Annegret Schneider, Grigorios Kotronoulas, Constantina Papadopoulou, Lisa McCann, Morven Miller, Jackie McBride, Zoe Polly, Simon Bettles, Alison Whitehouse, Nora Kearney, Roma Maguire

To examine the trajectories and predictors of state and trait anxiety in patients undergoing chemotherapy for breast or colorectal cancer. Secondary analysis of data collected as part of a large multi-site longitudinal study. Patients with breast or colorectal cancer completed validated scales assessing their state and trait anxiety levels (State-Trait Anxiety Inventory) and symptom burden (Rotterdam Symptom Checklist) at the beginning of each chemotherapy cycle. Longitudinal mixed model analyses were performed to test changes of trait and state anxiety over time and the predictive value of symptom burden and patients' demographic (age, gender) and clinical characteristics (cancer type, stage, comorbidities, ECOG performance status). Data from 137 patients with breast (60%) or colorectal cancer (40%) were analysed. Linear time effects were found for both state (χ(2) = 46.3 [df = 3]; p < 0.001) and trait anxiety (χ(2) = 17.708 [df = 3]; p = 0.001), with anxiety levels being higher at baseline and gradually decreasing over the course of chemotherapy. Symptom burden (β = 0.21; SD = 0.06; p = 0.001) predicted state anxiety throughout treatment, but this effect disappeared when accounting for trait anxiety scores before the start of chemotherapy (β = 0.85; SD = 0.05; p < 0.001). Patients' baseline trait anxiety was the only significant predictor of anxiety throughout treatment. Changes in the generally stable characteristic of trait anxiety indicate the profoundly life-altering nature of chemotherapy. The time point before the start of chemotherapy was identified as the most anxiety-provoking, calling for interventions to be delivered as early as possible in the treatment trajectory. Patients with high trait anxiety and symptom burden may benefit from additional support.