Validation of the NCI patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE) in women receiving treatment for metastatic breast cancer.

Abstract

9144 Background: NCI PRO-CTCAE is designed to enhance adverse event reporting by integrating patients’ self-report of the frequency (F), severity (S) and interference (I) with usual activities of 78 symptomatic treatment toxicities. This study examined the construct validity of a subset of PRO-CTCAE items in women with metastatic breast cancer (MBC). Methods: 207 women (70% aged 45-59 years; 94% White;71% college-educated) with HER2+ MBC who had received treatment in the past month were recruited from 6 U.S. breast cancer support groups and completed a web survey that collected 18 PRO-CTCAE symptoms, and the Rotterdam Symptom Checklist (RSC). Pairwise concordance among PRO-CTCAE symptom dimensions was examined using weighted Kappa and Bowker’s test for symmetry. Results: Respondents were a median of 47 months since MBC diagnosis and 61% rated their health-related quality of life (HRQL) as good to excellent. Symptom prevalence was similar for PRO-CTCAE and RSC, with respondents more likely to endorse mood disturbance on PRO-CTCAE (Anxiety/Worry 90%; Sad/Unhappy Feelings 86%) vs. RSC (Anxiety 63%; Depressed Mood 61%). There was parallel rank-ordering of fatigue, anxiety, insomnia, depression, difficulty concentrating and neuropathy as the symptoms that were most severe, interfered most and caused the greatest bother. Within PRO-CTCAE, pairwise agreement among F, S and I was moderate for most symptoms (κw=.42 to .54). Agreement between F and S was highest for pain, nausea and arm/leg swelling (κw=.61 to .80), and lowest for anxiety/worry and sad/unhappy feelings (κw= .27-.37). Except for arm/leg swelling, endorsement patterns by the dimensions of F, S and I were distinct (Bowker’s p all <.002). Across PRO-CTCAE symptoms, S was consistently higher than I (mean differences .12 to .72, all p<.01]). Means for S were consistently and significantly higher for those with impaired versus preserved HRQL, providing evidence of construct validity. Conclusions: This study supports the construct validity of PRO-CTCAE, and suggests that F, S and I dimensions offer non-overlapping information relative to 17 /18 PRO-CTCAE symptomatic toxicities.