Order effects in the assessment of quality of life in cancer patients.

ABSTRACT

Methodological studies and outcome research often include several health-related quality of life (HRQoL) measurement scales in one questionnaire. Psychological studies have previously demonstrated that changing the sequential order of measurement scales within a questionnaire can alter the pattern of responses. Little is known, however, about whether there are order effects on the assessment of HRQoL in cancer patients. Here we address this issue in a study of 190 Singaporean cancer patients who were assessed using two different HRQoL instruments placed in alternating sequence within a questionnaire package. Measurement properties of the instruments, including the number of missing values, means, variability, known-group validity and internal consistency, were compared in the two samples representing different presentation orders. The HRQoL instruments administered in different sequential orders appeared to be equivalent in several aspects. No major effect of presentation order on outcomes was shown. The reasons and implications of the absence of an order effect are discussed. We conclude that presentation order is unlikely to alter the responses to the two HRQoL instruments.