Highlights
- Access to Patient-Reported Outcome (PRO) data significantly improved inter-rater reliability (ICC) for 13 of 17 symptomatic adverse events in a multinational RCT.
- The most substantial improvements in assessment consistency occurred for subjective symptoms such as memory impairment, irritability, and concentration.
- The study supports the systematic integration of PROs into oncology clinical trials to mitigate provider under-reporting and improve the accuracy of safety data.
- While most symptoms showed improved reliability, some objective symptoms like diarrhea showed unexpected results, suggesting the need for nuanced interpretation of multi-source data.
The Limitations of Provider-Only Toxicity Grading
In the landscape of modern oncology, the Common Terminology Criteria for Adverse Events (CTCAE) serves as the universal language for reporting treatment-related toxicities. However, for decades, clinicians and researchers have recognized a fundamental flaw in this provider-centric model: the subjective nature of symptomatic adverse events. When a physician or nurse grades a patient’s fatigue, nausea, or cognitive function, they are essentially interpreting the patient’s experience through a professional lens that may be clouded by clinical bias, time constraints, or a lack of granular insight into the patient’s daily life.
Evidence has consistently shown that providers tend to under-report the frequency and severity of symptomatic adverse events compared to the patients themselves. This discrepancy is not merely a matter of differing perspectives; it has profound implications for drug safety profiles, dose-finding in Phase I/II trials, and the overall quality of life data that informs regulatory approvals and clinical guidelines. To address this, the integration of Patient-Reported Outcomes (PROs) has been proposed as a method to ‘ground truth’ these assessments. The central question of the recent multinational trial published in Lancet Oncology was whether providing these PRO data directly to clinicians at the point of care could improve the reliability and consistency of their CTCAE ratings.
Study Design and Methodology
This multinational, open-label, randomized controlled trial was conducted across 11 hospitals in ten countries, ensuring a diverse and representative cancer population. The study enrolled 1067 adults with various cancer diagnoses who were undergoing active treatment, including chemotherapy, immunotherapy, or radiotherapy. The broad inclusion criteria allowed for a ‘mixed cancer population’ that reflects the real-world complexity of oncology practice.
Intervention and Randomization
Patients were randomly assigned in a 1:1 ratio to either the intervention group or the control group. In the intervention group, providers (oncologists or trained nurses) were given access to the patient’s PRO data—specifically the European Organisation for Research and Treatment of Cancer (EORTC) QLQ-C30 and 16 additional items from the EORTC Item Library—before or during their CTCAE assessment. In the control group, providers performed their CTCAE ratings using traditional clinical interview methods without access to the PRO data.
Endpoints and Statistical Analysis
The primary endpoint was the inter-rater reliability of CTCAE ratings, measured by intraclass correlation coefficients (ICCs). To ensure rigorous data, two independent providers performed CTCAE ratings for each patient. The ICC is a statistical measure that describes how strongly units in the same group resemble each other; in this context, it measured the agreement between the two independent clinicians. Higher ICCs indicate greater reliability and less ‘noise’ in the toxicity data.
Key Findings: A Paradigm Shift in Reliability
The results of the trial provide compelling evidence for the value of PRO integration. Between 2020 and 2024, data from 1013 patients were analyzed. The findings revealed that inter-rater reliability was significantly higher in the intervention group for 13 of the 17 symptomatic adverse events evaluated. This suggests that when clinicians have access to the patient’s own report, their independent assessments become more consistent with one another, likely because they are basing their clinical judgment on a more standardized and accurate baseline of patient experience.
The Subjectivity Gap
The most dramatic improvements in reliability were seen in symptoms that are notoriously difficult to quantify through physical exam or laboratory tests. These included:
- Memory Impairment: ICC difference of 0.176 (p < 0.0001)
- Irritability: ICC difference of 0.161 (p < 0.0001)
- Concentration Impairment: ICC difference of 0.157 (p < 0.0001)
- Depression: ICC difference of 0.126 (p = 0.0012)
- Anxiety: ICC difference of 0.109 (p = 0.0018)
For these neuropsychiatric and cognitive symptoms, the provider’s traditional assessment is often a ‘guess’ based on a brief interaction. PRO data provides a structured history that anchors the clinician’s rating, leading to the observed increase in inter-rater agreement.
The Diarrhea Anomaly
Interestingly, the study found that for diarrhea, reliability was actually higher in the control group (ICC difference -0.066; p = 0.013). This outlier warrants closer inspection. Diarrhea is often graded based on the number of stools per day over baseline—a metric that is relatively objective. It is possible that PRO data, which might capture the patient’s distress or perceived severity of the diarrhea, introduced a subjective element that caused clinicians to deviate from the strict numerical grading criteria of the CTCAE, thereby reducing inter-rater consistency.
Non-Significant Differences
There were no significant differences in reliability for pain, rash, and peripheral sensory neuropathy. For rash, this is expected as it is a visual, objective finding. For pain, the lack of difference might suggest that clinicians are already highly attuned to asking about and documenting pain levels, or that the existing visual analog scales used in standard care already function similarly to PROs.
Expert Commentary and Clinical Implications
The findings of this trial have immediate implications for the design of future oncology clinical trials. Historically, the FDA and EMA have expressed interest in ‘Patient-Reported CTCAE’ (PRO-CTCAE) as a secondary endpoint. This study goes a step further by suggesting that PROs should not just be a secondary endpoint but a primary tool used to inform the ‘official’ provider-based CTCAE ratings.
By improving the ICC, PRO data essentially reduces the ‘measurement error’ in clinical trials. In a trial setting, lower measurement error means higher statistical power and a more accurate representation of the drug’s safety profile. For clinicians in daily practice, this data suggests that using structured patient questionnaires before a consultation can streamline the visit, ensuring that subtle but impactful symptoms like ‘brain fog’ or irritability are not missed or misgraded.
Addressing Study Limitations
While the results are robust, the open-label nature of the trial is a limitation. Providers knew whether they were seeing the PRO data, which could theoretically influence their effort in the assessment. Furthermore, the study focused on symptomatic events; it does not replace the need for objective laboratory and imaging-based toxicity monitoring. The challenge remains in how to integrate this into high-volume clinics without adding significant administrative burden to the healthcare team.
Conclusion
The integration of EORTC patient-reported outcome data into the CTCAE assessment process represents a significant advancement in oncology trial methodology. By bridging the gap between patient experience and provider assessment, PROs enhance the reliability of symptomatic adverse event detection, particularly for cognitive and emotional toxicities that are often undervalued in clinical reports. As oncology moves toward more patient-centered care, the ‘voice of the patient’ is proving to be not just an ethical necessity, but a statistical one as well.
Funding and ClinicalTrials.gov
This study was funded by the EORTC Quality of Life Group. The trial is registered at ClinicalTrials.gov, number NCT04066868.
References
- Wintner LM, et al. Inter-rater reliability of CTCAE assessments with or without EORTC patient-reported outcome data in a mixed cancer population: a multinational, open-label, randomised controlled trial. Lancet Oncol. 2026;27(2):233-242.
- Bentley TG, et al. Patient-reported outcomes in cancer clinical trials: a review of FDA approvals 2017-2022. J Natl Cancer Inst. 2023.
- Basch E. The missing voice of patients in drug-safety reporting. N Engl J Med. 2010;362(10):865-869.
