Response shift effects of quality of life assessments in breast cancer survivors.

Abstract

Measurements of quality of life (QoL) can be distorted by respondents adapting to new situations between measurement points and consequently having a changed frame of reference. To investigate this bias in breast cancer survivors, we compare their QoL with that of the general population and use two complementary methods for detecting this bias. Breast cancer survivors (n=308, response rate: 91%) were tested with the QoL questionnaire European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core 30. Response shift was examined with the thentest (retrospective judgements) and with the structural equation modelling (SEM) approach. Compared with the general population, breast cancer survivors showed impaired QoL in all functioning scales (Hedges' g: -0.56 to -0.93) and symptom scales (Hedges' g: 0.28-0.74).The thentest method indicated recalibration effects in several dimensions including social functioning. The SEM method detected a non-uniform recalibration effect for social functioning from pretest to posttest and from pretest to thentest, but no effect between thentest and posttest. Breast cancer survivors' QoL is clearly diminished. Comparing the two approaches for detecting response shift showed that it is also useful to apply SEM to retrospective judgements and that this can reveal response shift effects that would otherwise be overlooked.