Does The Study Population And The Use Of Proxy Respondent Have An Effect On The Latent Quality Of Life Constructs Measured By The CHU9D And The Pedsqltm 4.0? An Exploratory Factor Analysis

Objectives
An important psychometric property of instruments designed to measure health-related quality of life (HRQOL) is that they must accurately capture the latent HRQOL constructs for different subgroups within the instruments’ target population. This study examined the latent structures of two generic-paediatric-HRQOL measures [the non-preference-based Pediatric Quality of Life Inventory (PedsQL) and the preference-based Child Health Utility 9D (CHU9D)] when used in subgroups that differed according to age and type of respondent (self versus proxy-report).

Methods
Representative cross-sectional data were obtained from two 2014 cohorts of the Longitudinal Study of Australian Children (LSAC) [14-15yrs (n=3,247) and 10-11yrs (n=3,376)] and a separate 2013 independent study (community-cohort, 15-17yrs (n=755). CHU9D is self-reported across all cohorts, whereas PedsQL is proxy-reported by parents in the LSAC cohorts and self-reported in the community-cohort. Latent HRQOL constructs measured by the instruments were identified using exploratory factor analysis (EFA). The optimal number of factors for the EFA was determined using parallel analysis based on simulated polychoric correlation matrices.

Results
A five-factor structure was deemed optimal. In all three cohorts, the PedsQL dimensions loaded onto four distinct factors as the developer originally specified: ‘physical functioning (8 items), emotional functioning (5 items), social functioning (5 items) and school functioning (5 items). In the LSAC cohorts, all CHU9D dimensions loaded onto a separate latent factor, interpreted as general HRQOL while 8 CHU9D dimensions in the Community-cohort loaded onto emotional functioning and 1 onto school functioning.

Conclusions
Age differences did not seem to have an impact on the latent HRQOL constructs measured by the instruments but the use of proxy-respondents did. These results support the validity of using both instruments across a range of ages. That the CHU9D loaded onto a single factor in the LSAC cohorts reinforces the underlying premise of the CHU9D as a measure of health utility or overall HRQOL.