Living with cancer: "good" days and "bad" days--what produces them? Can the McGill quality of life questionnaire distinguish between them?

BACKGROUND

To determine the impact of care on quality of life (QOL), or to detect a change in QOL over time, measures of QOL must remain stable when QOL is stable (test-retest reliability) and change when QOL changes (responsiveness). This study addresses these issues for the McGill Quality of Life Questionnaire (MQOL). Unlike other studies that use disease status to indicate whether QOL has remained stable or changed, in this study the patient determines QOL stability or change. The authors also sought to clarify the determinants of "good" and "bad" days for oncology patients.

METHODS

Patients attending an oncology outpatient clinic or who were being treated by a palliative care service were asked to complete MQOL 4 times: on days they judged to be "good," "average," and "bad" and 2 days after the first completion. They also were asked to directly rate the change in their QOL during the intervals between MQOL completion and to report the most important determinants of their good and bad days.

RESULTS

The test-retest reliability of MQOL as measured by an intraclass correlation coefficient ranged from 0.69 to 0.78. All MQOL scores were significantly different on good, average, and bad days, except for the support subscale, in both clinical settings. Five domains were determinants of QOL: physical symptoms, physical functioning, psychologic well-being, existential well-being, and relationships.

CONCLUSIONS

MQOL's reliability and responsiveness suggest it can be used to determine changes in the QOL of groups. The results allow interpretation of changes in MQOL scores with respect to meaning of the change to oncology patients. This in turn is helpful to determine the sample size required in future studies. Some of the domains important to the QOL of oncology patients are not included in widely used measures of QOL.