Ulrike Ravens-Sieberer , 1 , 16 , Michael Erhart 1 , Luis Rajmil 2 , 3 , 4 , Michael Herdman 2 , 3 , 4 , Pascal Auquier 5 , Jeanet Bruil 6 , Mick Power 7 , Wolfgang Duer 8 , Thomas Abel 9 , Ladislav Czemy 10 , Joanna Mazur 11 , Agnes Czimbalmos 12 , Yannis Tountas 13 , Curt Hagquist 14 , Jean Kilroe 15 , the European KIDSCREEN Group
30 July 2010
To assess the criterion and construct validity of the KIDSCREEN-10 well-being and health-related quality of life (HRQoL) score, a short version of the KIDSCREEN-52 and KIDSCREEN-27 instruments.
The child self-report and parent report versions of the KIDSCREEN-10 were tested in a sample of 22,830 European children and adolescents aged 8–18 and their parents ( n = 16,237). Correlation with the KIDSCREEN-52 and associations with other generic HRQoL measures, physical and mental health, and socioeconomic status were examined. Score differences by age, gender, and country were investigated.
Correlations between the 10-item KIDSCREEN score and KIDSCREEN-52 scales ranged from r = 0.24 to 0.72 ( r = 0.27–0.72) for the self-report version (proxy-report version). Coefficients below r = 0.5 were observed for the KIDSCREEN-52 dimensions Financial Resources and Being Bullied only. Cronbach alpha was 0.82 (0.78), test–retest reliability was ICC = 0.70 (0.67) for the self- (proxy-)report version. Correlations between other children self-completed HRQoL questionnaires and KIDSCREEN-10 ranged from r = 0.43 to r = 0.63 for the KIDSCREEN children self-report and r = 0.22–0.40 for the KIDSCREEN parent proxy report. Known group differences in HRQoL between physically/mentally healthy and ill children were observed in the KIDSCREEN-10 self and proxy scores. Associations with self-reported psychosomatic complaints were r = −0.52 (−0.36) for the KIDSCREEN-10 self-report (proxy-report). Statistically significant differences in KIDSCREEN-10 self and proxy scores were found by socioeconomic status, age, and gender.
Our results indicate that the KIDSCREEN-10 provides a valid measure of a general HRQoL factor in children and adolescents, but the instrument does not represent well most of the single dimensions of the original KIDSCREEN-52. Test–retest reliability was slightly below a priori defined thresholds.