Response shift effects in patients' assessments of their quality of life after cardiac rehabilitation.

Abstract

The effect of intervention programs on health-related quality of life (HRQoL) can be underestimated due to response shift effects. This study aims to compare HRQoL between cardiac patients taking part in a rehabilitation program and the general population and to investigate changes in HRQoL in terms of response shift with two approaches. A sample of 282 cardiac rehabilitation inpatients (response rate: 58.9%) responded to the self-report quality of life questionnaire EORTC QLQ-C30 at baseline (during rehabilitation) and three months later (actual and retrospective judgment). Their HRQoL was compared to that of the general population. Response shift evaluation complemented the thentest with the structural equation modeling approach. Compared to the general population, patients showed impaired quality of life on all scales (Hedges' g between 0.31 and 1.57). The complementation of the thentest with the structural equation modeling approach revealed response shift effects in physical, emotional, cognitive, and social functioning. No effects were found in role functioning. The combination of both the thentest and the structural equation modeling approaches proved to be essential for obtaining comprehensive statistical evidence that response shift can distort measurements of change. Our results suggest that studies that use the thentest to evaluate the effectiveness of interventions should complement their analyses with the structural equation modeling approach to avoid biased effects.