Quality of Life Research (2018) 27:447–454
Evidence on the longitudinal construct validity of major generic
and utility measures of health-related quality of life in teens
John F. Dickerson
· David H. Feeny
· Gregory N. Clarke
· Alex L. MacMillan
· Frances L. Lynch
Accepted: 19 October 2017 / Published online: 17 November 2017
© Springer International Publishing AG 2017
Purpose To examine the longitudinal construct validity in the assessment of changes in depressive symptoms of widely used
utility and generic HRQL instruments in teens.
Methods 392 teens enrolled in the study and completed HRQL and diagnostic measures as part of the baseline interview.
HRQL measures included EuroQol (EQ-5D-3L), Health Utilities Index Mark 2 (HUI2) and Mark 3 (HUI3), Quality of Well-
Being Scale (QWB), Pediatric Quality of Life Inventory (PEDS-QL), RAND-36 (SF-6D), and Quality of Life in Depression
Scale (QLDS). Youth completed follow-up interviews 12 weeks after baseline. Sixteen youth (4.1%) were lost to follow-up.
We examined correlations between changes in HRQL instruments and the Children’s Depression Rating Scale-Revised
(CDRS-R) and assessed clinically meaningful change in multi-attribute utility HRQL measures using mean change (MC) and
standardized response mean (SRM) among youth showing at least moderate (20%) improvement in depression symptomology.
Results Spearman’s correlation coeﬃcients demonstrated moderate correlation between changes in CDRS-R and the HUI2
(r = 0.38), HUI3 (r = 0.42), EQ-5D-3L (r = 0.36), SF-6D (r = 0.39), and PEDS-QL (r = 0.39) and strong correlation between
changes in CDRS-R and QWB (r = 0.52) and QLDS (r = − 0.71). Eﬀect size results are also reported. Among multi-attribute
utility measures, all showed clinically meaningful improvements in the sample of youth with depression improvement (HUI2,
MC = 0.20, SRM = 0.97; HUI3, MC = 0.32, SRM = 1.17; EQ-5D-3L, MC = 0.08, SRM = 0.51; QWB, MC = 0.11, SRM = 0.86;
and SF-6D, MC = 0.12, SRM = 1.02).
Conclusions Findings support the longitudinal construct validity of included HRQL instruments for the assessment of change
in depression outcomes in teens. Results of this study can help inform researchers about viable instruments to include in
economic evaluations for this population.
Keywords Longitudinal construct validity · Health-related quality of life · Teens · Depression
Depression is a debilitating disorder that affects nearly
one in ﬁve youth during their adolescent years and has
substantial negative impact across many domains within
peoples’ lives [1–3]. Health-related quality of life (HRQL)
measures help to enumerate the diverse impact of depression
on an individual’s mental health as well as other domains,
such as physical health and social functioning [4, 5]. Find-
ings from studies evaluating treatment options for depression
that apply HRQL measures as part of the assessment allow
decision-makers to compare eﬃcacy and/or eﬀectiveness
and value across diﬀerent interventions and disorders .
Important considerations in the treatment of depression are
that depression is episodic and many treatment options are
informed by early response and continuation of response to
treatment [7–10]. Thus, HRQL measures must be respon-
sive to longitudinal changes in depression symptomology in
order to be useful in evaluations of depression interventions.
* John F. Dickerson
Center for Health Research - Kaiser Permanente, 3800 N.
Interstate Ave., Portland, OR 97227, USA
Department of Economics, McMaster University, Hamilton,
Health Utilities Incorporated, Dundas, ON, Canada
ThinkShout, Inc, 433 NW 4th Ave, Suite 100, Portland,
OR 97209, USA