the EQ-5D-Y domains (mobility, self-care, usual activities, pain/discomfort, sad/worry) and across the five domains for each participant and population. A sequential (first best then worst) heteroscedastic conditional logit was estimated to obtain part-worth utility values associated to each domain level in each population.

Result(s): 1,006 adults (range 18-over 55 years) and 1,000 adolescents (range 11-17 years) completed the online survey in February 2016. Pain/discomfort and being sad/worry were the most valued domains in both populations indicated by an evenly distribution of ESP in each domain. Not having problems to undertake usual activities was also valued by adolescents. The total ESP across the five domains suggested that adolescents exhibited a slightly larger proportion (16%) of choice inconsistencies with a total ESP<2 or total ESP>5 than adults (14%). The modelling results suggested that adolescents placed more weights on experiencing problems with mobility and pain/discomfort than adults. Adults place more weights on having problems with self-care and being worry/sad than adolescents.

Conclusion(s): Our results suggest that there seem to be age-related differences in elicitation values between adolescents and adults when valuing EQ-5D-Y health states corroborating previous exercises using other instruments. The similar levels of choice consistency between populations provide information about the feasibility of using the EQ-5D-Y as a potential instrument to obtain valuations using profile case BWS. Future work should explore whether our results hold using more sophisticated models to accommodate some of the preference heterogeneity observed in each population.