Patient Versus Caregiver: Correlation and Differences in Pediatric Quality of Life Using a Prospective Registry in a Large-Volume, Multisite Practice

Purpose: Patient-reported outcomes provide quality of life (QOL) data during and after radiation. When pediatric patients are unable to complete patient-reported outcomes, it is unknown whether caregiver responses are an accurate surrogate. We assessed whether caregiver scores for the Pediatric Quality of Life Inventory (PedsQL) Core and Brain Tumor Module questionnaires can substitute for missing child scores. Methods and Materials: From 2016 to 2018, pediatric patients treated with radiation were followed in a prospective, institutional registry. Child and caregiver Core and Tumor PedsQL surveys were obtained at pretreatment, end of treatment, and in regular follow-up. The differences between the 2 scores at each time point were quantified using a linear mixed-model and the level of agreement was estimated with intraclass correlation coefficient (ICC). An ICC 95% confidence interval (CI) lower limit exceeding 0.75 was considered an acceptable threshold for using caregiver scores as imputed values for missing child scores. Results: Ninety-one children completed 403 surveys. Caregivers underestimated QOL scores at baseline, but not at end of treatment or any follow-up time. The PedsQL Core total score had an ICC of 0.88 (95% CI, 0.81-0.92), and the emotional, physical, school, and social function subdomain scores were 0.81 (0.72-0.88), 0.72 (0.58-0.82), 0.79 (0.68-0.86), and 0.75 (0.62-0.83), respectively. The tumor total score ICC was 0.91 (0.85, 0.94), and each of the subdomains (cognitive problems, communication, movement and balance, nausea, pain and hurt, perceived physical appearance, procedural anxiety, treatment anxiety, and worry) had ICC lower bound 95% CI ≥0.75 except for communication (0.83, 0.74-0.89). Bland-Altman analysis demonstrated no visual change in discrepancy between child and caregiver estimates as overall QOL improved. Conclusions: Agreement between child- and caregiver-reported QOL was generally strong in the acute period after radiation, implying that caregiver scores may be imputed for child scores in future protocols and analyses of pediatric QOL.