Statistical analysis

Sociodemographic and clinical characteristics of patients and family caregivers were described by medians for continuous variables and proportions for categorical variables. The statistical significance of differences between patient age groups was assessed with the Wilcoxon rank sum test and Fisher’s exact test. To detect any potential effect of attrition on the outcomes of interest, we also calculated completion rates as the number of completed questionnaires divided by the number of expected questionnaires.

Changes over time from baseline in total and domain FACT-G scores were evaluated with one-way ANOVA with the Greenhouse-Geisser correction and a QoL and time interaction term. Repeated measures one-way ANOVA was also used for sub analyses of FACT-G domains. Changes of ≥5 points in FACT-G total scores and ≥ 2 points in domain scores were considered clinically meaningful [25]. The statistical significance of changes in patients’ treatment expectations from baseline to the second and third cycles of chemotherapy was assessed with Student’s t-test. Cohen’s kappa coefficient assessed agreement between patients’ and family caregivers’ treatment expectations, with a value of 0 indicating non-agreement and a value of 1 indicating perfect agreement [26]. Missing items in subscales were handled according to the FACIT Administration and Scoring Guidelines [27]; the subscale score was multiplied by the number of items in the subscale, then divided by the number of items answered. A p value < .05 was considered statistically significant. All analyses were conducted using STATA 15 [28].