Assessment of children's own health status using visual analogue scale and descriptive system of the EQ-5D-Y: linkage...

The objective of the study is to derive a scoring system for the EQ-5D child-friendly version (EQ-5D-Y) from a population-based sample of children based on their own health state assessments on the descriptive system and the visual analogue scale (VAS).

We used data from the 2008 and 2010 Raising Healthy Eating and Active Living Kids in Alberta surveys of grade five students aged primarily 10-11 years and their parents in the Canadian province of Alberta. We applied a random split sample approach. Two-thirds of the sample (modeling sample) was used for the estimation of the EQ-5D-Y VAS. The remaining one-third sample (validation sample) was used for the examination of the prediction accuracy of the estimation model. Multilevel linear regression was used to estimate EQ-5D-Y VAS score and produce EQ-5D-Y index values from the children's self-rated states described in the EQ-5D-Y and the accompanying VAS values.

The mean EQ-5D-Y VAS-based index value was 0.891 (SD 0.133) in the modeling sample and 0.885 (SD 0.134) in the validation sample. The own VAS-based index generally showed logical consistency, with lower values for health states that were logically worse. There was no statistically significant difference between the observed and the predicted VAS values in this sample.

This study elicited a value set for health states defined by the EQ-5D-Y that is based on respondents' own VAS ratings from a large representative sample of Canadian children. Future study is needed to validate this type of value set among different age groups of children.