Evaluating the use of the EORTC PRO measures for improving inter-rater reliability of CTCAE ratings

This international trial showed that giving clinicians access to patient-reported outcome (PRO) data improves how consistently they rate treatment side effects in cancer care using CTCAE.

Study aim and design
The study investigated whether PRO data from the EORTC QLQ-C30 and EORTC Item Library could enhance inter-rater reliability of CTCAE ratings for symptomatic adverse events (AEs). A total of 1,067 adults with various cancers receiving chemotherapy, immunotherapy, or radiotherapy were enrolled in 11 hospitals across 10 countries and randomised to CTCAE assessment with or without access to PRO data. The data of 1,013 patients was analysed.

What was measured
Two independent providers (oncologists or trained nurses) graded 17 CTCAE AEs on the same day, and agreement between their ratings was quantified using intraclass correlation coefficients (ICCs). The primary endpoint was inter-rater reliability; secondary endpoints included how often symptoms of any grade were detected and how agreement differed by symptom type.

Key findings
Inter-rater reliability was at least fair in all domains and significantly higher in the PRO-informed group for 13 of 17 symptoms, especially psychological and cognitive problems such as memory and concentration impairment, irritability, anxiety, and depression. For most of these symptoms, clinicians in the intervention group more often agreed on grade 1 or 2 toxicities, indicating better detection of low-grade but clinically relevant burdens.

Implications for trials
These results support integrating PRO data into routine toxicity assessment to complement clinician judgement and provide a more patient-centred picture of treatment harm. Incorporating PROs into CTCAE assessment can strengthen safety monitoring and interpretation of cancer clinical trial outcomes, particularly for symptoms that are not easily observed in routine consultations.

The study is closed.

The study explores whether two different clinicians assess more accurately the symptom burden of a patient, if they previously receive a standardized summary of how the patient perceives their symptoms. This is necessary, as sometimes symptoms go unnoticed and two independent clinicians rate the symptoms of one single patient very differently. If the patient’s answers to special questionnaires are used to guide the ratings by clinicians, this can influence e.g. which side effects are listed in drug package inserts and if patients receive treatment for these side effects as part of their routine care.