What is the baseline inter-rater agreement in oncology adverse event grading?

Published studies find substantial variation. Two reviewers identifying CTCAE toxicities in the same 100 radiation oncology notes agreed at a weighted kappa of 0.59 and an unweighted kappa of 0.68 (Fairchild et al., Int J Radiat Oncol Biol Phys 2020). On attribution, in the placebo arms of two phase III trials 47% and 50% of adverse events were recorded as related to the study drug, and when the same event recurred in the same patient its attribution changed at least once in 31% to 36% of cases (Hillman et al., J Clin Oncol 2010). Authors of an attribution workshop of regulators, sponsors and investigators described processes for determining causality as often 'sub-optimal, unreliable, and inefficient' (George et al., Cancer Treat Rev 2019). Burna AI's own agreement with expert clinicians is being measured in a concordance study of 1,600 cases against an adjudicated two-clinician reference panel, reading out in December 2026; the design is at burna.ai/validation.