Abstract
The status and prognostic value of the disagreement between physician and patient assessments of symptomatic adverse events (AEs) remain unclear for patients with metastatic colorectal cancer treated with first-line cetuximab plus chemotherapy. Paired data on patient-reported outcomes using the EORTC QLQ-C30 and physician-reported outcomes using the NCI-CTCAE for eight symptomatic AEs (fatigue, pain, insomnia, dyspnea, constipation, appetite loss, nausea/vomiting, and diarrhea) were collected from a prospective trial assessing the relationships between treatment efficacy, AEs, and quality of life. The overall agreement rates between patient and physician reporting at 4 weeks ranged from 40.2% to 76.5% for 129 patients. The level of agreement based on Cohen's κ statistics was slight to poor for dyspnea, pain, fatigue, and insomnia, while it was moderate to fair for the remaining AEs. No clinicopathological characteristics of disagreement were found. The underreporting by physicians ranged from 12.5% (nausea/vomiting) to 56.7% (fatigue). The 2-year overall survival (OS) rate was more favorable for patients with high agreement than for those with low agreement (71.2% vs. 46.5%, p = .016), and the agreement status was an independent factor of OS (HR, 2.31; 95% CI, 1.13-4.71; p = .022). For patients who were reported as asymptomatic by the physician, the presence of patient-reported symptoms resulted in a trend toward poor prognostic outcomes for appetite loss, dyspnea, diarrhea, and constipation. These findings provide the clinical importance of the monitoring of patient-reported symptoms that can be complementary to physician-reported data to ensure more accurate clinical outcomes.