Assessing agreement between terminally ill cancer patients' reports of their quality of life and family caregiver and palliative care physician proxy ratings.

CONTEXT: Proxy ratings, if valid, may provide an alternative approach to evaluating patient quality of life (QoL) at the end of life.
OBJECTIVES: To examine agreement between terminally ill cancer patients' self-reported QoL and proxy assessment of patient QoL by their family caregiver (FCG) and palliative care physicians (PCPs) at two time points.
METHODS: Patients admitted to an acute palliative care unit and their FCGs and PCPs completed the McGill Quality of Life Questionnaire (MQOL) at Days 3 and 6 after admission. Response bias and response precision were examined at the individual and group levels. Furthermore, we examined patient factors affecting agreement and responsiveness of proxy MQOL scores to changes in patients' QoL between Days 3 and 6.
RESULTS: Statistically and clinically significant mean differences were detected between the patient and both proxy groups' reports of QoL on Day 3, with the magnitude of the differences decreasing somewhat by Day 6. Proxies underestimated patients' QoL compared with patients' self-report. Response precision based on intraclass correlation values and proportion of approximate agreement was poor to fair at both time points. Agreement was better for patients with greater physical burden and more cognitive difficulties. Proxies' responsiveness to change from Day 3 to Day 6 was low, and proxies were not able to detect minimally important changes in QoL.
CONCLUSION: The findings suggest that moderate agreement between patient and proxy ratings of QoL develops over time but that precision at the individual level, which is more clinically relevant, is less reliable. New strategies for improving proxy reliability are needed.